text
stringlengths 87
880k
| pmid
stringlengths 1
8
| accession_id
stringlengths 9
10
| license
stringclasses 2
values | last_updated
stringlengths 19
19
| retracted
stringclasses 2
values | citation
stringlengths 22
94
| decoded_as
stringclasses 2
values | journal
stringlengths 3
48
| year
int32 1.95k
2.02k
| doi
stringlengths 3
61
| oa_subset
stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|
==== Front
PLoS MedPLoS MedpmedplosmedPLoS Medicine1549-12771549-1676Public Library of Science San Francisco, USA 1623199510.1371/journal.pmed.0020393PerspectivesInfectious DiseasesEpidemiology/Public HealthHealth PolicyHIV/AIDSSexual HealthUrologyHIV Infection/AIDSMedicine in Developing CountriesMen's HealthClinical trialsPublic HealthDoes Male Circumcision Prevent HIV Infection? PerspectiveSiegfried Nandi Nandi Siegfried is a South African Nuffield Medical Fellow at the University of Oxford, Oxford, United Kingdom. E-mail: [email protected]
Competing Interests: NS is the lead author of a Cochrane systematic review of observational studies of this topic and has advocated for the need for randomised trials to assess the effects of male circumcision. She has no financial or other competing interests in male circumcision.
11 2005 25 10 2005 2 11 e393Copyright: © 2005 Nandi Siegfried.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Randomized, Controlled Intervention Trial of Male Circumcision for Reduction of HIV Infection Risk: The ANRS 1265 Trial
A Landmark Paper in HIV Research?
The First Randomised Trial of Male Circumcision for Preventing HIV: What Were the Ethical Issues?
Siegfried discusses the first reported randomized controlled trial of whether circumcision protects against HIV, published in PLoS Medicine.
==== Body
Given the devastating mortality and morbidity associated with HIV/AIDS, many potential prevention measures against HIV infection have been explored. Male circumcision is one of these, and seven reviews of the literature [1–7], including two meta-analyses [4,5], have been published. However, as pointed out in the Cochrane systematic review of the subject, all studies to date were observational, and many were of poor quality. In the absence of any experimental evidence, no causal relationship between male circumcision and HIV prevention could be confidently assumed [7].
The First Randomised Trial
In this issue of PLoS Medicine, Auvert and colleagues report results from the first completed trial of male circumcision for reducing HIV infection in South African heterosexual men [8]. The authors conducted a randomised and blindly evaluated trial in a semiurban area near Johannesburg in which the background HIV prevalence rate among heterosexual men was 4.4%. Between July 2002 and February 2004, they randomised 3,274 men, with 1,617 undergoing medical circumcision at the beginning of the trial (the intervention group) and 1,657 remaining uncircumcised (the control group). The men were followed up at three clinic visits at 3, 12, and 21 months, and were tested for HIV at each visit.
At an interim analysis done after all participants had completed the 12-month clinic visit, the Data Monitoring and Safety Board stopped the trial on the basis of the interim results. The results showed that, after excluding those men who were HIV-positive at the beginning of the trial (n = 146), 20 of the circumcised men became infected with HIV during the trial compared with 49 men in the control group. The risk of acquiring HIV infection was significantly reduced by 60% in the men who had undergone circumcision (incidence rate ratio = 0.4; p < 0.001). All men in the control group were then offered circumcision.
Strengths of the Study
One strength of this study is that participants were drawn from the general population, increasing the generalisability of the findings. In addition, the relatively low loss to follow-up (7.9% overall) demonstrates that trials of this nature can be adequately conducted in poorly resourced settings.
The researchers, the participants, and their supporters should be congratulated for attempting and successfully completing a trial as complex as this. The interim analysis was planned, and used an appropriate statistical stopping rule, thus reducing the chance of randomly exaggerated treatment effects [9].
Concerns about the Randomisation
The researchers did not report how the randomisation sequence was generated, and they used an unusual form of allocation to comparison groups. As participants requested to be actively involved in the allocation process, they were invited to choose an envelope from a pre-prepared box of ten envelopes. After each allocation, this box was then refilled with envelopes from another box containing five envelopes for the intervention group and five for the control group. Although the envelopes were equally distributed in the second box, this was not necessarily the case with the participant box. This unequal distribution partly explains why the numbers in the comparison groups differ by a total of 40.
Perhaps of more concern is that allocation concealment may have been inadequate given that the centre manager was responsible for filling both boxes of envelopes, and potentially could have subverted the treatment allocation, thus introducing selection bias. Inadequate allocation concealment has been shown to be associated with exaggerated treatment effects [10,11]. Despite this, as the baseline characteristics of both comparison groups are similar and the sensitivity analyses of the results are robust, the effect of these quality parameters is probably negligible.
Ethical Concerns
Although it is unlikely to have affected the results, the trialists decided not to inform participants of their HIV status, neither at the beginning or end nor during the trial. The authors argue that at the time of the trial, antiretrovirals were unavailable in the public health sector in South Africa, and that the participants received intensive counselling on how to avoid contracting and transmitting HIV. The researchers chose not to inform those who were HIV-positive at the beginning because they were concerned that exclusion from the trial would be stigmatising. Men were encouraged to attend voluntary counselling and testing (VCT) in a nearby public clinic or in a special VCT centre in the same building as the investigation centre. Men who tested positive for other sexually transmitted infections that were treatable locally were referred to the local clinic.
Avoiding stigmatisation is an important consideration, but it could be argued that if withholding HIV status was the only feasible option, then the conditions were not suitable to conduct such a trial. In fact, referring participants to the local VCT centre or the clinic for treatment of other sexually transmitted infections may have resulted in more stigmatisation, as this referral would have been visible to others rather than only to those within the trial processes. Some might argue that the nondisclosure of HIV status fails the test of beneficence (the obligation to prevent and remove harms and to promote the good of a person by minimising the risks incurred to the research participant and maximising the benefits to them and others). Not only were participants affected by nondisclosure, but so were their partners. It is unlikely that this approach would be tolerated in a more developed setting [12].
Policy Implications
The trialists suggest that circumcision could be rapidly incorporated into national plans of countries where circumcision is not widely practised, while recognising that promotion of circumcision may also lead to undesirable outcomes such as undermining condom promotion. They are right to argue that we need to seriously consider circumcision as a potential prevention method, but it seems wise to be more cautious in making recommendations for policy. Within- and between-country differences in culture, religion, and social norms will need to be very carefully considered before implementing circumcision programmes. Crucially, the results of two additional trials underway in Uganda and Kenya are awaited. Considering the results of all three trials together is likely to provide us with stronger evidence to guide policy.
I am deeply grateful to Michael Parker and Iain Chalmers for their advice and comments on the draft paper.
Citation: Siegfried N (2005) Does male circumcision prevent HIV infection? PLoS Med 2(11): e393.
Abbreviation
VCTvoluntary counselling and testing
==== Refs
References
Moses S Plummer FA Bradley JE Ndiya-Achola JO Nagelkerke NJD The association between lack of male circumcision and risk for HIV infection: A review of the epidemiological data Sex Transm Dis 1994 21 201 210 7974070
DeVincenzi ID Mertens T Male circumcision: A role in HIV prevention? AIDS 1994 8 153 160 8043224
Moses S Bailey RC Donald AR Male circumcision: Assessment of health benefits and risks Sex Transm Dis 1998 74 368 373
Howe RV Circumcision and HIV infection: Review of the literature and meta-analysis Int J STD AIDS 1999 10 8 16 10215123
Weiss HA Quigley MA Hayes R Male circumcision and risk of HIV infection in sub-Saharan Africa: A systematic review and meta-analysis AIDS 2000 14 2361 2370 11089625
Bailey RC Plummer FA Moses S Male circumcision and HIV prevention: Current knowledge and future research directions Lancet Infect Dis 2001 1 223 230 11871509
Siegfried N Muller M Deeks J Volmink J Egger M HIV and male circumcision—A systematic review with assessment of the quality of studies Lancet Infect Dis 2005 5 165 173 15766651
Auvert B Taljaard D Lagarde E Sobngwi-Tambekou J Sitta R Randomized, controlled intervention trial of male circumcision for reduction of HIV infection risk: The ANRS 1265 trial PLoS Med 2005 2 e298 10.1371/journal.pmed.0020298 16231970
Schulz KF Grimes DA Multiplicity in randomised trials II: Subgroup and interim analyses Lancet 2005 365 1657 1661 15885299
Schulz KF Chalmers I Hayes RJ Altman DG Empirical evidence of bias. Dimensions of methodological quality associated with estimates of treatment effects in controlled trials JAMA 1995 273 408 412 7823387
Kunz R Vist G Oxman A Randomisation to protect against selection bias in health-care trials Cochrane Database Syst Rev 2002 2002 CD001855
Clark P AIDS research in developing countries: Do the ends justify the means? Med Sci Monit 2002 8 ED5 ED16 12218951
|
16231995
|
PMC1262558
|
CC BY
|
2021-01-05 10:39:20
|
no
|
PLoS Med. 2005 Nov 25; 2(11):e393
|
utf-8
|
PLoS Med
| 2,005 |
10.1371/journal.pmed.0020393
|
oa_comm
|
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-2311617152910.1186/1471-2105-6-231Research ArticleA linear memory algorithm for Baum-Welch training Miklós István [email protected] Irmtraud M [email protected] MTA-ELTE Theoretical Biology and Ecology Group, Pázmány Péter sétány 1/c 1117 Budapest, Hungary.2 European Bioinformatics Institute, Wellcome Trust Genome Campus, Cambridge CB10 1SD, UK.2005 19 9 2005 6 231 231 24 6 2005 19 9 2005 Copyright © 2005 Miklós and Meyer; licensee BioMed Central Ltd.2005Miklós and Meyer; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background:
Baum-Welch training is an expectation-maximisation algorithm for training the emission and transition probabilities of hidden Markov models in a fully automated way. It can be employed as long as a training set of annotated sequences is known, and provides a rigorous way to derive parameter values which are guaranteed to be at least locally optimal. For complex hidden Markov models such as pair hidden Markov models and very long training sequences, even the most efficient algorithms for Baum-Welch training are currently too memory-consuming. This has so far effectively prevented the automatic parameter training of hidden Markov models that are currently used for biological sequence analyses.
Results:
We introduce the first linear space algorithm for Baum-Welch training. For a hidden Markov model with M states, T free transition and E free emission parameters, and an input sequence of length L, our new algorithm requires O(M) memory and O(LMTmax (T + E)) time for one Baum-Welch iteration, where Tmax is the maximum number of states that any state is connected to. The most memory efficient algorithm until now was the checkpointing algorithm with O(log(L)M) memory and O(log(L)LMTmax) time requirement. Our novel algorithm thus renders the memory requirement completely independent of the length of the training sequences. More generally, for an n-hidden Markov model and n input sequences of length L, the memory requirement of O(log(L)Ln-1 M) is reduced to O(Ln-1 M) memory while the running time is changed from O(log(L)Ln MTmax + Ln(T + E)) to O(Ln MTmax (T + E)).
An added advantage of our new algorithm is that a reduced time requirement can be traded for an increased memory requirement and vice versa, such that for any c ∈ {1, ..., (T + E)}, a time requirement of Ln MTmax c incurs a memory requirement of Ln-1 M(T + E - c).
Conclusion
For the large class of hidden Markov models used for example in gene prediction, whose number of states does not scale with the length of the input sequence, our novel algorithm can thus be both faster and more memory-efficient than any of the existing algorithms.
==== Body
Background
Hidden Markov Models (HMMs) are widely used in Bioinformatics [1], for example, in protein sequence alignment, protein family annotation [2,3] and gene-finding [4,5].
When an HMM consisting of M states is used to annotate an input sequence, its predictions crucially depend on its set of emission probabilities ε and transition probabilities . This is for example the case for the state path with the highest overall probability, the so-called optimal state path or Viterbi path [6], which is often reported as the predicted annotation of the input sequence.
When a new HMM is designed, it is usually quite easy to define its states and the transitions between them as these typically closely reflect the underlying problem. However, it can be quite difficult to assign values to its emission probabilities ε and transition probabilities . Ideally, they should be set up such that the model's predictions would perfectly reproduce the known annotation of a large and diverse set of input sequences.
The question is thus how to derive the best set of transition and emission probabilities from a given training set of annotated sequences. Two main scenarios have to be distinguished [1]:
(1) If we know the optimal state paths that correspond to the known annotation of the training sequences, the transition and emission probabilities can simply be set to the respective count frequencies within these optimal state paths, i.e. to their maximum likelihood estimators. If the training set is small or not diverse enough, pseudo-counts have to be added to avoid over-fitting.
(2) If we do not know the optimal state paths of the training sequences, either because their annotation is unknown or because their annotation does not unambiguously define a state path in the HMM, we can employ an expectation maximisation (EM) algorithm [7] such as the Baum-Welch algorithm [8] to derive the emission and transition probabilities in an iterative procedure which increases the overall log likelihood of the model in each iteration and which is guaranteed to converge at least to a local maximum. As in case (1), pseudo-counts or Dirichlet priors can be added to avoid over-fitting when the training set is small or not diverse enough.
Methods and results
Baum-Welch training
The Baum-Welch algorithm defines an iterative procedure in which the emission and transition probabilities in iteration n + 1 are set to the number of times each transition and emission is expected to be used when analysing the training sequences with the set of emission and transition probabilities derived in the previous iteration n.
Let Ti,jn
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGubavdaqhaaWcbaGaemyAaKMaeiilaWIaemOAaOgabaGaemOBa4gaaaaa@3306@ denote the transition probability for going from state i to state j in iteration n, Ein(y)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGfbqrdaqhaaWcbaGaemyAaKgabaGaemOBa4gaaOGaeiikaGIaemyEaKNaeiykaKcaaa@33E2@ the emission probability for emitting letter y in state i in iteration n, P(X) the probability of sequence X, and xk the kth letter in input sequence X which has length L. We also define Xk as the sequence of letters from the beginning of sequence X up to sequence position k, (x1, ...xk). Xk is defined as the sequence of letters from sequence position k + 1 to the end of the sequence, (xk+1, ...xL).
For a given set of training sequences, S, the expectation maximisation update for transition probability Ti,jn
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGubavdaqhaaWcbaGaemyAaKMaeiilaWIaemOAaOgabaGaemOBa4gaaaaa@3306@, Ti,jn+1
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGubavdaqhaaWcbaGaemyAaKMaeiilaWIaemOAaOgabaGaemOBa4Maey4kaSIaeGymaedaaaaa@34D8@ can then be written as
Ti,jn+1=∑X∈Sti,jn(X)/P(X)∑j′∑X∈Sti,jn,(X)/P(X)where ti,jn(X):=∑k=1Lfn(Xk,i)Ti,jnEjn(xk+1)bn(Xk+1,j) (1)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaafaqaaeGabaaabaGaemivaq1aa0baaSqaaiabdMgaPjabcYcaSiabdQgaQbqaaiabd6gaUjabgUcaRiabigdaXaaakiabg2da9maalaaabaaccaGae8xeIu+aaSbaaSqaaiabdIfayjabgIGiolabdofatbqabaGccqWG0baDdaqhaaWcbaGaemyAaKMaeiilaWIaemOAaOgabaGaemOBa4gaaOGaeiikaGIaemiwaGLaeiykaKIaei4la8IaemiuaaLaeiikaGIaemiwaGLaeiykaKcabaGae8xeIu+aaSbaaSqaaiqbdQgaQzaafaaabeaakiab=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@AC24@
The superfix n on the quantities on the right hand side indicates that they are based on the transition probabilities Ti,jn
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGubavdaqhaaWcbaGaemyAaKMaeiilaWIaemOAaOgabaGaemOBa4gaaaaa@3306@ and emission probabilities Ein(xk+1)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGfbqrdaqhaaWcbaGaemyAaKgabaGaemOBa4gaaOGaeiikaGIaemiEaG3aaSbaaSqaaiabdUgaRjabgUcaRiabigdaXaqabaGccqGGPaqkaaa@3747@ of iteration n. f(Xk, i): = P(x1, ...xk, s(xk) = i) is the so-called forward probability of the sequence up to and including sequence position k, requiring that sequence letter xk is read by state i. It is equal to the sum of probabilities of all state paths that finish in state i at sequence position k. The probability of sequence X, P(X), is therefore equal to f(XL, End). b(Xk, i): = P(xk+1, ...xL|s(xk) = i) is the so-called backward probability of the sequence from sequence position k + 1 to the end, given that the letter at sequence position k, xk, is read by state i. It is equal to the sum of probabilities of all state paths that start in state i at sequence position k.
For a given set of training sequences, S, the expectation maximisation update for emission probability Ein(y)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGfbqrdaqhaaWcbaGaemyAaKgabaGaemOBa4gaaOGaeiikaGIaemyEaKNaeiykaKcaaa@33E2@, Ein+1(y)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGfbqrdaqhaaWcbaGaemyAaKgabaGaemOBa4Maey4kaSIaeGymaedaaOGaeiikaGIaemyEaKNaeiykaKcaaa@35B4@, is
Ein+1(y)=∑X∈Sein(y,X)/P(X)∑y'∑X∈Sein(y',X)/P(X)where ein(y,X):=∑k=1Lδxk,yfn(Xk,i)bn(Xk,i) (2)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@A18B@
δ is the usual delta function with δxk,y
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqaH0oazdaWgaaWcbaGaemiEaG3aaSbaaWqaaiabdUgaRbqabaWccqGGSaalcqWG5bqEaeqaaaaa@33E7@ = 1 if xk = y and δxk,y
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqaH0oazdaWgaaWcbaGaemiEaG3aaSbaaWqaaiabdUgaRbqabaWccqGGSaalcqWG5bqEaeqaaaaa@33E7@ = 0 if xk ≠ y. As before, the superfix n on the quantities on the right hand side indicates that they are calculated using the transition probabilities Ti,jn
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGubavdaqhaaWcbaGaemyAaKMaeiilaWIaemOAaOgabaGaemOBa4gaaaaa@3306@ and emission probabilities Ein(xk+1)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGfbqrdaqhaaWcbaGaemyAaKgabaGaemOBa4gaaOGaeiikaGIaemiEaG3aaSbaaSqaaiabdUgaRjabgUcaRiabigdaXaqabaGccqGGPaqkaaa@3747@ of iteration n.
The forward and backward probabilities fn (Xk, i) and bn(Xk, i) can be calculated using the forward and backward algorithms [1] which are introduced in the following section.
Baum-Welch training using the forward and backward algorithm
The forward algorithm proposes a procedure for calculating the forward probabilities f(Xk, i) in an iterative way. f(Xk, i) is the sum of probabilities of all state paths that finish in state i at sequence position k.
The recursion starts with the initialisation
f(X0,i)={1 i=Start0 i≠Start
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGMbGzcqGGOaakcqWGybawdaWgaaWcbaGaeGimaadabeaakiabcYcaSiabdMgaPjabcMcaPiabg2da9maaceaabaqbaeqabiqaaaqaaiabigdaXiaaykW7caaMc8UaaGPaVlabdMgaPjabg2da9iabdofatjabdsha0jabdggaHjabdkhaYjabdsha0bqaaiabicdaWiaaykW7caaMc8UaaGPaVlabdMgaPjabgcMi5kabdofatjabdsha0jabdggaHjabdkhaYjabdsha0baaaiaawUhaaaaa@54AC@
where Start is the number of the start state in the HMM. The recursion proceeds towards higher sequence positions
f(Xk+1,i)=∑j=1Mf(Xk,j)Tj,iEi(xk+1)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGMbGzcqGGOaakcqWGybawdaWgaaWcbaGaem4AaSMaey4kaSIaeGymaedabeaakiabcYcaSiabdMgaPjabcMcaPiabg2da9maaqahabaGaemOzaygaleaacqWGQbGAcqGH9aqpcqaIXaqmaeaacqWGnbqta0GaeyyeIuoakiabcIcaOiabdIfaynaaBaaaleaacqWGRbWAaeqaaOGaeiilaWIaemOAaOMaeiykaKIaemivaq1aaSbaaSqaaiabdQgaQjabcYcaSiabdMgaPbqabaGccqWGfbqrdaWgaaWcbaGaemyAaKgabeaakiabcIcaOiabdIha4naaBaaaleaacqWGRbWAcqGHRaWkcqaIXaqmaeqaaOGaeiykaKcaaa@549C@
and terminates with
P(X)=P(XL)=f(XL,End)=∑j=1Mf(XL,j)Tj,End
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGqbaucqGGOaakcqWGybawcqGGPaqkcqGH9aqpcqWGqbaucqGGOaakcqWGybawdaWgaaWcbaGaemitaWeabeaakiabcMcaPiabg2da9iabdAgaMjabcIcaOiabdIfaynaaBaaaleaacqWGmbataeqaaOGaeiilaWIaemyrauKaemOBa4MaemizaqMaeiykaKIaeyypa0ZaaabCaeaacqWGMbGzaSqaaiabdQgaQjabg2da9iabigdaXaqaaiabd2eanbqdcqGHris5aOGaeiikaGIaemiwaG1aaSbaaSqaaiabdYeambqabaGccqGGSaalcqWGQbGAcqGGPaqkcqWGubavdaWgaaWcbaGaemOAaOMaeiilaWIaemyrauKaemOBa4Maemizaqgabeaaaaa@5975@
where End is the number of the end state in the HMM. The recursion can be implemented as a dynamic programming procedure which works its way through a two-dimensional matrix, starting at the start of the sequence in the Start state and finishing at the end of the sequence in the End state of the HMM.
The backward algorithm calculates the backward probabilities b(Xk, i) in a similar iterative way. b(Xk, i) is the sum of probabilities of all state paths that start in state i at sequence position k. Opposed to the forward algorithm the backward algorithm starts at the end of the sequence in the End state and finishes at the start of the sequence in the Start state of the HMM.
The backward algorithm starts with the initialisation
b(XL,i)={1 i=End0 i≠End
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGIbGycqGGOaakcqWGybawdaahaaWcbeqaaiabdYeambaakiabcYcaSiabdMgaPjabcMcaPiabg2da9maaceaabaqbaeqabiqaaaqaaiabigdaXiaaykW7caaMc8UaaGPaVlabdMgaPjabg2da9iabdweafjabd6gaUjabdsgaKbqaaiabicdaWiaaykW7caaMc8UaaGPaVlabdMgaPjabgcMi5kabdweafjabd6gaUjabdsgaKbaaaiaawUhaaaaa@4ED8@
and continues towards lower sequence positions with the recursion
b(Xk,i)=∑j=1MEi(xk)Ti,jb(Xk+1,j)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGIbGycqGGOaakcqWGybawdaahaaWcbeqaaiabdUgaRbaakiabcYcaSiabdMgaPjabcMcaPiabg2da9maaqahabaGaemyrau0aaSbaaSqaaiabdMgaPbqabaaabaGaemOAaOMaeyypa0JaeGymaedabaGaemyta0eaniabggHiLdGccqGGOaakcqWG4baEdaWgaaWcbaGaem4AaSgabeaakiabcMcaPiabdsfaunaaBaaaleaacqWGPbqAcqGGSaalcqWGQbGAaeqaaOGaemOyaiMaeiikaGIaemiwaG1aaWbaaSqabeaacqWGRbWAcqGHRaWkcqaIXaqmaaGccqGGSaalcqWGQbGAcqGGPaqkaaa@52A7@
and terminates with
P(X)=b(X1,Start)=∑j=1MTStart,jb(X1,j)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGqbaucqGGOaakcqWGybawcqGGPaqkcqGH9aqpcqWGIbGycqGGOaakcqWGybawdaahaaWcbeqaaiabigdaXaaakiabcYcaSiabdofatjabdsha0jabdggaHjabdkhaYjabdsha0jabcMcaPiabg2da9maaqahabaGaemivaq1aaSbaaSqaaiabdofatjabdsha0jabdggaHjabdkhaYjabdsha0jabcYcaSiabdQgaQbqabaGccqWGIbGycqGGOaakcqWGybawdaahaaWcbeqaaiabigdaXaaakiabcYcaSiabdQgaQjabcMcaPaWcbaGaemOAaOMaeyypa0JaeGymaedabaGaemyta0eaniabggHiLdaaaa@5894@
As can be seen in the recursion steps of the forward and backward algorithms described above, the calculation of f(Xk+1, i) requires at most Tmax previously calculated elements f(Xk, j) for j ∈ {1, ..M}. Tmax is the maximum number of states that any state of the model is connected to. Likewise, the calculation of b(Xk, i) refers to at most Tmax elements b(Xk+1, j) for j ∈ {1, ..M}.
In order to continue the calculation of the forward and backward values f(Xk, i) and b(Xk, i) for all states i ∈ {1, ..M} along the entire sequence, we thus only have to memorise M elements.
Baum-Welch training using the checkpointing algorithm
Unit now, the checkpointing algorithm [11-13] was the most efficient way to perform Baum-Welch training. The basic idea of the checkpointing algorithm is to perform the forward and backward algorithm by memorising the forward and backward values only in O(L)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGpbWtcqGGOaakdaGcaaqaaiabdYeambWcbeaakiabcMcaPaaa@30CA@ columns along the sequence dimension of the dynamic programming table. The checkpointing algorithm starts with the forward algorithm, retaining only the forward values in O(L)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGpbWtcqGGOaakdaGcaaqaaiabdYeambWcbeaakiabcMcaPaaa@30CA@ columns. These columns partition the dynamic programming table into O(L)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGpbWtcqGGOaakdaGcaaqaaiabdYeambWcbeaakiabcMcaPaaa@30CA@ separate fields. The checkpointing algorithm then invokes the backward algorithm which memorises the backward values in a strip of length O(L)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGpbWtcqGGOaakdaGcaaqaaiabdYeambWcbeaakiabcMcaPaaa@30CA@ as it moves along the sequence. When the backward calculation reaches the boundary of one field, the pre-calculated forward values of the neighbouring checkpointing column are used to calculate the corresponding forward values for that field. The forward and backward values of that field are then available at the same time and are used to calculate the corresponding values for the EM update.
The checkpointing algorithm can be further refined by using embedded checkpoints. With an embedding level of k, the forward values in O(L1k)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGpbWtcqGGOaakcqWGmbatdaahaaWcbeqaamaaleaameaacqaIXaqmaeaacqWGRbWAaaaaaOGaeiykaKcaaa@3348@ columns of the initial calculation are memorised, thus defining O(L/L1k)=O(Lk−1k)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciGacaGaaeqabaqabeGadaaakeaacqWGpbWtcqGGOaakcqWGmbatcqGGVaWlcqWGmbatdaahaaWcbeqaamaaleaameaacqaIXaqmaeaacqWGRbWAaaaaaOGaeiykaKIaeyypa0Jaem4ta8KaeiikaGIaemitaW0aaWbaaSqabeaadaWcbaadbaGaem4AaSMaeyOeI0IaeGymaedabaGaem4AaSgaaaaakiabcMcaPaaa@3F40@ long fields. When the memory-sparse calculation of the backward values reaches the field in question, the forward algorithm is invoked again to calculate the forward values for O(L1k)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGpbWtcqGGOaakcqWGmbatdaahaaWcbeqaamaaleaameaacqaIXaqmaeaacqWGRbWAaaaaaOGaeiykaKcaaa@3348@ additional columns within that field. This procedure is iterated k times within the thus emerging fields. In the end, for each of the O(L1k)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGpbWtcqGGOaakcqWGmbatdaahaaWcbeqaamaaleaameaacqaIXaqmaeaacqWGRbWAaaaaaOGaeiykaKcaaa@3348@-long k-sub-fields, the forward and backward values are simultaneously available and are used to calculate the corresponding values for the EM update. The time complexity of this algorithm for one Baum-Welch iteration and a given training sequence of length L is O(kLMTmax + L(T + E)), since k forward and 1 backward algorithms have to be invoked, and the memory complexity is O(kL1kM)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGpbWtcqGGOaakcqWGRbWAcqWGmbatdaahaaWcbeqaamaaleaameaacqaIXaqmaeaacqWGRbWAaaaaaOGaemyta0KaeiykaKcaaa@35CA@. For k = log(L), this amounts to a time requirement of O(log(L)LMTmax + L(T + E)) and a memory requirement of O(log(L)M), since L1log(L)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGmbatdaahaaWcbeqaamaaleaameaacqaIXaqmaeaacyGGSbaBcqGGVbWBcqGGNbWzcqGGOaakcqWGmbatcqGGPaqkaaaaaaaa@35F7@ = e.
Baum-Welch training using the new algorithm
It is not trivial to see that the quantities Ti,jn+1
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGubavdaqhaaWcbaGaemyAaKMaeiilaWIaemOAaOgabaGaemOBa4Maey4kaSIaeGymaedaaaaa@34D8@ and Ein+1(y)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGfbqrdaqhaaWcbaGaemyAaKgabaGaemOBa4Maey4kaSIaeGymaedaaOGaeiikaGIaemyEaKNaeiykaKcaaa@35B4@ of Equations 1 and 2 can be calculated in an even more memory-sparse way as both, the forward and the corresponding backward probabilities are needed at the same time in order to calculate the terms fn(Xk,i)Ti,jnEin(xk+1)bn(Xk+1,j)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGMbGzdaahaaWcbeqaaiabd6gaUbaakiabcIcaOiabdIfaynaaBaaaleaacqWGRbWAaeqaaOGaeiilaWIaemyAaKMaeiykaKIaemivaq1aa0baaSqaaiabdMgaPjabcYcaSiabdQgaQbqaaiabd6gaUbaakiabdweafnaaDaaaleaacqWGPbqAaeaacqWGUbGBaaGccqGGOaakcqWG4baEdaWgaaWcbaGaem4AaSMaey4kaSIaeGymaedabeaakiabcMcaPiabdkgaInaaCaaaleqabaGaemOBa4gaaOGaeiikaGIaemiwaG1aaWbaaSqabeaacqWGRbWAcqGHRaWkcqaIXaqmaaGccqGGSaalcqWGQbGAcqGGPaqkaaa@52D1@ in ti,jn(X)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWG0baDdaqhaaWcbaGaemyAaKMaeiilaWIaemOAaOgabaGaemOBa4gaaOGaeiikaGIaemiwaGLaeiykaKcaaa@363B@ and δxk,yfn(Xk,i)bn(Xk,i)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqaH0oazdaWgaaWcbaGaemiEaG3aaSbaaWqaaiabdUgaRbqabaWccqGGSaalcqWG5bqEaeqaaOGaemOzay2aaWbaaSqabeaacqWGUbGBaaGccqGGOaakcqWGybawdaWgaaWcbaGaem4AaSgabeaakiabcYcaSiabdMgaPjabcMcaPiabdkgaInaaCaaaleqabaGaemOBa4gaaOGaeiikaGIaemiwaG1aaWbaaSqabeaacqWGRbWAaaGccqGGSaalcqWGPbqAcqGGPaqkaaa@4742@ in ein(y,X)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGLbqzdaqhaaWcbaGaemyAaKgabaGaemOBa4gaaOGaeiikaGIaemyEaKNaeiilaWIaemiwaGLaeiykaKcaaa@363B@ of Equations 1 and 2. A calculation of these quantities for each sequence position using a memory-sparse implementation (that would memorise only M values at a time) both for the forward and backward algorithm would require L-times more time, i.e. significantly more time. Also, an algorithm along the lines of the Hirschberg algorithm [9,10] cannot be applied as we cannot halve the dynamic programming table after the first recursion.
We here propose a new algorithm to calculate the quantities Ti,jn+1
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGubavdaqhaaWcbaGaemyAaKMaeiilaWIaemOAaOgabaGaemOBa4Maey4kaSIaeGymaedaaaaa@34D8@ and Ein+1(y)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGfbqrdaqhaaWcbaGaemyAaKgabaGaemOBa4Maey4kaSIaeGymaedaaOGaeiikaGIaemyEaKNaeiykaKcaaa@35B4@ which are required for Baum-Welch training. Our algorithm requires O(M) memory and O(LMTmax (T + E)) time rather than O(log(L)M) memory and O(log(L{LMTmax + L(T + E)) time.
The trick for coming up with a memory efficient algorithm is to realise that
• ti,jn(X)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWG0baDdaqhaaWcbaGaemyAaKMaeiilaWIaemOAaOgabaGaemOBa4gaaOGaeiikaGIaemiwaGLaeiykaKcaaa@363B@ and ein(y,X)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGLbqzdaqhaaWcbaGaemyAaKgabaGaemOBa4gaaOGaeiikaGIaemyEaKNaeiilaWIaemiwaGLaeiykaKcaaa@363B@ in Equations 1 and 2 can be interpreted as a weighted sum of probabilities of state paths that satisfy certain constraints and that
• the weight of each state path is equal to the number of times that the constraint is fulfilled.
For example, ti,jn(X)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWG0baDdaqhaaWcbaGaemyAaKMaeiilaWIaemOAaOgabaGaemOBa4gaaOGaeiikaGIaemiwaGLaeiykaKcaaa@363B@ in the numerator in Equation 1 is the weighted sum of probabilities of state paths for sequence X that contain at least one i → j transition, and the weight of each such state path in the sum is the number of times this transition occurs in the state path.
We now show how ti,jn(X)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWG0baDdaqhaaWcbaGaemyAaKMaeiilaWIaemOAaOgabaGaemOBa4gaaOGaeiikaGIaemiwaGLaeiykaKcaaa@363B@ in Equation 1 can be calculated in O(M) memory and O(LMTmax) time. As the superfix n is only there to remind us that the calculation of ti,jn(X)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWG0baDdaqhaaWcbaGaemyAaKMaeiilaWIaemOAaOgabaGaemOBa4gaaOGaeiikaGIaemiwaGLaeiykaKcaaa@363B@ is based on the transition and emission probabilities of iteration n and as this index does not change in the calculation of ti,jn
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWG0baDdaqhaaWcbaGaemyAaKMaeiilaWIaemOAaOgabaGaemOBa4gaaaaa@3346@, we discard it for simplicity sake in the following.
Let ti, j (Xk, l) denote the weighted sum of probabilities of state paths that finish in state l at sequence position k of sequence X and that contain at least one i → j transition, where the weight for each state path is equal to its number of i → j transitions.
Theorem 1: The following algorithm calculates ti, j (X) in O(M) memory and O(LMTmax) time. ti, j (X) is the weighted sum of probabilities of all state paths for sequence X that have at least one i → j transition, where the weight for each state path is equal to its number of i → j transitions.
The algorithm starts with the initialisation
f(X0,m) ={1 m=Start0 m≠Startti,j(X0,m)=0
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@6B92@
and proceeds via the following recursion
f(Xk+1,m) =∑n=1Mf(Xk,n)Tn,mEm(xk+1)ti,j(Xk+1,m)={∑n=1Mti,j(Xk,n)Tn,mEm(xk+1) m≠jf(Xk,i)Ti,mEm(xk+1)+ m=j∑n=1Mti,j(Xk,n)Tn,mEm(xk+1) (3)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@F84B@
and finishes with
P(X)=f(XL,End) =∑n=1Mf(XL,n)Tn,Endti,j(X)=ti,j(XL,End)={∑n=1Mti,j(XL,n)Tn,Endf(XL,i)Ti,End+∑n=1Mti,End(Xk,n)Tn,EndEnd≠jEnd=j (4)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@D08E@
Proof:
(1) It is obvious that the recursion requires only O(M) memory as the calculation of all f(Xk+1, m) values with m ∈ {1, ..M} requires only access to the M previous f(Xk, n) values with n ∈ {1, ..M}. Likewise, the calculations of all ti, j(Xk+1, m) values with m ∈ {1, ..M} refer only to M elements ti, j(Xk, n) with n ∈ {1, ..M}. We therefore have to remember only a thin "slice" of ti, j and f values at sequence position k in order to be able to calculate the ti, j and f values for the next sequence position k + 1. The time requirement to calculate ti, j is O(LMTmax): for every sequence position and for every state in the HMM, we have to sum at most Tmax terms in order to calculate the backward and forward terms.
(2) The f(Xk, m) values are identical to the previously defined forward probabilities and are calculated in the same way as in the forward algorithm.
(3) We now prove by induction that ti, j(Xk, l) is equal to the weighted sum of probabilities of state paths that have at least one i → j transition and that finish at sequence position k in state l, the weight of each state path being equal to its number of i → j transitions.
Initialisation step (sequence position k = 0): ti, j(X0, m) = 0 is true as the sum of probabilities of state paths that finish in state m at sequence position 0 and that have at least one i → j transition is zero. Induction step k → k + 1: We now show that if Equation 3 is true for sequence position k, it is also true for k + 1. We have to distinguish two cases:
(i) case m = j:
ti,j(Xk+1,m)=f(Xk,i)Ti,jEj(xk+1)+ (5)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWG0baDdaWgaaWcbaGaemyAaKMaeiilaWIaemOAaOgabeaakiabcIcaOiabdIfaynaaBaaaleaacqWGRbWAcqGHRaWkcqaIXaqmaeqaaOGaeiilaWIaemyBa0MaeiykaKIaeyypa0JaemOzayMaeiikaGIaemiwaG1aaSbaaSqaaiabdUgaRbqabaGccqGGSaalcqWGPbqAcqGGPaqkcqWGubavdaWgaaWcbaGaemyAaKMaeiilaWIaemOAaOgabeaakiabdweafnaaBaaaleaacqWGQbGAaeqaaOGaeiikaGIaemiEaG3aaSbaaSqaaiabdUgaRjabgUcaRiabigdaXaqabaGccqGGPaqkcqGHRaWkcaWLjaGaaCzcaiabcIcaOiabiwda1iabcMcaPaaa@569C@
∑n=1Mti,j(Xk,n)Tn,jEj(Xk+1) (6)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaadaaeWbqaaiabdsha0naaBaaaleaacqWGPbqAcqGGSaalcqWGQbGAaeqaaOGaeiikaGIaemiwaG1aaSbaaSqaaiabdUgaRbqabaGccqGGSaalcqWGUbGBcqGGPaqkaSqaaiabd6gaUjabg2da9iabigdaXaqaaiabd2eanbqdcqGHris5aOGaemivaq1aaSbaaSqaaiabd6gaUjabcYcaSiabdQgaQbqabaGccqWGfbqrdaWgaaWcbaGaemOAaOgabeaakiabcIcaOiabdIfaynaaBaaaleaacqWGRbWAcqGHRaWkcqaIXaqmaeqaaOGaeiykaKIaaCzcaiaaxMaacqGGOaakcqaI2aGncqGGPaqkaaa@516A@
The first term, see right hand side of 5, is the sum of probabilities of state paths that finish at sequence position k + 1 and whose last transition is from i → j. The second term, see 6, is the sum of probabilities of state paths that finish at sequence position k + 1 and that already have at least one i → j transition. Note that the term in 6 also contains a contribution for n = i. This ensures that the weight of those state path that already have at least one i → j transition is correctly increased by 1. The sum, ti, j(Xk+1, m), is therefore the weighted sum of probabilities of state paths that finish in sequence position k + 1 and contain at least one i → j transition. Each state path's weight in the sum is equal to its number of i → j transitions.
(ii) case m ≠ j:
∑n=1Mti,j(Xk,n)Tn,jEj(Xk+1) (6)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaadaaeWbqaaiabdsha0naaBaaaleaacqWGPbqAcqGGSaalcqWGQbGAaeqaaOGaeiikaGIaemiwaG1aaSbaaSqaaiabdUgaRbqabaGccqGGSaalcqWGUbGBcqGGPaqkaSqaaiabd6gaUjabg2da9iabigdaXaqaaiabd2eanbqdcqGHris5aOGaemivaq1aaSbaaSqaaiabd6gaUjabcYcaSiabdQgaQbqabaGccqWGfbqrdaWgaaWcbaGaemOAaOgabeaakiabcIcaOiabdIfaynaaBaaaleaacqWGRbWAcqGHRaWkcqaIXaqmaeqaaOGaeiykaKIaaCzcaiaaxMaacqGGOaakcqaI2aGncqGGPaqkaaa@516A@
The expression on the right hand side is the weighted sum of probabilities of state paths that finish in sequence position k + 1 and contain at least one i → j transition.
We have therefore shown that if Equation 3 is true for sequence position k, it is also true for sequence position k + 1. This concludes the proof of theorem 1. □
It is easy to show that ei(y, X) in Equation 2 can also be calculated in O(M) memory and O(LMTmax) time in a similar way as ti, j(X). Let ei(y, Xk, l) denote the weighted sum of probabilities of state paths that finish at sequence position k in state l and for which state i reads letter y at least once, the weight of each state path being equal to the number of times state i reads letter y. As in the calculation of ti, j(X) we again omit the superfix n as the calculation of ei(y, X) is again entirely based on the transition and emission probabilities of iteration n.
Theorem 2: ei(y, X) can be calculated in O(M) memory and O(LMTmax) time using the following algorithm. ei(y, X) is the weighted sum of probabilities of state paths for sequence X that read letter y in state i at least once, the weight of each state path being equal to the number of times letter y is read by state i.
Initialisation step:
f(X0,m) ={1 m=Start0 m≠Startei(y,X0,m)=0
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@7036@
Recursion:
f(Xk+1,m)=∑n=1Mf(Xk,n)Tn,mEm(xk+1)ei(y,Xk+1,m)={∑n=1Mei(y,Xk,n)Tn,mEm(xk+1)if m≠i or xk+1≠yf(Xk,i)Ti,mEm(xk+1)+∑n=1Mei(y,Xk,n)Tn,mEm(xk+1)if m=i and xk+1=y
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@FD7B@
Termination step:
P(X)=f(XL,End)=∑n+1Mf(XL,n)Tn,End (7)ei(y,X)=ei(y,XL,End)=∑n+1Mei(y,XL,n)Tn,End
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@890A@
Proof: The proof is strictly analogous to the proof of theorem 1.
The above theorems have shown that ti, j(X) and ei(y, X) can each be calculated in O(M) memory and O(LMTmax) time. As there are T transition parameters and E emission parameters to be calculated in each Baum-Welch iteration, and as these T + E values can be calculated independently, the time and memory requirements for each iteration and a set of training sequences whose sum of sequence lengths is L using our new algorithm are
• O(M) memory and O(LMTmax (T + E)) time, if all parameter estimates are calculated consecutively
• O(M(T + E)) memory and O(LMTmax) time, if all parameter estimates are calculated in parallel
• more generally, O(Mc) memory and O(LMTmax (T + E - c)) time for any c ∈ {1,..., (T + E)}, if c of T + E parameters are to be calculated in parallel
Note that the calculation of P(X) is a by-product of each ti, j(X) and each ei(y, X) calculation, see Equations 4 and 7, and that T is equal to the number of free transition parameters in the HMM which is usually smaller than the number of transitions probabilities. Likewise, E is the number of free emission parameters in the HMM which may differ from the number of emission probabilities when the probabilities are parametrised.
Discussion and Conclusion
We propose the first linear-memory algorithm for Baum-Welch training. For a hidden Markov model with M states, T free transition and E free emission parameters, and an input sequence of length L, our new algorithm requires O(M) memory and O(LMTmax (T + E)) time for one Baum-Welch iteration as opposed to O(log(L)M) memory and O(log(L)LMTmax + L(T + E)) time using the checkpointing algorithm [11-13], where Tmax is the maximum number of states that any state is connected to. Our algorithm can be generalised to pair-HMMs and, more generally, n-HMMs that analyse n input sequences at a time in a straightforward way. In the n-HMM case, our algorithm reduces the memory and time requirements from O(log(L)Ln-1 M) memory and O(log(L)Ln MTmax + Ln(T + E)) time to O(Ln-1 M) memory and O(Ln MTmax (T + E))) time. An added advantage of our new algorithm is that a reduced time requirement can be traded for an increased memory requirement and vice versa, such that for any c ∈ {1,..., (T + E)}, a time requirement of Ln MTmax c incurs a memory requirement of Ln-1 M(T + E - c). For HMMs, our novel algorithm renders the memory requirement completely independent of the sequence length. Generally, for n-HMMs and all T + E parameters being estimated consecutively, our novel algorithm reduces the memory requirement by a factor log(L) and the time requirement by a factor log(L)/(T +E) + 1/(MTmax). For all hidden Markov models whose number of states does not depend on the length of the input sequence, this thus amounts to a significantly reduced memory requirement and – in cases where the number of free parameters and states of the model (i.e. T + E) is smaller than the logarithm of sequence lengths – even to a reduced time requirement.
For example, for an HMM that is used to predict human genes, the training sequences have a mean length of at least 2.7·104 bp which is the average length of a human gene [14]. Using our new algorithm, the memory requirement for Baum-Welch training is reduced by a factor of about 28 ≈ e* In (2.7·104) with respect to the most memory-sparse version of the checkpointing algorithm.
Our new algorithm makes use of the fact that the numerators and denominators of Equations 1 and 2 can be decomposed in a smart way that allows a very memory-sparse calculation. This calculation requires only one uni-directional scan along the sequence rather than one or more bi-directional scans, see Figure 1. This property gives our algorithm the added advantage that it is easier to implement as it does not require programming techniques like recursive functions or checkpoints.
Figure 1 Pictorial description of the new algorithm for pair-HMMs. This figure shows a pictorial description of the differences between the forward-backward algorithm (a) and our new algorithm (b) for the Baum-Welch training of a pair-HMM. Each large rectangle corresponds to the projection of the three-dimensional dynamic programming matrix (spanned by the two input sequences X and Y and the states of the HMM) onto the sequence plane. (a) shows how the numerator in Equation 1 is calculated at the pair of sequence positions indicated by the black square using the standard forward and backward algorithm. (b) shows how our algorithm simultaneously calculates a strip of forward values f(Xk, Yq, m) and a strip of ti, j(XkYq, m) values at sequence position k in sequence X in order to estimate ti, j in Equation 1.
Baum-Welch training is only guaranteed to converge to a local optimum. Other optimisation techniques have been developed in order to find better optima. One of the most successful methods is simulated annealing (SA) [1,15]. SA is essentially a Markov chain Monte Carlo (MCMC) in which the target distribution is sequentially changed such that the distribution gets eventually trapped in a local optimum. One can give proposal steps a higher probability as they are approaching locally better points. This can increase the performance of the MCMC method, especially in higher dimensional spaces [16]. One could base the candidate distribution on the expectations such that proposals are more likely to be made near the EM updates (calculated with our algorithm). There is no need to update all the parameters in one MCMC step, modifying a random subset of parameters yields also an irreducible chain. The last feature makes SA significantly faster than Baum-Welch updates as we need to calculate expectations only for a few parameters using SA. In that way, our algorithm could be used for highly efficient parameter training: using our algorithm to calculate the EM updates in only linear space and using SA instead of the Baum-Welch algorithm for fast parameter space exploration.
Typical biological sequence analyses these days often involve complicated hidden Markov models such as pair-HMMs or long input sequences and we hope that our novel algorithm will make Baum-Welch parameter training an appealing and practicable option.
Other commonly employed methods in computer science and Bioinformatics are stochastic context free grammars (SCFGs) which need O(L2 M) memory to analyse an input sequence of length L with a grammar having M non-terminal symbols [1]. For a special type of SCFGs, known as covariance models in Bioinformatics, M is comparable to L, hence the memory requirement is O(L3). This has recently been reduced to O(L2 log(L)) using a divide-and-conquer technique [17], which is the SCFG analogue of the Hirschberg algorithm for HMMs [9]. However, as the states of SCFGs can generally impose long-range correlations between any pair of sequence positions, it seems that our algorithm cannot be applied to SCFGs in the general case.
Authors' contributions
The algorithm is the result of a brainstorming session of the authors on the Genome campus bus back to Cambridge city centre on the evening of the 17th February 2005. Both authors contributed equally.
Acknowledgements
The authors would like to thank one referee for the excellent comments. I.M. is supported by a Békésy György postdoctoral fellowship. Both authors wish to thank Nick Goldman for inviting I.M. to Cambridge.
==== Refs
Durbin R Eddy S Krogh A Mitchison G Biological sequence analysis 1998 Cambridge University Press
Krogh A Brown M Mian IS Sjölander K Haussler D Hidden Markov models in biology: Applications to protein modelling J Mol Biol 1994 235 1501 1531 8107089 10.1006/jmbi.1994.1104
Eddy S HMMER: Profile hidden Markov models for biological sequence analysis 2001
Meyer IM Durbin R Comparative ab initio prediction of gene structures using pair HMMs Bioinformatics 2002 18 1309 1318 12376375 10.1093/bioinformatics/18.10.1309
Meyer IM Durbin R Gene structure conservation aids similarity based gene prediction Nucleic Acids Research 2004 32 776 783 14764925 10.1093/nar/gkh211
Viterbi A Error bounds for convolutional codes and an assymptotically optimum decoding algorithm IEEE Trans Infor Theor 1967 260 269 10.1109/TIT.1967.1054010
Dempster AP Laird NM Rubin DB Maximum likelihood from incomplete data via the EM algorithm J Roy Stat Soc B 1977 39 1 38
Baum LE An equality and associated maximization technique in statistical estimation for probabilistic functions of Markov processes Inequalities 1972 3 1 8
Hirschberg DS A linear space algorithm for computing maximal common subsequences Commun ACM 1975 18 341 343 10.1145/360825.360861
Myers EW Miller W Optimal alignments in linear space CABIOS 1988 4 11 17 3382986
Grice JA Hughey R Speck D Reduced space sequence alignment CABIOS 1997 13 45 53 9088708
Tarnas C Hughey R Reduced space hidden Markov model training Bioinformatics 1998 14 4001 406 10.1093/bioinformatics/14.5.401
Wheeler R Hughey R Optimizing reduced-space sequence analysis Bioinformatics 2000 16 1082 1090 11159327 10.1093/bioinformatics/16.12.1082
International Human Genome Sequencing Consortium Initial sequencing and analysis of the human genome Nature 2001 409 860 921 11237011 10.1038/35057062
Kirkpatrick S Gelatt CD JrVecchi MP Optimization by Simulated Annealing Science 1983 220 671 680
Roberts GO Rosenthal JS Optimal scaling of discrete approximations to Langevin diffusions J R Statist Soc B 1998 60 255 268 10.1111/1467-9868.00123
Eddy S A memory-efficient dynamic programming algorithm for optimal alignment of a sequence to an RNA secondary structure BMC Bioinformatics 2002 3 18 12095421 10.1186/1471-2105-3-18
|
16171529
|
PMC1262623
|
CC BY
|
2021-01-04 16:27:27
|
no
|
BMC Bioinformatics. 2005 Sep 19; 6:231
|
utf-8
|
BMC Bioinformatics
| 2,005 |
10.1186/1471-2105-6-231
|
oa_comm
|
==== Front
PLoS PathogPLoS PathogppatplpaplospathPLoS Pathogens1553-73661553-7374Public Library of Science San Francisco, USA 1624470910.1371/journal.ppat.001001705-PLPA-RA-0098R2plpa-01-02-04Research ArticleCell BiologyParasitologyEukaryotesNoneIdentification of the Moving Junction Complex of Toxoplasma gondii: A Collaboration between Distinct Secretory Organelles MJ Complex of
Toxoplasma gondiiAlexander David L 1Mital Jeffrey 2Ward Gary E 2Bradley Peter 13Boothroyd John C 1*
1 Department of Microbiology and Immunology, Stanford University, Stanford, California, United States of America
2 Department of Microbiology and Molecular Genetics, University of Vermont, Burlington, Vermont, United States of America
3 Department of Microbiology and Immunology, University of California, Los Angeles, Los Angeles, California, United States of America
Haldar Kasturi EditorNorthwestern University Medical School, United States of America* To whom correspondence should be addressed. E-mail: [email protected] 2005 21 10 2005 1 2 e1714 7 2005 12 9 2005 Copyright: © 2005 Alexander et al.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.Apicomplexan parasites, including Toxoplasma gondii and Plasmodium sp., are obligate intracellular protozoa. They enter into a host cell by attaching to and then creating an invagination in the host cell plasma membrane. Contact between parasite and host plasma membranes occurs in the form of a ring-shaped moving junction that begins at the anterior end of the parasite and then migrates posteriorly. The resulting invagination of host plasma membrane creates a parasitophorous vacuole that completely envelops the now intracellular parasite. At the start of this process, apical membrane antigen 1 (AMA1) is released onto the parasite surface from specialized secretory organelles called micronemes. The T. gondii version of this protein, TgAMA1, has been shown to be essential for invasion but its exact role has not previously been determined. We identify here a trio of proteins that associate with TgAMA1, at least one of which associates with TgAMA1 at the moving junction. Surprisingly, these new proteins derive not from micronemes, but from the anterior secretory organelles known as rhoptries and specifically, for at least two, from the neck portion of these club-shaped structures. Homologues for these AMA1-associated proteins are found throughout the Apicomplexa strongly suggesting that this moving junction apparatus is a conserved feature of this important class of parasites. Differences between the contributing proteins in different species may, in part, be the result of selective pressure from the different niches occupied by these parasites.
Synopsis
Among the world's most important pathogens are a group known as the Apicomplexa. These are single-celled, eukaryotic parasites that cause a range of diseases including malaria and some AIDS opportunistic infections, such as toxoplasmosis and cryptosporidiosis. The group shares several properties: first, they are all intracellular parasites that require a host cell in which to grow; second, they all have an extraordinary collection of structures at their front end, the eponymous apical complex; and third, during invasion, each forms an intimate association with the host cell surface. This ring of contact, which migrates down the parasite as invasion proceeds, is termed the moving junction (MJ). Until now, the composition of the MJ has been a complete mystery. Here, the authors identify four proteins that apparently make up the MJ in Toxoplasma gondii and show that the structure is apparently conserved throughout the Apicomplexa, including in the malaria parasites. Surprisingly, forming the MJ appears to be a collaboration between two, completely different secretion organelles within the apical complex. Detailed study of the MJ complex will shed light on what adaptations each parasite has evolved for the hosts and the cell type they infect. It may also represent an important target for prevention and treatment.
Citation:Alexander DL, Mital J, Ward GE, Bradley P, Boothroyd JC (2005) Identification of the moving junction complex of Toxoplasma gondii: A collaboration between distinct secretory organelles. PLoS Pathog 1(2): e17.
==== Body
Introduction
Toxoplasma gondii is an obligate intracellular parasite that belongs to the phylum Apicomplexa. All members of this phylum exhibit a similar process for invading into host cells [1]. The process begins with gliding motility that utilizes a reversible attachment to migrate over the surface of the host cell, perhaps to find a susceptible site for entry. The parasites then establish an intimate association, involving reorientation to put the apical secretory structures in contact with the host membrane [2]. This reorientation is coincident with a transient spike in host plasma membrane conductance, consistent with a break in the bilayer [3]. Next, the specialized secretory organelles in the parasite's apex, called micronemes and rhoptries, secrete their contents [1,4]. A moving junction (MJ), where the host and parasite plasma membranes are in intimate contact, is then formed.
The MJ is associated with a visible constriction around the invading parasite that migrates from anterior to posterior end during invasion. This results in the creation of a parasitophorous vacuolar membrane (PVM) derived from invagination of the host plasma membrane [3–6]. The entire process is highly orchestrated and driven by the parasite's actin and myosin machinery [7–11].
Among other functions, the MJ apparently serves to exclude most host membrane proteins from entering the PVM [4]. Transmission electron micrographs of invading Plasmodium merozoites and Toxoplasma tachyzoites have shown the MJ to be associated with electron-dense structures reminiscent of tight junctions, but the composition of these is not known [5,6]. Numerous proteins involved in invasion have been identified, but only one of these, MCP-1 of P. falciparum, has been shown to be associated with the MJ [12,13]. This protein has no signal peptide or transmembrane domain and is presumed to interact with the MJ complex from the inside of the parasite; hence the key surface molecules that form this complex have remained a mystery.
Apical Membrane Antigen 1 (AMA1) is among the proteins directly implicated in host cell invasion by several members of the phylum, including Toxoplasma, Plasmodium, and Babesia [14–16]. AMA1 is unique to the Apicomplexa and has been localized to the micronemes of developing intracellular parasites and to the apical surface of extracellular parasites just prior to invasion [17–19]. During invasion, AMA1 migrates across the tachyzoite surface from where it is proteolytically shed in a soluble form [18–21]. These observations have led to the hypothesis that AMA1 is one of the parasite adhesins that transiently engage host receptors during the invasion process [22]. This model is supported by the observation that anti-AMA1 antibodies block invasion at a stage following the initial attachment to the host plasma membrane in Toxoplasma [18], Plasmodium [2] and Babesia [16]. As predicted from these data, TgAMA1 is an essential gene, and conditional knockout parasites are significantly impaired in their ability to invade [23].
TgAMA1 and its homologues from other Apicomplexa have the structural characteristics of type-I transmembrane proteins. The ectoplasmic region contains 16 invariant cysteine residues that are present in all AMA1 sequences, indicating a conserved overall topology among the homologues [24,25]. AMA1 from P. vivax has been crystallized, and analysis of the structure suggests a receptor-binding role in invasion, requiring inter-domain interaction [26].
To better understand the precise role of TgAMA1, we sought to identify the parasite proteins with which it associates. We describe here three such proteins, at least two of which, RON2 and RON4, surprisingly, derive from the rhoptry necks, not the micronemes, from which AMA1 is released. At least one, RON4, is secreted upon attachment to host cells, where it then precisely and predominantly localizes to the MJ as it migrates down the length of the invading parasite. In wild-type parasites, only a small amount of TgAMA1 is associated with RON4 at the MJ, but in engineered parasites with reduced amounts of TgAMA1 this proportion rises and a clear concentration of TgAMA1 is evident at the MJ. This crucial complex appears to be conserved across the Apicomplexa.
Results
Identification of AMA1-Associating Proteins
To determine the proteins associating with TgAMA1, immunoprecipitation from detergent extracts of extracellular Toxoplasma tachyzoites was performed using an independent monoclonal antibody (mAb) recognizing one of two distinct TgAMA1 epitopes. The mAb B3.90 [19] detects an epitope in the ectodomain whereas mAb CL22 [18] was raised to a dodecapeptide at the C-terminus of TgAMA1. Each of these antibodies immunoprecipitated TgAMA1 (~70kDa) and near identical profiles of TgAMA1-associating proteins (AAPs; Figure 1A). These AAPs, with apparent sizes of ~145, 130, 110, and 45 kDa, were stably associated with TgAMA1 at high salt concentrations (500 mM NaCl), suggesting a highly specific interaction (unpublished data). The specificity of the AAP association was further demonstrated by immunoblotting the material that co-precipitated with anti-TgAMA1 and probing for the highly abundant MIC2, GRA7, ROP1, or ROP2/3/4 proteins. None of these proteins were detected (unpublished data).
Figure 1 Identification of TgAMA1-Associating Proteins
(A) Monoclonal antibodies CL22 and B3.90 were used to immunoprecipitate TgAMA1 and associating proteins from RIPA lysates of wild-type tachyzoites. The resulting material was resolved by SDS-PAGE followed by Coomassie staining to detect the proteins. Controls were immunoprecipitation with beads alone or an irrelevant mAb specific for the rhoptry bulb protein, ROP1. Each lane represents the product of 107 parasites. The polypeptides specifically associating with TgAMA1 are denoted on the right by their estimated molecular masses in kDa: p145, p130, p110, and p45. Molecular mass markers (in kDa) are indicated on the left of each panel.
(B) mAb CL22 was used to immunoprecipitate TgAMA1 and associating proteins from ΔAMA/AMA1-myc parasites without (−) or with (+)Atc.
(C) Immunoblot of the lanes shown in Figure 1B probed with antibody specific for TgAMA1.
Further evidence for the specificity of the AAPs for TgAMA1 was obtained using a Toxoplasma line in which the endogenous TgAMA1 locus has been disrupted and several tetracycline-repressible myc-tagged copies have been introduced (ΔAMA1/AMA1-myc; [23]). In the absence of added drug, this parasite line expresses ~10% of the constitutive TgAMA1 levels, which is a sufficient level for essentially wild-type levels of invasion and growth. Tetracycline treatment of these parasites reduces TgAMA1 levels to ~0.25% of wild-type amounts. Although such parasites have no apparent defect in motility or initial attachment to the host cell, they are almost completely blocked in their ability to invade [23]. A specific role for TgAMA1 in invasion is further indicated by the observation that if tetracycline is added to cultures in which the parasites have already entered a host cell, there is no effect on further growth within that cell or on egress, but subsequent invasion is blocked. Immunoprecipitation of the AAPs from the ΔAMA1/AMA1-myc line grown in the absence of tetracycline resulted in recovery of the same profile of AAP bands (Figure 1B). As expected from its lower level of expression, markedly less TgAMA1 was recovered on a per-parasite basis, relative to wild-type parasites. For the four AAPs, however, nearly identical amounts were obtained (Figure 1 and unpublished data). As a result, the stoichiometry appeared to go from an excess of TgAMA1 over the AAPs in wild-type parasites to something closer to equal levels in the engineered parasites. This point is discussed further below.
Under tetracycline-repressed conditions, immunoprecipitation with anti-TgAMA1 antibodies yielded much reduced TgAMA1 and the four AAPs (Figure 1B), although all were still detectable by Coomassie staining (Figure 1B) and/or immunoblot (Figure 1C). A variable number of bands below ~50 kDa were also observed in these gels, but apart from AAP45, they were not reproducibly detected in wild-type parasites (Figure 1A) and their abundance was not consistently reduced in the TgAMA1 knock-down conditions [Figure 1B, +Atc (anhydrotetracycline)]. Thus, they were not further pursued.
Mass Spectrometry-Based Proteomic Analysis of TgAMA1 and the AAPs
To determine the identity of the four AAPs, gel slices from the immunoprecipitation experiments were excised, digested with trypsin, and the eluted peptides were subjected to liquid chromatography-electrospray ionization-ion trap mass spectrometry (LC-MS/MS) analysis (Table 1). This analysis also confirmed that the ~70 kDa band was TgAMA1. Included among the identified peptides for this protein was one with an N-terminal sequence not derived from tryptic cleavage (TgAMA1 amino acids 40–54, Table 1). This likely derives from the pro-peptide cleavage of TgAMA1 described previously [18] and indicates cleavage of 17 amino acids to the C-terminal side of the predicted site for removal of the signal peptide. This N-terminal sequence for mature TgAMA1 was confirmed by Edman sequencing of the same immunoprecipitated TgAMA1 (unpublished data).
Table 1 Identification of TgAMA1-Associating Proteins by LC-MS/MS
Mass spectrometric analysis of the three largest AAPs revealed multiple peptides corresponding to Toxoplasma TwinScan predicted sequences Ts0698, Ts5626, and Ts4705 for AAP145, AAP130, and AAP110, respectively. Whether analyzed as a total gel slice representing the sizes from ~100 kDa to ~150 kDa, or as individually excised bands, these three predicted proteins and TgAMA1 were the only proteins for which multiple peptides were identified that met the confidence criteria described in the Materials and Methods (Table 1).
Ts0698 corresponds to a protein of ~157 kDa, with a predicted signal sequence and three transmembrane domains. Proteomic analysis of the Toxoplasma rhoptries also recently identified this protein [27], and using antibodies to recombinant protein, it was shown to derive from the constricted, anterior portion of the rhoptries termed rhoptry necks. Ts0698 was one of four such neck proteins identified and was named RON2. The complete open reading frame was determined by cDNA sequencing; this confirmed the TwinScan prediction of a protein of ~157 kDa. Allowing for removal of a signal peptide, the predicted size (~153 kDa) is slightly larger than that predicted from the mobility on SDS-PAGE (~145 kDa). This leaves open the possibility that it may be processed by proteolytic cleavage to its mature size, a common property of many rhoptry proteins [28] although the discrepancy could also be a result of anomalous migration. There were no apparent functional motifs beyond the signal peptide and three putative hydrophobic transmembrane domains. RON2 is paralogous to two other Toxoplasma-predicted proteins, Ts3110 and Ts0430, but none of the peptides detected here exactly match these other two predicted proteins, and so the identity of p145 could be unambiguously assigned to RON2.
Ts5626 corresponds to a protein that TwinScan predicts should be ~78 kDa with a signal peptide and no predicted transmembrane domains. Proteomic analysis of the Toxoplasma rhoptries also identified this protein and, using antibodies to recombinant protein showed it to also be a rhoptry neck protein [27]. It was, therefore, dubbed RON4. The complete open reading frame was determined by cDNA sequencing. This indicated that the Ts5626 gene, in fact, corresponds to a predicted protein considerably larger (~107 kDa) than the TwinScan algorithm predicted (the TwinScan prediction miscalled some of the exon/intron boundaries, which is not uncommon). Allowing for removal of a signal peptide, the predicted size (~105 kDa) is still much smaller than that suggested by its mobility on SDS-PAGE (~130 kDa). The retarded migration in SDS-PAGE analysis may be explained by the 44 amino acid repeat sequence and the charges within this repeat [27]. There is a paralogous protein identified in the T. gondii database, Ts2928, but none of the peptides detected here exactly match that protein and so the identification of AAP130 as RON4 was unambiguous. No functional motifs, other than a signal peptide, were observed.
Ts4705 predicts a protein of ~179 kDa. This protein was also identified in the rhoptry proteome analysis although neither its size nor its localization were definitively determined [27]. It therefore received no gene name and its true origin remains to be confirmed although all but one of 15 novel proteins that were definitively characterized in the Bradley et al. study proved to be from the rhoptry bulb or neck. It is therefore highly likely that Ts4705 corresponds to a rhoptry protein and given its association with RON2 and RON4 it most likely derives from the rhoptry neck. The discrepancy between the apparent (~110 kDa) and predicted (~179 kDa) sizes for this protein could indicate post-translational processing, as mentioned above for RON2. The identification of Ts4705 peptides in AAP45 is consistent with this possibility (see below) but some of the discrepancy could also be due to abnormal mobility or errors in the TwinScan prediction.
In the excised gel band corresponding to AAP45, two significant peptide hits were identified corresponding to Toxoplasma actin (TgACT1), as well as seven peptides from the predicted C-terminus of Ts4705. Immunoblots with antibodies specific to TgACT1 confirmed that actin was present in the TgAMA1-immmunoselected material (unpublished data). Importantly, however and unlike the Coomassie staining of AAP45, there was no difference in the anti-actin signal using TgAMA1-immunoprecipitated material from untreated or tetracycline-treated ΔAMA1/AMA1-myc parasites and comparable signals were also seen in preparations obtained by immunoprecipitation using antibodies to an irrelevant control protein, ROP1 (unpublished data). These results strongly argue against a specific association of TgACT1 with TgAMA1. Consistent with this, TgACT1 has been proposed to interact with aldolase [29] and Toxofilin [30] yet no peptides corresponding to either of these two proteins were detected by MS, and neither were detected on immunoblots of material immunoprecipitated with anti-TgAMA1 (unpublished data).
The Ts4705 peptides identified in p45 were all derived from the C-terminal region, with no overlap to peptides identified from p110 (Table 1). This is consistent with a proteolytic processing of the Ts4705 predicted polypeptide into p110 and p45 both of which are associated, directly or indirectly, with TgAMA1.
Functional Analysis of TgAMA1 and RON Protein Interactions
The finding that the micronemal protein TgAMA1 associates with at least two rhoptry neck proteins was unexpected and begged the question of whether these proteins associate with TgAMA1 in intact parasites or only when given the artificial opportunity of parasite lysis. To determine whether there is an in vivo interaction between the RONs and TgAMA1, therefore, we used chemical cross-linking of live, cultured parasites followed by immunoprecipitation from parasite extracts denatured so as to disrupt most non-covalent interactions (boiled in the presence of detergent). The cross-linker used was the membrane-impermeable, thiol-cleavable 3,3′-Dithiobis (sulfosuccinimidylpropionate) (DTSSP) and the parasites were from a culture in which invasion was synchronized with a high potassium block [31] that was released simultaneously with the addition of the cross-linker. Under these conditions, any proteins that are exposed to the medium and are within twelve angstroms of each other should be cross-linked, including those that are put out onto the surface during invasion.
Following two washes and quenching of any remaining cross-linker, the fibroblast monolayer and attached parasites were lysed and boiled in 1% SDS and then equilibrated into RIPA buffer. Immunoprecipitation with anti-TgAMA1 was performed, as before, followed by DTT reduction and analysis by SDS-PAGE. Immunoblotting revealed RON4 and RON2 specifically associating with significant amounts of TgAMA1 but only if cross-linked before the precipitation (Figure 2A). The specificity of the cross-linking was demonstrated by showing that the control rhoptry protein ROP1 was not co-immunoprecipitated by any of the heterologous antisera and neither did anti-ROP1 antibodies co-precipitate detectable amounts of any of the other three proteins (Figure 2B). Reciprocal experiments using immunoprecipitation with anti-RON4 yielded TgAMA1 (Figure 2C). Some TgAMA1 is detected in the absence of DTSSP, but cross-linking significantly increased the amount of TgAMA1 co-precipitating with RON4. Interestingly, RON2 was pulled down with anti-RON4 regardless of the presence of cross-linker suggesting a covalent or other strong association between these two proteins (Figure 2C). No co-precipitation of TgAMA1 and RON2, or RON4 was seen using heat-denatured lysates from extracellular parasites, with or without prior cross-linking (not shown). These results indicate that an association between TgAMA1 and RONs occurs during the process of invasion and is surface-exposed. That is, the association is not simply an artifact of parasite lysis although this may increase the amount of association that is seen ( Figure 1).
Figure 2 Immunoprecipitation from Lysates of Parasites Subjected to Chemical Cross-Linking in Live Cultures
Tachyzoites were allowed to synchronously invade fibroblast monolayers in the absence (−) or presence (+) of the homobifunctional cross-linking agent DTSSP. Invasion and cross-linking were allowed to proceed for 15 min, then the reaction was quenched and invasion stopped. Harvested parasites were extracted in 1% SDS and heat denatured, and then processed for immunoprecipitation with antibodies to TgAMA1 (A), ROP1 (B), or RON4 (C). These immunoprecipitates were separated by reducing SDS-PAGE and analyzed by immunoblotting to determine the presence or absence of RON2, RON4, TgAMA1, or ROP1 in the co-precipitating material.
To explore the association further, tachyzoites that were intracellular and ones that were entering and exiting fibroblast monolayers were analyzed by indirect immunofluorescence (IIF) to identify where in the cell TgAMA1, RON4 and RON2 might be associating. Unfortunately, the antisera available for RON2 gave an IIF signal only using methanol-fixation, and not the formaldehyde-fixation conditions needed to preserve good morphology for IIF. While this did allow identification of RON2 as a rhoptry neck protein [27], no signal was observed on invading parasites indicating that either this protein is released into the medium during invasion or, more likely, the structures it is within are not stabilized with this fixative (unpublished data).
Antibodies to RON4, however, worked well using formaldehyde-fixation and so co-localization with TgAMA1 could be assessed. Using intracellular parasites, and as predicted from the literature, anti-RON4 and anti-TgAMA1 antibodies gave the closely apposed but clearly distinct IIF patterns typical of rhoptry necks and micronemes, respectively (Figure 3). RON4 was also detected in the parasitophorous vacuole (Figure 3, arrows) as previously reported [27]. No TgAMA1 is seen in the vacuolar space. These results confirmed that TgAMA1 and the RON2/4 proteins are initially in distinct compartments within the parasite.
Figure 3 Localization of RON4 and TgAMA1 in Intracellular Tachyzoites
DIC and deconvolution processing of IIF were used to image a PV containing eight intracellular parasites that had been fixed with formaldehyde and permeabilized with triton X-100. TgAMA1 was localized with mAb CL22 and Texas-red-conjugated goat-anti-mouse antiserum. RON4 was localized via rabbit antiserum to recombinant RON4 followed by goat-anti-rabbit antiserum conjugated with fluorescein isothyocyante (FITC) (green). The images shown represent an extended focus projection through five 0.1-μm sections after iterative deconvolution. RON4 staining within the vacuole in the regions between the parasites is indicated with an arrow.
Analysis of AAP Localization in Invading Parasites
The above results predicted that the association of TgAMA1 and RON proteins happens only upon initiation of invasion when the micronemes and rhoptries discharge. To examine this, we used IIF and the invasion-synchronization method described above. At the very start of invasion, antibodies to RON4 can be seen staining the apical tip of parasites that are in intimate contact with the host cell and a very small, circumferential ring (Figure 4A–4D). RON4 is not detected on the surface of unattached parasites (unpublished data). Antibodies to surface antigen 1, SAG1, show bright staining under non-permeabilized conditions and this can be used to identify the portion of an invading parasite outside of the host cell up to the point of contact with the host plasma membrane [32]. Subsequent permeabilization and staining for RON4 shows this protein localizes precisely with the limit of SAG1 staining (Figure 4A–4D). As the parasite enters the host cell, this ring of RON4 staining migrates down the length of the parasite, coincident with the constriction that demarcates the MJ (arrows) between host and parasite (Figure 4E–4H). Fully invaded parasites show a focal staining at the posterior end of the parasite where the PVM appears to contact the host plasma membrane (Figure 4I–4L). There is also some apical staining within the parasite, the intensity of which varies depending on the degree of saponin permeabilization (Figure 4I–4L and unpublished data). The RON4 staining of the MJ, including the final posterior focus is fully apparent in the absence of detergent permeabilization (Figure 4M–4T) consistent with the observation that RON4 is outside the portion of the MJ that keeps antibodies out of the nascent parasitophorous vacuole (PV). The focal surface staining at the site of invasion is only observed within about six min after release from the high potassium block to initiate invasion; after that, and likely when the PV has dissociated from the host plasma membrane [3], this posterior focal staining is no longer detected (unpublished data).
Figure 4 Localization of RON4 During Invasion
Tachyzoites were allowed to invade following a potassium shift as described in the text. DIC (A, E, I, M, and Q) and deconvolution IIF were then used to image formaldehyde-fixed parasites as described in Figure 3. The images presented are extended focus projections through ten 0.2-μm sections. Prior to detergent permeabilization, SAG1 was detected with the mAb DG52 followed by goat anti-mouse antiserum conjugated to FITC (green) (B, F, J, N, and R). RON4 detection was with rabbit antisera raised to recombinant RON4 and Texas red-conjugated goat-anti-rabbit antibody; this was done after permeabilization by the addition of saponin (C, G, and K) or without such permeabilization (O, S). The merged images are (D, H, L, P, and T). (A–D) represent a parasite that has just begun invasion as indicated by SAG1 staining over about 80% of the parasite. (E–H) show a parasite about halfway in. (I–L) show a parasite that appears to be fully inside (no SAG1 staining) showing small posterior cap of RON4 and apical staining consistent with rhoptry necks. (M–T) show parasites stained without prior permeabilization. (M–P) also show a parasite fully in, whereas (Q–T) show a parasite at about halfway in. The MJ is indicated by an arrow where it is clearly apparent in the DIC image.
The surface accessibility of RON4 to antibodies suggested that anti-RON4 antiserum might be also inhibitory for invasion, similar to the findings for some antibodies against TgAMA1 and PfAMA1 [2,18,33]. This was tested by pre-incubation of parasites with anti-RON4 antisera and/or inclusion of the antisera upon adding parasites to the human foreskin fibroblasts (HFFs). No significant reduction in invasion efficiency was observed (unpublished data); hence, binding of antibodies to RON4 either does not inhibit the invasion process or the titer of the antibodies used was insufficient to exert an effect.
IIF of invading wild-type parasites using mAb CL22 and mAb B3.90 to detect TgAMA1 confirmed the reported surface localization of TgAMA1 on invading parasites and showed a marked concentration on the posterior region up to the MJ (Figure 5A and 5B). Co-staining with anti-RON4 showed overlap of the two signals at the MJ (Figure 5C and 5D) although the majority of TgAMA1 was not coincident with RON4. Nevertheless, rotation of the three dimensional rendering makes clear the overlap of TgAMA1 throughout the RON4 ring at the MJ (Figure 5E–H). While these results show some co-localization of RON4 and TgAMA1 at the MJ, the majority of TgAMA1 is clearly posterior to this position.
Figure 5 Localization of TgAMA1 and RON4 During Invasion
Tachyzoites were allowed to invade following a potassium shift as for Figure 4. DICA, E, and I) and deconvolution IIF were then used to image formaldehyde-fixed parasites as described in Figure 3, except that the monolayers were permeabilized at the outset and a mAb specific for the cytoplasmic tail of TgAMA1 (CL22) was used in place of anti-SAG1. The images shown here are three-dimensional reconstructions from forty 0.1-μm sections.
(A–D) show a wild-type parasite that is about 40% inside the host cell (the outside portion shows a bright, posterior cap of CL22 staining). (E) shows the axis of rotation for the reconstruction, and the new view of the IIF images present in (B–D) are shown in the corresponding panels (F–H). Images (I–L) show a ΔAMA1/AMA1-myc parasite with low surface AMA1 signal (no posterior cap) that is about 80% invaded into its vacuole. The MJ is indicated by an arrow.
The increase in the stoichiometry of RON4-association with TgAMA1 observed with the ΔAMA1/AMA1-myc parasites suggested that their association might be more apparent in IIF using these parasites. This line is fully invasion-competent, yet it has, on average, only ~10% of the wild-type amount of TgAMA1 [23]. Hence, the overall TgAMA1 staining should be less, but that which is observed may be concentrated in regions on the surface most critical for invasion. Repeating the IIF on invading ΔAMA1/AMA1-myc parasites confirmed the reduced level of TgAMA1 staining relative to wild-type but revealed that approximately one in twenty of these parasites have a level that is substantially lower than the average (i.e., <<10% of wild-type) (Figure 5I and 5J). This reduction could be due to loss of one or more copies of the TgAMA1-myc sequence or exhaustion of the store of TgAMA1 during a previous, abortive invasion step. Regardless, examination of these parasites showed a distinct concentration of TgAMA1 at the MJ that precisely co-localized with RON4 (Figure 5I–5L). Together with the immunoprecipitation data above, these results clearly indicate that TgAMA1 and RON4 form a complex (likely with RON2 and TwinScan 4705) that tracks with and, in the case of RON4, defines the MJ in invading parasites.
Analysis of RON4 Localization Following Ionophore-Induced Egress
Egress involves many of the same phenomena as invasion, including the constriction and an apparent MJ. To address whether the complex identified above might also be present in exiting parasites, we repeated the IIF using ionophore-induced egress [34]. Figure 6B shows that RON4 is indeed detected in rings coincident with constriction as parasites pass through either the PV or the host plasma membrane (solid arrows) as well as at the extreme apical tip of the parasite (open arrow) and in the collapsed parasitophorous vacuole. The ring structure was not detected in the absence of detergent permeabilization (unpublished data), suggesting it is located on the cytoplasmic side of the host cell plasma membrane. TgAMA1 on the apical surface of parasites that protrude from the infected cell is readily detected without detergent permeabilization using mAb B3.90 (unpublished data). Following triton permeabilization, the TgAMA1 staining extends below the constriction observed by differential interference contrast (DIC) light microscopy and can be detected adjacent to and occasionally intersecting with the RON4 ring (Figure 6C and 6D). TgAMA1 is not superimposable throughout the circumferential RON4 staining, contrary to what was observed during invasion.
Figure 6 Localization of RON4 and TgAMA1 During Egress
DIC (A) and IIF (B, C, and D) were used to image wild-type tachyzoites fixed with formaldehyde one minute after calcium-ionophore treatment to induce egress, and permeabilized with triton X-100. (B) and (C) show staining with anti-RON4 (FITC) and anti-TgAMA1 (Texas Red), respectively. (D) is the merged image. Antibody and IIF details are as for Figure 4.
To address the role of TgAMA1 in egress, ΔAMA1/AMA1-myc parasites were grown in the presence of Atc to suppress TgAMA1 levels, and then treated with calcium-ionophore to induce egress. Interestingly, unlike the impairment in invasion, there was no effect on egress under these conditions and RON4 localization to rings in the exiting parasites was unaffected (unpublished data). These results indicate that RON4 can function independently of TgAMA1, and that TgAMA1 is not essential for egress. RON4 may be recruited to the MJ wherever the parasites transit membranes, whereas TgAMA1 is apparently required only for invasion.
Localization of RON4 at the Apical Tip of Parasites in Intimate Contact with Host Cells
Following tetracycline treatment, the majority of ΔAMA1/AMA1-myc parasites will attach to host cells but fail to invade [23]. This is associated with a defect in release of certain rhoptry proteins into the host cell based on studies of the well-characterized ROP1 and ROP2/3/4 markers. We asked whether RON4 secretion was also disrupted in the invasion-defective parasites. IIF using antibodies to RON4 showed that for the attached but un-invaded parasites RON4 is secreted to the apical tip of the parasite (Figure 7) even though no TgAMA1 is detectable (unpublished data). This RON4 staining, however, is generally amorphous and bleb-like with no formation of a ring-like MJ. In rare instances, parasites with nascent RON4 rings are observed in the absence of detectable AMA1 although such parasites appear to be severely retarded in their invasion based on the very short distance they have penetrated into the host cells, despite being given an extended time for invasion. Hence, it appears that TgAMA1 is necessary both for the formation and subsequent movement of the MJ complex.
Figure 7 RON4 localization in ΔAMA1/AMA-myc Tachyzoites Following Tetracycline Repression and Abortive Invasion
DIC (A) and deconvolution IIF (B–D) were used to image formaldehyde-fixed cultures. Images represent a projection through ten 0.2-μm sections. Intracellular ΔAMA1/AMA-myc tachyzoites were grown in Atc-containing media for 30 hr, syringe-released and then synchronized for invasion using a potassium shift as described in the methods. After allowing invasion for 20 min, the cover-slips were fixed and processed for IIF without detergent permeabilization. SAG1 staining ([B] FITC) shows an extracellular parasite in contact with the HFF monolayer. (C) shows the RON4 (Texas Red staining). (D) is a merge of (B) and (C).
RON4 May Be an Essential Protein
To further investigate the function of RON4, attempts were made to delete the RON4 gene. A knock-out vector was constructed to replace the entire RON4 gene with the hypoxanthine-xanthine-guanine phosphoribosyl transferase (HXGPRT) selectable marker. RHΔHXGPRT tachyzoites were transfected with a linearized knock-out cassette and carried through mycophenolic acid/xanthine selection. Despite several attempts, PCR analysis of genomic DNA isolated from drug-resistant populations indicated that the vector had not replaced the endogenous RON4 gene. Although a negative result in such experiments is not definitive, we have been successful in knock-out attempts with many other genes and in at least one of the cases where such attempts have failed [18], subsequent studies confirmed the gene to be essential [23]. Although such negative results are not definitive, they are consistent with the RON4 gene being essential, and suggest that RON4 is an indispensable component of the MJ complex created by these obligate intracellular parasites.
Discussion
In this report, we describe the coordinated secretion of rhoptry necks and micronemes to form a complex of apparently four proteins, TgAMA1, RON4, RON2 and Ts4705, at least of two of which interact at the MJ of invading parasites. LeBrun et. al. have obtained very similar findings to these, i.e., RON4 is at the moving junction as is part of a stable complex with RON2 and Ts4705[35]. Additional RON proteins have been identified as part of a rhoptry proteome analysis [27] leaving open the possibility that these, and perhaps other proteins, also contribute to the MJ complex but, if so, they do not appear to form a stable association with the TgAMA1-associated complex described here.
TgAMA1 and RON4 co-localize with the MJ at the ring of contact between parasite and host plasma membranes. The MJ is a boundary that excludes antibodies from entering the nascent PV, and the fact that RON4 can be detected at the MJ in the absence of detergent permeabilization indicates that it is on the exterior of the MJ both with respect to the parasite and host plasma membranes and the antibody-excluding portion of the MJ. Some of the MJ proteins are likely below this antibody exclusion zone and/or buried beneath the parasite or host cell membranes [4]. This may ultimately prove to be the location of the RON2 and Ts4705 proteins.
In published electron micrographs of Plasmodium [5] and Toxoplasma caught in the act of invading red blood cells, which Toxoplasma can do in vitro [6], the MJ is associated with electron-dense structures of unknown composition and function (although the cytosolic MCP-1 protein may be involved in the Plasmodium parasite [36]). In Plasmodium, the electron-dense ring is associated with clearance of the merozoite surface coat as the parasite enters the red blood cell [32]. A similar phenomenon is observed in Toxoplasma with GPI-anchored surface antigens, such as SAG1, but only if bivalent antibodies are used to loosely cross-link the proteins [5,32,37]. These observations suggest a model whereby the MJ serves as an area of continuous attachment and detachment of the two parallel membranes that somehow, as the two membranes pass through, sieves out selected membrane proteins [38]. This sieving process serves to push toward the rear or even completely remove parasite ligands like MIC2 that are transiently engaged with host receptors and thus explaining the lack of close association between the parasite and host membranes anterior of the MJ [29,39]. The absence of MIC2 in front of the MJ also argues against continued micronemal secretion after invasion has commenced although we cannot exclude the possibility that selected microneme contents (like TgAMA1) somehow continue to be secreted. It seems more likely that the presence of TgAMA1 in front of the MJ is due to the MJ specifically allowing excess TgAMA1 molecules that are not needed for MJ formation (as well as all GPI-anchored surface proteins) to pass through the sieving complex.
Our data can be integrated with the existing literature to produce the model shown in Figure 8 emphasizing the coordinated secretion and migration of TgAMA1 (green) and RON proteins (red). Micronemes are shown discharging through the rhoptry neck as has been suggested from the restricted access to the apical surface of the parasite [40] and based on the redistribution of PfAMA1 from micronemes to rhoptries in Plasmodium schizonts [20]. Secretion of RON4 follows the reorientation of the parasite to put its apical, secretory end in contact with the host membrane; this coincides with the formation of an exclusionary junction between the parasite and host cell plasma membranes. In addition to formation of the junction, one of the earliest events in invasion is the injection of rhoptry bulb proteins into the host cytoplasm where they are transiently detected in “evacuoles” [41]. After this initial stage, the parasite moves into the host cell apparently driven by migration of the MJ to the posterior end. This stage is also associated with the creation of a PV in front of the invading parasite. Once the parasite has fully entered and the PV is essentially complete, a residual focus of RON4 remains exposed on the host cell surface at the junction of PVM and host plasma membrane. This RON4 focus is lost once the PV detaches from the host plasma membrane. This is consistent with the observation of Suss-Toby et al. [3] who demonstrated a delay in the pinching off and separation of the PVM from the host plasma membrane, analogous to the constricted pit stage of receptor-mediated endocytosis. During this penultimate stage, the PVM is topologically just a very deep invagination of the host plasma membrane although the PV interior is inaccessible to macromolecules like antibodies throughout the entire process. Removal or dissociation of the RON4-containing proteinaceous cap may represent the limiting step for membrane fission and final separation of the PV.
Figure 8 Secretion and Redistribution of TgAMA1 and RON4 During Invasion
In this schematic representation, TgAMA1 (green) is secreted to the surface of parasites as they glide across the host surface. Upon reorientation, the amount of TgAMA1 on the surface is increased. A tight association with the host cell membrane is established with the secretion of RON4 (red), and rhoptry bulb constituents (grey) are detected in e-vacuoles within host cytoplasm. The MJ is established and either microneme secretion ceases or some surface proteins are able to pass through the MJ (e.g., TgAMA1, as is known for the GPI-anchored surface antigen, SAG1). RON4 migrates toward the posterior end of the parasite as a discrete ring at the constricted interface with the host cell membrane. TgAMA1 is distributed non-uniformly across the parasite surface on both sides of the MJ with a circumferential concentration at the RON4 ring. The ring containing RON4, TgAMA1, and likely RON2 and Ts4705 migrates the length of the parasite until the parasite is fully enveloped within the parasitophorous vacuole. RON4, but not TgAMA1, is then found on the host cell surface at the junction of the PV and host PM. Release of this MJ complex from the host PM allows the release of the parasite-containing PV into the host cytoplasm.
MIC2 could play a role in forming the MJ complex, but apart from its posterior cap distribution, it shows no specific association with the MJ in IIF experiments (unpublished data; [42]). It is possible that, as with TgAMA1 on wild-type parasites, the failure to specifically detect an accumulation of MIC2 at the MJ by IIF could be due to the strong, distributed signal across the parasite surface behind the MJ. As they stand, however, the above observations suggest that the TgAMA1/RON4 complex at the MJ functions independently of MIC2.
One implication of our model is a slight elaboration of previous models for sequential organelle-secretion during invasion. These were based, in part, on the analysis of rhoptry bulb proteins and held that micronemes release first, then rhoptries and lastly dense granules [1]. Our data argue for the simultaneous secretion of micronemes and at least the rhoptry necks; the rhoptry bulbs may indeed come later. Release of the neck proteins may be activated by the same calcium-responsive pathways controlling microneme secretion [43], and the two may even be physically connected if the micronemes discharge via the rhoptry necks, as previously speculated [14,40].
Our model leaves open how TgAMA1 and the MJ complex link to the cytoskeletal machinery. The fact that invasion is inhibited by Cytochalasin-D treatment clearly indicates a role for actin in the entry process [44,45], and myosin-A has also been implicated [7,46], but whether this reflects a role for these molecules in MJ movement or some other aspect of invasion (e.g., the posterior migration of the MIC2 complex) has yet to be determined. The type-1 transmembrane topology of TgAMA1 and multiple transmembrane domains of RON2 suggest one or other of these could bridge the MJ to cytoskeletal motors. Other type-1 microneme proteins, such as MIC2 in Toxoplasma and TRAP in Plasmodium, bridge to actin through aldolase [29,47,48] but neither actin nor aldolase were found specifically associated in our TgAMA1- or RON4-immunoprecipitations suggesting that if such a link is involved, it is not stable enough to be detected in these conditions. This suggests an indirect or different linkage to the cytoskeletal motor is involved in MJ vs. MIC2 migration.
The differences in linkage may also be reflected in the parasites speed when gliding vs. invading: gliding motility by Toxoplasma tachyzoites is an order of magnitude faster than invasion [49,50]. Likewise, analysis of Plasmodium sporozoite motility in vivo shows a major difference in speed during gliding across or even through hepatocytes vs. invasion into a vacuole [51]. The observation that RON4 rings fail to organize and migrate in the absence of TgAMA1 and yet normal gliding motility is unaffected by the absence of this protein [23] further indicates that invasion and gliding are very distinct processes.
RON4 but not TgAMA1 was associated with the MJ during ionophore-induced egress indicating that RON4 can function independently of TgAMA1. Consistent with this, IIF on the bradyzoite stages present in tissue cysts from a chronically infected mouse brain, revealed RON4 in the rhoptry necks and lumenal face of the cyst wall, but no TgAMA1 (unpublished data). This suggests that bradyzoites may utilize an alternative protein to coordinate with RON4 and undergo invasion. MIC7 and MIC9 are bradyzoite-specific microneme proteins that could fulfill such a role [52,53].
BLAST analysis reveals homologues in the predicted protein database of Plasmodium falciparum (http://plasmodb.org) for each of the TgAMA1-associated proteins: RON2\′s homologue is Pf14_0495 (e-value ~10−44), RON4 corresponds to Pf11_0168 (e-value ~10−9), and Ts4705 corresponds to MAL8P1.73 (e-value ~10−9). The Plasmodium homologs for RON2 and RON4 are predicted to be somewhat larger in size, but overall these proteins have similar predicted structural topology except that the homologue for Ts4705 lacks a predicted signal sequence; it does, however, have a strong hydrophobic stretch near the N-terminus of the predicted protein that may serve this function in reality. Homologues for RON2 and RON4 are also found in expressed sequence tag sequences for Neospora caninum. Consistent with this, our RON4 antisera detect distinct rings at the moving junction of invading N. caninum (unpublished data). The microneme protein TgAMA1 is highly conserved among Apicomplexa and the finding that all of the associating rhoptry neck proteins are also conserved strongly suggests that coordinated secretion from these two compartments is a common mechanism whereby the components of the moving junction are assembled at the interface between host and parasite plasma membranes.
In addition to its clear similarity to Pf14_0495, RON2 also has weak BLAST similarity (e-value ~10−8) to another published protein in P. falciparum, PfRhopH1 [54]. This latter protein is part of a complex of three high molecular weight proteins within the rhoptry necks of P. falciparum [55]. None have yet been found at the MJ but the homology to RON2 and their presence in the necks makes this possibility tantalizing. Although the other two partners of PfRhopH1, PfRhopH2 and PfRhopH3, have no detectable homologues in the Toxoplasma genome, it could be that they represent further components of the MJ and that sequence drift prevents identification through simple BLAST analysis. The different host and cellular niches of these parasites could easily produce the kind of evolutionary pressure necessary for such rapid drift. Of particular interest is the extraordinarily broad host range of Toxoplasma compared to most of the other Apicomplexa. It is tempting to speculate that specific adaptations of the MJ might be involved, e.g., in providing the parasite its own anchor in the host cell membrane and thus essentially all of the machinery necessary to invade without demanding anything of the host cell except its membrane. Which aspects of Toxoplasma's MJ constituents are unique will be interesting to explore as an explanation for this parasite's universal taste.
Materials and Methods
Parasite culture.
T. gondii tachyzoites lacking the RHΔHXGPRT gene were used for generation of lysates and the immunofluorescence analysis shown here [56,57]. The tetracycline-repressible TgAMA1-expressing parasite line, ΔAMA1/AMA1-myc, and parental line retaining the endogenous TgAMA1 gene, AMA1/AMA1-myc, were derived in the TgTA line [11,23]. Immunofluorescence localization analysis was also carried out in the Prugniaud strain [57] and did not differ from the results described here for the RHΔHXGPRT strain. Tachyzoites were grown in monolayers of HFF, exactly as previously described [18]. The ΔAMA1/AMA1-myc and parental AMA1/AMA1-myc were grown in media containing 50 μg/ml mycophenolic acid, 50 μg/ml xanthine and 20 μM chloramphenicol [46,58]. ΔAMA1/AMA1-myc parasites were depleted of TgAMA1-myc by adding fresh media containing 1.0 μg/ml (Clontech, Palo Alto, California, United States) to HFF monolayers infected at an MOI of 5, 1 hr prior to changing the media. Parasites were then grown for 24–30 hr in the presence of Atc and harvested by syringe-release. Under these conditions TgAMA1-myc was reduced to ~0.25% of constitutive TgAMA1 amounts as assessed by immunoblot analysis.
Immunological procedures.
The mAbs specific for TgAMA1 were CL22 [18] and B3.90 [19]. For immunoprecipitation of TgAMA1, these monoclonal antibodies were bound and coupled using dimethyl pimelimidate to fastflow Protein G sepharose (Amersham Biosciences, Little Chalfont, United Kingdom) [59].
Tachyzoite lysates were generated by extracting 5 × 109 extracellular parasites in 10 ml of TEN buffer (50 mM Tris [pH 8.0], 5 mM EDTA, 150 mM NaCl) containing RIPA detergents [1% NP40, 0.5% Deoxycholate, 0.01% SDS, and Complete protease inhibitor, (Roche Diagnostics, Mannheim, Germany)] on ice for 30 min and removing insoluble material by centrifugation at 3,000 × g for 20 min.
These extracts were incubated with the indicated antibodies coupled to protein G sepharose for 4 hr at 4 °C, followed by three washes (15-min each) in RIPA buffer and three washes in Tris-buffered saline (TBS; 50 mM Tris [pH 8.0], 150 mM NaCl). Bound polypeptides were eluted with 0.1 M triethylamine [pH 11.5] (using five successive elutions), and lyophilized to concentrate the eluate and remove the triethylamine. The 106 parasite equivalents were fractionated by SDS-PAGE for immunoblot analysis, or, similarly, 108 parasites for Coomassie brilliant blue staining.
Cross-linking of surface proteins.
A potassium buffer shift was used to synchronize invasion as described elsewhere [31]. Briefly, tachyzoites were harvested, washed, and resuspended in Endo Buffer (10 mM K2SO4, 2.5 mM Mg2SO4, 1 mM glucose, 5 mM Tris [pH 8.2], and 3.5 mg/ml BSA). They were then added in a volume of 10 ml to HFF monolayers in 150-mm dishes and allowed to settle at 37 °C for 15 min. The Endo buffer was removed, and 10 ml of PBS, at 37 °C, containing 50 μM DTSSP, was added to simultaneously initiate invasion and cross-linking. This incubation was carried out for 20 min at 37 °C. The HFF monolayers were washed twice with TBS to remove and quench the cross-linker. These monolayers with attached and invaded parasites were harvested, denatured by heating to 95 °C for 5 min in 1% SDS, then equilibrated in RIPA buffer, and processed as described for immunoprecipitation analysis. Under these conditions, the DTSSP cross-linker reduced invasion by ~25% as assessed by SAG1 antibody accessibility.
Protein analysis and mass spectrometry.
Individual bands from Coomassie stained SDS-PAGE gels were excised, treated with trypsin, and extracted for LC-MS/MS analysis. This analysis was performed at the Stanford University mass spectrometry facility (http://mass-spec.stanford.edu/), and the peptide fragment fingerprint data were subjected to a combined database search of the TwinScan-predicted proteins from the Toxoplasma genome database (ToxoDB; http://ToxoDB.org), and the NCBI-nr protein database by using the Bioworks MS/MS ion search program(www.thermo.com/) and Xtandem search included in the Scaffold analysis software (www.proteomesoftware.com). Peptides were ranked by an X-corr >3.2 and a D-CN >0.25, and the individual spectra examined for ion coverage and overall quality.
Nucleic acid techniques.
The generation of complete coding region cDNAs for RON2 and RON4 were as described previously, as was the use of recombinant versions of these proteins to produce antibodies in mice [27].
Constructs for the disruption of RON4 were generated using HXGPRT selection and the vector pMini-GFP.ht [60]. For the knockout construct, a PCR fragment from genomic bases −3941 to −1939, relative to the start codon, were digested with Kpn and Apa1, and ligated into pMini-GFP.ht similarly cut. Into the resulting construct, a PCR fragment of genomic bases 12,053–14,276 (again relative to the ATG start codon) digested with Not1 and SpeI was cloned. The primers used to expand the flanking regions were 5′-CGGGGTACCCCAATCAAAATCCGCAATAGCC-3′, and 5′-CGGGTGCACCAGGTGACCCGTCCATAC −3′ for the upstream flank and 5′-GGACTAGTTTGCCTTGTTTCGCCTTAC-3′, and 5′-AAGCGGCCGCTGTTTCCCTTTGAACTCTGCCAC-3′ for the downstream flank.
The resulting vector pKORON4 was linearized with NotI, and 50 μg of DNA was electroporated into RHΔHXGPRT strain parasites by standard methods, and selection of HXGPRT parasites was performed as previously described [58]. Four independent transformations were carried out. Ten days after transformation and MPA/xanthine selection, DNA was isolated from selected populations and PCR used to screen for the presence of knockout parasites. An HXGPRT-internal primer (5′-GTGGCGATTCTCATCGACTT-3′) and a primer representing the region upstream of RON4 (5′- CTTCTTCGGTTCCTCGTTAG-3′) were used for PCR detection of integration into the upstream flanking region.
Microscopy.
Analysis of tachyzoite-invasion was performed following potassium buffer shift to synchronize invasion essentially as described above. Specifically, 106 parasites in Endo buffer were added onto HFF monolayers grown on 20-mm cover-slips in 24-well plates and incubated for 15 min at 37 °C to settle and contact the monolayer. Invasion was initiated by exchanging the buffer to HBSS, supplemented with 1% FBS at 37 °C followed by incubation for either 1 min to capture partially invaded parasites, or 20 min for a predominantly fully invaded population. Tachyzoites were washed in PBS, and fixed with 3.5% formaldehyde in 150 mM phosphate buffer [pH 7.2], washed and processed for indirect immunofluorescence as described previously [18]. Alternatively, fixation was carried out using 100% methanol at −20 °C for 2 min.
Specific antibody staining was developed with appropriate Alexa488- or Alexa594-secondary antibodies (Molecular Probes, Eugene Oregon, United States). Phase and fluorescence images were captured at 100× on an Olympus BX60 and a Hamamatsu Orca100 CCD, and were pseudo-colored and merged using Image pro-plus 2.0 software (Mediacybernetics, Silver Spring, Maryland, United States). Where indicated, serial Z-stacks images were collected at 100× on a motorized Zeiss (Thornwood, New York, United States) Axiovert 200M equipped for DIC light microscopy. In these cases, fluorescence images were captured with a Hamamatsu Orca2 CCD camera (Hamamatsu, Hamamatsu City, Japan) and were deconvolved by using an iterative algorithm and calculated point spread function, pseudo-colored, and merged using Openlab 4.02 and Velocity 3.01 software (Improvision, Lexington, Massachusetts, United States).
We thank Stephen Cheng for his work on Ts4705, Sandeep Ravindran for help with developing the evacuole experiments, and David J. Ferguson for the RON4 and TgAMA1 IIF on in vivo bradyzoites. This work was supported by grants from the NIH to DLA (F32AI10552), JCB (AI21423 and AI45057), GW (AI063276), and grants to PJB from the American Cancer Society (PF-99–018–01-MBC) and the Ellison Medical Foundation (ID-NS-0162–04). Preliminary genomic and/or cDNA sequence data was accessed via http://ToxoDB.org and/or http://www.tigr.org/tdb/t_gondii/. Genomic data were provided by The Institute for Genomic Research (supported by the NIH grant AI05093), and by the Sanger Center (Wellcome Trust). Expressed sequence tag sequences were generated by Washington University (NIH grant 1R01AI045806-01A1).
Competing interests. The authors have declared that no competing interests exist.
Author contributions. DLA conceived, designed, and performed the experiments. DLA, JM, PB, and JCB analyzed the data. JM, GEW, and PB contributed reagents/materials/analysis tools. DLA and JCB wrote the paper.
Abbreviations
AAPTgAMA1-associating protein
AMA1apical membrane antigen 1
Atcanhydrotetracycline
DICdifferential interference contrast
DTSSP3,3′-Dithiobis (sulfosuccinimidylpropionate)
FITCfluorescein isothyocyante
HFFhuman foreskin fibroblasts
IIFindirect immunofluorescence
LC-MS/MSliquid chromatography-electrospray ionization-ion trap mass spectrometry
mAbmonoclonal antibody
MJmoving junction
PVMparasitophorous vacuolar membrane
PVparasitophorous vacuole
HXGPRThypoxanthine-xanthine-guanine phosphoribosyl transferase
RONrhoptry neck protein
SAG1surface antigen 1
TgACT1Toxoplasma actin
TgAMA1T. gondii version of AMA1
==== Refs
References
Carruthers VB Sibley LD 1997 Sequential protein secretion from three distinct organelles of Toxoplasma gondii accompanies invasion of human fibroblasts Eur J Cell Biol 73 114 123 9208224
Mitchell GH Thomas AW Margos G Dluzewski AR Bannister LH 2004 Apical membrane antigen 1, a major malaria vaccine candidate, mediates the close attachment of invasive merozoites to host red blood cells Infect Immun 72 154 158 14688092
Suss-Toby E Zimmerberg J Ward GE 1996 Toxoplasma invasion: The parasitophorous vacuole is formed from host cell plasma membrane and pinches off via a fission pore Proc Natl Acad Sci U S A 93 8413 8418 8710885
Mordue DG Desai N Dustin M Sibley LD 1999 Invasion by Toxoplasma gondii establishes a moving junction that selectively excludes host cell plasma membrane proteins on the basis of their membrane anchoring J Exp Med 190 1783 1792 10601353
Aikawa M Miller LH Johnson J Rabbege J 1978 Erythrocyte entry by malarial parasites. A moving junction between erythrocyte and parasite J Cell Biol 77 72 82 96121
Michel R Schupp K Raether W Bierther FW 1980 Formation of a close junction during invasion of erythrocytes by Toxoplasma gondii in vitro Int J Parasitol 10 309 313 7451022
Dobrowolski JM Carruthers VB Sibley LD 1997 Participation of myosin in gliding motility and host cell invasion by Toxoplasma gondii
Mol Microbiol 26 163 173 9383198
Morrissette NS Sibley LD 2002 Cytoskeleton of apicomplexan parasites Microbiol Mol Biol Rev 66 21 38 11875126
Sibley LD Hakansson S Carruthers VB 1998 Gliding motility: An efficient mechanism for cell penetration Curr Biol 8 R12 R14 9427622
Wetzel DM Hakansson S Hu K Roos D Sibley LD 2003 Actin filament polymerization regulates gliding motility by apicomplexan parasites Mol Biol Cell 14 396 406 12589042
Meissner M Schluter D Soldati D 2002 Role of Toxoplasma gondii myosin A in powering parasite gliding and host cell invasion Science 298 837 840 12399593
Soldati D Foth BJ Cowman AF 2004 Molecular and functional aspects of parasite invasion Trends Parasitol 20 567 574 15522666
Klotz FW Hadley TJ Aikawa M Leech J Howard RJ 1989 A 60-kDa Plasmodium falciparum protein at the moving junction formed between merozoite and erythrocyte during invasion Mol Biochem Parasitol 36 177 185 2671726
Bannister LH Hopkins JM Dluzewski AR Margos G Williams IT 2003
Plasmodium falciparum apical membrane antigen 1 (PfAMA-1) is translocated within micronemes along subpellicular microtubules during merozoite development J Cell Sci 116 3825 3834 12902400
Triglia T Healer J Caruana SR Hodder AN Anders RF 2000 Apical membrane antigen 1 plays a central role in erythrocyte invasion by Plasmodium species Mol Microbiol 38 706 718 11115107
Gaffar FR Yatsuda AP Franssen FF de Vries E 2004 Erythrocyte invasion by Babesia bovis merozoites is inhibited by polyclonal antisera directed against peptides derived from a homologue of Plasmodium falciparum apical membrane antigen 1 Infect Immun 72 2947 2955 15102807
Narum DL Thomas AW 1994 Differential localization of full-length and processed forms of PF83/AMA-1 an apical membrane antigen of Plasmodium falciparum merozoites Mol Biochem Parasitol 67 59 68 7838184
Hehl AB Lekutis C Grigg ME Bradley PJ Dubremetz JF 2000
Toxoplasma gondii homologue of plasmodium apical membrane antigen 1 is involved in invasion of host cells Infect Immun 68 7078 7086 11083833
Donahue CG Carruthers VB Gilk SD Ward GE 2000 The Toxoplasma homolog of Plasmodium apical membrane antigen-1 (AMA-1) is a microneme protein secreted in response to elevated intracellular calcium levels Mol Biochem Parasitol 111 15 30 11087913
Healer J Crawford S Ralph S McFadden G Cowman AF 2002 Independent translocation of two micronemal proteins in developing Plasmodium falciparum merozoites Infect Immun 70 5751 5758 12228305
Howell SA Withers-Martinez C Kocken CH Thomas AW Blackman MJ 2001 Proteolytic processing and primary structure of Plasmodium falciparum apical membrane antigen-1 J Biol Chem 276 31311 31320 11399764
Carruthers VB Blackman MJ 2005 A new release on life: Emerging concepts in proteolysis and parasite invasion Mol Microbiol 55 1617 1630 15752188
Mital J Meissner M Soldati D Ward GE 2005 Conditional expression of Toxoplasma gondii apical membrane antigen-1 (TgAMA1) demonstrates that TgAMA1 plays a critical role in host cell invasion Mol Biol Cell 16 4341 4349 16000372
Nair M Hinds MG Coley AM Hodder AN Foley M 2002 Structure of domain III of the blood-stage malaria vaccine candidate, Plasmodium falciparum apical membrane antigen 1 (AMA1) J Mol Biol 322 741 753 12270711
Hodder AN Crewther PE Matthew ML Reid GE Moritz RL 1996 The disulfide bond structure of Plasmodium apical membrane antigen-1 J Biol Chem 271 29446 29452 8910611
Pizarro JC Vulliez-Le Normand B Chesne-Seck ML Collins CR Withers-Martinez C 2005 Crystal structure of the malaria vaccine candidate apical membrane antigen 1 Science 308 408 411 15731407
Bradley PJ Ward C Cheng SJ Alexander DL Coller S 2005 Proteomic analysis of rhoptry organelles reveals many novel constituents for host-parasite interactions in Toxoplasma gondii
J Biol Chem E-pub ahead of print
Soldati D Lassen A Dubremetz JF Boothroyd JC 1998 Processing of Toxoplasma ROP1 protein in nascent rhoptries Mol Biochem Parasitol 96 37 48 9851605
Jewett TJ Sibley LD 2003 Aldolase forms a bridge between cell surface adhesins and the actin cytoskeleton in apicomplexan parasites Mol Cell 11 885 894 12718875
Poupel O Boleti H Axisa S Couture-Tosi E Tardieux I 2000 Toxofilin, a novel actin-binding protein from Toxoplasma gondii, sequesters actin monomers and caps actin filaments Mol Biol Cell 11 355 368 10637313
Kafsack BF Beckers C Carruthers VB 2004 Synchronous invasion of host cells by Toxoplasma gondii
Mol Biochem Parasitol 136 309 311 15478810
Dubremetz JF Rodriguez C Ferreira E 1985
Toxoplasma gondii : Redistribution of monoclonal antibodies on tachyzoites during host cell invasion Exp Parasitol 59 24 32 3881269
Dutta S Haynes JD Moch JK Barbosa A Lanar DE 2003 Invasion-inhibitory antibodies inhibit proteolytic processing of apical membrane antigen 1 of Plasmodium falciparum merozoites Proc Natl Acad Sci U S A 100 12295 12300 14526103
Black MW Arrizabalaga G Boothroyd JC 2000 Ionophore-resistant mutants of Toxoplasma gondii reveal host cell permeabilization as an early event in egress Mol Cell Biol 20 9399 9408 11094090
Lebrun M Michelin A El Hajj H Poncet J Bradley PJ 2005 The rhoptry neck protein RON4 relocalizes at the moving junction during Toxoplasma gondii invasion Cell Microbiol In press.
Hudson-Taylor DE Dolan SA Klotz FW Fujioka H Aikawa M 1995
Plasmodium falciparum protein associated with the invasion junction contains a conserved oxidoreductase domain Mol Microbiol 15 463 471 7783617
Grimwood J Smith JE 1995
Toxoplasma gondii : Redistribution of tachyzoite surface protein during host cell invasion and intracellular development Parasitol Res 81 657 661 8570580
Charron AJ Sibley LD 2004 Molecular partitioning during host cell penetration by Toxoplasma gondii
Traffic 5 855 867 15479451
Brecht S Carruthers VB Ferguson DJ Giddings OK Wang G 2001 The toxoplasma micronemal protein MIC4 is an adhesin composed of six conserved apple domains J Biol Chem 276 4119 4127 11053441
Preiser P Kaviratne M Khan S Bannister L Jarra W 2000 The apical organelles of malaria merozoites: Host cell selection, invasion, host immunity and immune evasion Microbes Infect 2 1461 1477 11099933
Hakansson S Charron AJ Sibley LD 2001 Toxoplasma evacuoles: A two-step process of secretion and fusion forms the parasitophorous vacuole Embo J 20 3132 3144 11406590
Carruthers VB Sherman GD Sibley LD 2000 The Toxoplasma adhesive protein MIC2 is proteolytically processed at multiple sites by two parasite-derived proteases J Biol Chem 275 14346 14353 10799515
Carruthers VB Moreno SN Sibley LD 1999 Ethanol and acetaldehyde elevate intracellular [Ca2+] and stimulate microneme discharge in Toxoplasma gondii
Biochem J 342 379 386 10455025
Miller LH Aikawa M Johnson JG Shiroishi T 1979 Interaction between cytochalasin B-treated malarial parasites and erythrocytes. Attachment and junction formation J Exp Med 149 172 184 105074
Dobrowolski JM Sibley LD 1996 Toxoplasma invasion of mammalian cells is powered by the actin cytoskeleton of the parasite Cell 84 933 939 8601316
Meissner M Brecht S Bujard H Soldati D 2001 Modulation of myosin A expression by a newly established tetracycline repressor-based inducible system in Toxoplasma gondii
Nucleic Acids Res 29 E115 11713335
Herm-Gotz A Weiss S Stratmann R Fujita-Becker S Ruff C 2002
Toxoplasma gondii myosin A and its light chain: A fast, single-headed, plus-end-directed motor Embo J 21 2149 2158 11980712
Buscaglia CA Coppens I Hol WG Nussenzweig V 2003 Sites of interaction between aldolase and thrombospondin-related anonymous protein in plasmodium Mol Biol Cell 14 4947 4957 14595113
Morisaki JH Heuser JE Sibley LD 1995 Invasion of Toxoplasma gondii occurs by active penetration of the host cell J Cell Sci 108 2457 2464 7673360
Hakansson S Morisaki H Heuser J Sibley LD 1999 Time-lapse video microscopy of gliding motility in Toxoplasma gondii reveals a novel, biphasic mechanism of cell locomotion Mol Biol Cell 10 3539 3547 10564254
Frevert U Engelmann S Zougbede S Stange J Ng B 2005 Intravital observation of Plasmodium berghei sporozoite infection of the liver PLoS Biol 3 e192. DOI: 10.1371/journal.pbio.0030192 15901208
Gross U Bohne W Soete M Dubremetz JF 1996 Developmental differentiation between tachyzoites and bradyzoites of Toxoplasma gondii
Parasitol Today 12 30 33 15275305
Tomavo S Fortier B Soete M Ansel C Camus D 1991 Characterization of bradyzoite-specific antigens of Toxoplasma gondii
Infect Immun 59 3750 3753 1894373
Sam-Yellowe TY Florens L Wang T Raine JD Carucci DJ 2004 Proteome analysis of rhoptry-enriched fractions isolated from Plasmodium merozoites J Proteome Res 3 995 1001 15473688
Sam-Yellowe TY Del Rio RA Fujioka H Aikawa M Yang JC 1998 Isolation of merozoite rhoptries, identification of novel rhoptry-associated proteins from Plasmodium yoelii, P. chabaudi, P. berghei, and conserved interspecies reactivity of organelles and proteins with P. falciparum rhoptry-specific antibodies Exp Parasitol 89 271 284 9676705
Donald RG Roos DS 1993 Stable molecular transformation of Toxoplasma gondii : A selectable dihydrofolate reductase-thymidylate synthase marker based on drug-resistance mutations in malaria Proc Natl Acad Sci U S A 90 11703 11707 8265612
Zenner L Foulet A Caudrelier Y Darcy F Gosselin B 1999 Infection with Toxoplasma gondii RH and Prugniaud strains in mice, rats and nude rats: Kinetics of infection in blood and tissues related to pathology in acute and chronic infection Pathol Res Pract 195 475 485 10448664
Black M Seeber F Soldati D Kim K Boothroyd JC 1995 Restriction enzyme-mediated integration elevates transformation frequency and enables co-transfection of Toxoplasma gondii
Mol Biochem Parasitol 74 55 63 8719245
Harlow E Lane D 1988 Antibodies: A laboratory manual Cold Spring Harbor (New York) Cold Spring Harbor Laboratory Press 495 p.
Arrizabalaga G Ruiz F Moreno S Boothroyd JC 2004 Ionophore-resistant mutant of Toxoplasma gondii reveals involvement of a sodium/hydrogen exchanger in calcium regulation J Cell Biol 165 653 662 15173192
|
16244709
|
PMC1262624
|
CC BY
|
2021-01-05 12:11:27
|
no
|
PLoS Pathog. 2005 Oct 21; 1(2):e17
|
utf-8
|
PLoS Pathog
| 2,005 |
10.1371/journal.ppat.0010017
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1623197310.1371/journal.pbio.0030356Research ArticleBioinformatics/Computational BiologyCell BiologyImmunologySystems BiologyBiochemistryMus (Mouse)MammalsVertebratesAnimalsEukaryotesModeling T Cell Antigen Discrimination Based on Feedback Control of Digital ERK Responses Modeling T Cell Ligand DiscriminationAltan-Bonnet Grégoire
1
Germain Ronald N [email protected]
1
1Lymphocyte Biology Section, Laboratory of Immunology, National Institute of Allergy and Infectious Disease, National Institutes of Health, Bethesda, Maryland, United States of AmericaMarrack Philippa Academic EditorNational Jewish Medical and Research Center/Howard Hughes Medical InstituteUnited States of America11 2005 25 10 2005 25 10 2005 3 11 e35616 3 2005 22 8 2005 Copyright: © 2005 Altan-Bonnet and Germain.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Experimentally Validated Model Accounts for T Cells' Discriminating Ways
T-lymphocyte activation displays a remarkable combination of speed, sensitivity, and discrimination in response to peptide–major histocompatibility complex (pMHC) ligand engagement of clonally distributed antigen receptors (T cell receptors or TCRs). Even a few foreign pMHCs on the surface of an antigen-presenting cell trigger effective signaling within seconds, whereas 1 × 105–1 × 106 self-pMHC ligands that may differ from the foreign stimulus by only a single amino acid fail to elicit this response. No existing model accounts for this nearly absolute distinction between closely related TCR ligands while also preserving the other canonical features of T-cell responses. Here we document the unexpected highly amplified and digital nature of extracellular signal-regulated kinase (ERK) activation in T cells. Based on this observation and evidence that competing positive- and negative-feedback loops contribute to TCR ligand discrimination, we constructed a new mathematical model of proximal TCR-dependent signaling. The model made clear that competition between a digital positive feedback based on ERK activity and an analog negative feedback involving SH2 domain-containing tyrosine phosphatase (SHP-1) was critical for defining a sharp ligand-discrimination threshold while preserving a rapid and sensitive response. Several nontrivial predictions of this model, including the notion that this threshold is highly sensitive to small changes in SHP-1 expression levels during cellular differentiation, were confirmed by experiment. These results combining computation and experiment reveal that ligand discrimination by T cells is controlled by the dynamics of competing feedback loops that regulate a high-gain digital amplifier, which is itself modulated during differentiation by alterations in the intracellular concentrations of key enzymes. The organization of the signaling network that we model here may be a prototypic solution to the problem of achieving ligand selectivity, low noise, and high sensitivity in biological responses.
Computational modeling provide a model which accounts for the high sensitivity, ligand selectivity, and low noise in TCR signaling. Several predictions of the model are supported by experimental data.
==== Body
Introduction
The functions of the adaptive immune system are regulated by intracellular signals arising from the interaction of clonally distributed, somatically generated receptors on T or B lymphocytes with antigens derived from invading infectious organisms [1,2]. The antigen receptors (T cell receptors or TCRs) on most conventional CD4+ and CD8+ T lymphocytes recognize short peptides extracted from pathogen proteins and displayed on cell surfaces in association with integral membrane proteins encoded by the major histocompatibility complex (peptide–MHC molecule ligands or pMHCs) [3]. Because the cellular machinery that creates pMHCs does not distinguish in most cases between pathogen proteins and host proteins, the surface of a cell that is being scanned by TCRs is typically a mosaic of self- and foreign-pMHC ligands [4]. This imposes a critical task on the T-cell recognition and intracellular signaling machinery, which is to avoid triggering functional responses to the highly abundant self-pMHCs while fostering rapid, highly sensitive, and specific responses to low densities of non-self-pMHCs on the same membrane. One major factor contributing to this discrimination by mature T cells is the elimination during thymic development of many immature cells possessing TCRs that are highly reactive with self-pMHCs [5,6]. However, this cellular selection itself depends on the capacity of the TCR to make fine distinctions between closely related pMHC structures when transducing signals that regulate cell survival and differentiation—distinctions that also must be made by mature, post-thymic T cells.
Two models have been put forward to account for the exquisite discrimination capacity of T cells. The first model is based on the idea that agonist pMHCs capable of functional T-cell activation induce a specific conformational change in the TCR complex [7–9]. The second model suggests that the signaling machinery of the T cell employs kinetic thresholding based on the lifetime of pMHC–TCR complexes to discriminate agonist pMHCs from non-agonist pMHCs [10–12]. Two experimental observations cannot be explained by the former model. First, among all the X-ray crystallographic structures of TCRs in complex with different pMHC ligands [13,14], none have displayed a change in conformation that is specific for an agonist pMHC in comparison to a non-agonist pMHC. Some investigators have proposed that a conformational change takes place in the signaling CD3 or ζ chains associated with the αβ ligand-binding subunits of the TCR [7–9,15,16], but the structures of these subunits in combination with the TCR remain to be solved, and convincing evidence for this hypothesis has yet to be reported. Second, and more significantly, the potency of pMHCs in activation of T cells endowed with a particular TCR is modulated during intrathymic differentiation [17]. Developing T cells (thymocytes) signal and respond functionally to self-pMHCs that are non-agonists for T cells in the periphery [18,19]. Hence, T cells with a given TCR can respond differently to the same set of pMHCs. This observation challenges explanations of ligand discrimination in T-cell activation on the basis of pMHC-specific conformational changes in TCRs, because responses to a given ligand differ in immature and mature cells despite the identity of the antigen-receptor structure.
In the second model, the signaling machinery of the T cell employs kinetic thresholding based on the lifetime of pMHC–TCR complexes to discriminate agonist pMHCs from non-agonist pMHCs [20,21]. The kinetic-proofreading concept, as first detailed for T cells by McKeithan [10] and then elaborated in several later variants of this original model [11,22,23], postulates that small differences in the longevity of pMHC–TCR associations are amplified into large differences in downstream signaling output by a signal-transduction pathway with many steps, each of which requires continued ligand–receptor interaction to occur. Indeed, the only biophysical parameter reported in multiple studies to correlate with the quality of T-cell activation is the lifetime of the pMHC–TCR complex. Measurements based on surface plasmon resonance with soluble TCR and pMHC suggest that the dissociation rate, rather than the association rate, of the complex is most sensitive to pMHC structure and relates best with biological potency of the ligand [24]. In one study, Kersh et al. [25] reported that a single amino acid substitution converted an agonist pMHC into a weak-agonist pMHC with a 2 × 104-fold reduction in biological potency, while only decreasing 5-fold the lifetime of the corresponding pMHC–TCR complex (from 10.8 to 2.3 s at room temperature). Thus, modest biophysical differences in the pMHC–TCR interaction appear to result in exquisite functional pMHC discrimination by T cells.
T cells not only show this capacity to distinguish among closely related ligand structures, but also have a response to antigen that is fast, extremely sensitive, and frequently digital in nature. A single pMHC is sufficient to trigger a calcium response [23] or cytotoxic activity [26] in primed T cells. Measurements of the early phosphorylation of the TCRζ chains [27] and of the calcium response of T cells [23,28] demonstrated that T-cell signaling occurs on a very short timescale (within as few as 15 s) after antigen-presenting cell (APC)–T-cell contact. Functional activation of T cells, after hours of contact with APC, is typically characterized at the individual cell level by an all-or-none response, whether measured as cytokine gene activation [29] or proliferation. These considerations make the classical kinetic-proofreading schemes for TCR signaling unsatisfactory because these models provide adequate ligand discrimination only at the expense of sensitivity or speed of response [12,30] and fail to account for a T cell's digital response to receptor engagement [23] (see Protocol S1 for a quantitative analysis of the limitations of classical kinetic-proofreading schemes in replicating the known features of TCR signaling).
The primary aim of the present study was to develop a detailed, quantitative model of early TCR signaling that accounts for these conjoint characteristics of a T cell's response to antigen. The model incorporates direct measurements of key concentrations of signaling molecules and of pMHC–TCR ligand interactions. Experiments conducted simultaneously with the model building revealed that T cells exhibit an unexpected property in their TCR-dependent signaling, namely a digital extracellular signal-regulated kinase (ERK) response that involves an extremely high level of input amplification. By constructing our model around a kinetic-proofreading scheme modified through the inclusion of two competing feedback pathways previously proposed to sharpen the discrimination threshold between closely related TCR ligands [27], we show how spurious activation of this explosive ERK amplifier by abundant non-agonist ligands can be prevented, while retaining sensitivity to low numbers of agonist ligands. The validity of our model was examined directly by cell-based experiments testing three predictions: the rapid increase of the signaling response time when the number of ligands is decreased, the hierarchy of antagonism in T-cell signaling, and the tunability of ligand discrimination during T-cell clonal expansion. The biological responses all fit well with the predictions of the model, providing strong support for the conclusion that this differential feedback scheme represents the core network of reactions that guide T-cell signaling responses to ligands and accounts for the characteristics of this key immune event.
Results
Highly Amplified, Digital ERK Responses Induced by Agonist pMHC
To develop and test a predictive model of T-cell activation, we needed quantitative measurements of the early signaling events associated with TCR engagement by pMHC. In addition, because functional data on T-cell activation show a sharp distinction between agonist ligands (that elicit T-cell responses even at low densities) and structurally related non-agonist ligands (that do not elicit such responses even at high densities), we specifically sought to identify a feature of the proximal T-cell signaling pathway that reflects this ability to discriminate, in a nearly absolute manner, between agonist and non-agonist pMHCs. We focused initially on the cascade activating the ERKs. These enzymes are members of the mitogen-activated protein kinase (MAPK) family [20,31] and are particularly attractive candidates for participating in such digital discrimination because prior studies have emphasized the importance of this pathway in both regulation of TCR signaling [27] and in functional responses [32], while other work with Xenopus oocytes has shown that the organization of this enzyme cascade can produce an ultrasensitive response associated with cell-fate decisions [33,34].
As a model system, we chose to examine the ERK-phosphorylation response of OT-1 CD8+ T cells upon activation with peptide-pulsed APCs. The transporter associated with antigen processing (TAP)–deficient lymphoma RMA-S was used as the APC because this cell line does not efficiently load self-peptides into newly synthesized major histocompatibility complex class I molecules, but does effectively present exogenously added peptide via those class I molecules that reach the cell surface [35]. Thus, peptide-pulsed RMA-S cells present a homogeneous pMHC-ligand display spanning several decades in number without an appreciable pool of co-expressed self-ligands [35]. Although there is evidence that self-recognition can synergize with contemporaneous agonist recognition in the activation of at least some CD4+ T cells [36], a previous study failed to demonstrate a substantial role for self-ligands in the activation of OT-1 T cells using RMA-S APC [37], and we have confirmed the latter findings (unpublished data). The absence of such self-pMHCs on RMA-S APC and the evidence against such ligands contributing to T-cell activation in this T cell–APC combination allowed us to simplify both our model of TCR signaling and the corresponding experimental studies. In addition, a monoclonal antibody (25D1.16) to an agonist pMHC ligand for the OT-1 T cells (SIINFEKL-Kb) was available [38], permitting direct quantitation of the absolute number of agonist ligands generated at each concentration of pulsing peptide (Figure S1).
Using RMA-S APCs with calibrated numbers of ligands to activate T cells, the ERK-phosphorylation response was measured by intracellular staining combined with flow cytometry. We found that the dual phosphorylation of ERK necessary for the activity of this kinase can be detected in an individual T cell when as few as ten agonist SIINFEKL-Kb ligands are presented on average by the APCs, with 10% of T cells showing a robust response after 3 min of T cell–APC contact (Figure 1A). These studies also revealed a previously unappreciated aspect of the T-cell ERK-signaling response: after 3 min of contact with APCs, the pattern of staining is strictly bimodal, i.e., the ERK response of T cells is essentially digital. Control experiments confirmed that the intracellular staining protocol we used was capable of detecting doubly phosphorylated, kinase-active ERK (ppERK) levels within T cells that were lower or higher than the fixed level seen among the responding cells in Figure 1A (Figure S2), indicating that the quantitatively constant nature of this signaling response is an inherent property of the T cells and not an artifact of the measurement technique. This bimodal distribution could be fitted as a sum of two discrete log-normal distributions, indicating that the individual cell ppERK response is switch-like with a nearly infinite Hill coefficient (Figure 1B). We also measured the number of ERK molecules phosphorylated during T-cell activation, using purified phosphorylated ERK for calibration. This analysis showed that the digital response of a T cell is macroscopic with the phosphorylation of 100,000 ERK proteins (Figures S2 and S3), revealing that the ERK pathway in T cells acts as a high-gain digital amplifier.
Figure 1 Quantitation of Speed, Sensitivity, and Specificity of the Digital ppERK Response in Naïve T Cells
(A) Distribution of ERK phosphorylation (measured by flow cytometry) among individual naïve OT-1 T cells after 3 min of activation by RMA-S APCs at different levels of presentation of the agonist pMHC SIINFEKL-Kb.
(B) Fit of the distribution of ppERK responses among OT-1 T cells activated with an average of 130 SIINFEKL-Kb ligands on the surface of each RMA-S APC. The fit (a sum of two log-normal distributions) is statistically adequate (χ2 = 1.72 for 128 points, and three fitting parameters).
(C) Theoretical effect of biological variation (“noise”) in ligand presentation by APCs and the responsiveness of individual T cells on the steepness of the dose response of a population of T cells. The ppERK response of an individual T cell is essentially digital (infinite Hill coefficient), but the low observed Hill coefficient (1.9) for the dose response of real T cells at a population level can be explained by taking into account the noise in ligand presentation (CV = 50%) and the possible noise in the activation threshold of the T cells (CV = 75%).
(D) Experimental ppERK dose response of naïve OT-1 T cells activated for 3 min with peptide-pulsed RMA-S cells, plotted as the percentage of responding cells. The Hill coefficient measured for this dose response is 1.9 ± 0.1 (n = 3). The threshold for activation (midpoint) is 24 ± 4 SIINFEKL-Kb on each RMA-S APC. Because the T cell's surface area is three times less than that of an RMA-S cell, as few as eight SIINFEKL-Kb ligands may be sufficient to trigger a full ppERK response if a full surface sweep of the RMA-S membrane by the T cell is not accomplished before signaling takes place.
(E) Dose response for ERK phosphorylation among naïve OT-1 T cells, after 3 min of activation by RMA-S APC pulsed with SIINFEKL peptide variants. The peptide SIINFEKL is a known agonist for OT-1 T cells, whereas EIINFEKL and SIIRFEKL are non-agonists. The percentage of responding cells is plotted as a function of the number of peptide-Kb ligands presented on the surface of each RMA-S APC.
In contrast to the single-cell results, the dose-response curve for ERK activation of a population of OT-1 T cells can be fitted with a Hill coefficient of 1.9 ± 0.1 (n = 3) and an EC50 = 24 ± 4 (n = 3) SIINFEKL-Kb per RMA-S (Figure 1C). The apparent discrepancy between the infinite Hill coefficient determined at the individual cell level and the shallower dose response measured at the population level can be understood by taking into account the distribution of ligand densities on APCs and variations in responsiveness among individual T cells (“biological noise”) (Figures 1D and S1; Protocol S2). Based on these data, the average threshold for the digital ppERK response in OT-1 T cells is 24 SIINFEKL-Kb presented per RMA-S. Because the surface area of naïve OT-1 T cells is three times less than the surface of the RMA-S used as APCs in our experiments, the absolute threshold to trigger the phosphorylation of 100,000 ERK molecules within 3 min of T cell–APC contact may be as few as eight SIINFEKL-Kb ligands.
This ERK-phosphorylation response of OT-1 T cells is also specific. When non-agonist peptide variants (such as EIINFEKL or SIIRFEKL), as defined by functional response measurements, are presented on the surface of the APC, no phosphorylation of ERK above the background could be detected, even with 1 × 105 pMHCs per APC (Figure 1E). Moreover, ERK phosphorylation in T cells after 3 min of activation also correlated with the functional specificity of activation when assayed by CD69 upregulation, interferon gamma (IFNγ) expression, or cytotoxicity (Figure S4). Because the only known differences between SIINFEKL-Kb (agonist) and the non-stimulatory EIINFEKL-Kb or SIIRFEKL-Kb ligands for the OT-1 TCR are the lifetimes of the pMHC–TCR interactions (31.5, 10.7, or 6.3 s, respectively, at room temperature [39]), our data indicate that modest differences in ligand–receptor interaction are translated into robust discrimination by the digital ppERK response of a T cell's signaling machinery.
Model of the Early Events in T-Cell Activation
The large and rapid signal amplification associated with ERK phosphorylation in T cells, coupled with a capacity of these cells to discriminate over four orders of magnitude of ligand density between two pMHCs that differ in TCR-binding lifetime by less than 5-fold, raises major questions about how this can be accommodated by traditional kinetic-proofreading schemes. Superficially, it would seem that the system should be extremely sensitive to noise, with even very poor ligands for the TCR eventually “sneaking through” [40] and causing the digital ERK response to be activated. To deal with this problem, some form of filtration or noise suppression is needed, which in signal processing would typically be mediated by a negative-feedback system [41]. Recently, the SH2 domain-containing tyrosine phosphatase (SHP-1) has been shown to play such a role in TCR signaling [27,42], and some models of TCR discrimination have incorporated this feedback to limit responses to high levels of weakly binding ligands [23,43]. However, such negative feedback alone would also act to diminish the sensitivity of the system to otherwise agonist (stimulatory) ligands. The discovery of a positive-feedback loop involving ERK-1 that protects TCRs from the inhibitory effects of SHP-1 [27] provides a possible solution to this dilemma involving sensitivity. However, because no explicit model has yet tested whether the combination of these two divergent feedback pathways with a proofreading-based scheme would quantitatively account for the key characteristics of TCR signaling, we set out to construct such a model and test its predictive capacity.
In Figure 2A, we present a simplified block diagram summarizing the key kinetic components of our model. Interaction of a pMHC with a TCR yields successive steps of phosphorylation of the TCR complex (e.g., of the associated CD3/ζ chains via activation of the Src family kinase Lck). Dissociation of the pMHC–TCR complex is assumed to permit the rapid dephosphorylation of TCR-complex components by a highly abundant active phosphatase (e.g., CD45 [44]), which is not explicitly simulated here. Kinetic proofreading of pMHC–TCR interactions is based on the assumption that phosphorylation requires pMHC–TCR contact and that dephosphorylation rapidly reverses these events upon ligand dissociation owing to the action of abundant phosphatases. The phosphorylated TCR complexes can activate two divergent pathways. Beginning shortly after TCR engagement occurs, the phosphatase SHP-1 is tyrosine phosphorylated by active Lck [45]. The resulting pSHP-1 binds stably to Lck-containing TCR complexes via interaction with the kinase's SH2 domain. This docked SHP-1 becomes enzymatically activated upon further tyrosine phosphorylation in the TCR complex, leading to dephosphorylation of the Lck, CD3/ζ, and associated ZAP-70 kinase components of the signaling complex [27,42]. This mechanism of action permits pSHP-1 to act as a spreading negative feedback by decorating unengaged Lck-containing TCR complexes and quickly deactivating the receptor when ligand engagement initiates phosphorylation events within that complex [46].
Figure 2 Computer Model of the Early Events of T-Cell Activation
(A) Sketch of model. Differential positive-/negative-feedback loops are added to a kinetic-proofreading scheme of pMHC–TCR interaction. At early times, phosphorylated TCR complexes activate SHP-1 (a tyrosine phosphatase), which provides a negative-feedback effect by dephosphorylating components within the TCR complex. Upon TCR engagement by an agonist-quality ligand, but with a time delay, the MAPK (ERK) cascade is activated and provides a positive-feedback effect by protecting the TCR complex from binding and dephosphorylation by SHP-1.
(B) Explicit model of core module of the early events of TCR signaling (see Figure S6 for an expanded view of the model).
(C) Table of the number and corresponding cytoplasmic concentrations of the signaling components involved in the model. An asterisk indicates molecules whose number and concentration are estimated.
(D) Output of the computer simulation. After 3 min of simulated time, the TCR signaling machinery produces a sensitive and specific ppERK response. There is also a sharp transition in the ppERK response depending on the quality of the pMHC ligands (as measured by the lifetime, τ, of their interaction with TCR). Four categories of ligands can be defined from the simulation. For pMHCs whose t is above 15 s, a complete ppERK response is obtained with as few as ten ligands; these are the strong agonists. For pMHCs whose τ is between 3 and 15 s, a ppERK response is obtained when sufficient numbers of ligands are present; these are the weak agonists. pMHCs whose τ is below 3 s fail to trigger a ppERK response; these are non-agonists. Finally, because different combinations of feedback control are triggered by each category of ligands, ligands whose τ is below 1 s do not trigger negative feedback efficiently. These may constitute the majority of self-ligands, preventing self-recognition from depressing responses to full agonists [69].
At slightly later times after TCR engagement, TCR complexes can proceed to full phosphorylation and trigger a kinase cascade activating ERK. Active ERK, in turn, acts as a positive feedback by serine phosphorylation of the Lck in TCR complexes, a biochemical modification that prevents the kinase from binding to pSHP-1 [27]. Hence, our model network for TCR signaling can be summarized as a kinetic proofreading of the pMHC–TCR interactions, triggering the MAPK cascade as a high-gain digital amplifier, with a rapid-onset analog SHP-1-mediated negative feedback and a slower digital ERK-1-dependent positive feedback modulating the triggering threshold. The evidence for signal spreading that we have reported previously [46], in concert with the digital nature of the ERK response, enables this counter-intuitive arrangement of an early arising negative feedback and delayed positive feedback to support effective signaling.
For the purpose of comparing computer simulations with experiments, we implemented explicit chemical reactions using parameters derived as much as possible from direct measurements rather than fitting, although the latter was necessary for many of the enzymatic rates that have not been measured in a cellular context (Figures 2B, S3, and S5; Protocol S3). The only quantitative parameter distinguishing different pMHC ligands in our model is the lifetime of their interaction with TCR. The expression levels of signaling molecules were determined by quantitative immunoblotting or flow cytometry and their concentration calculated based on the measured cytoplasmic volume of naïve T cells (15 fl) (Figures 2C and S5). These measurements underscore the fact that T cells contain large concentrations of signaling molecules (>3 μMol), which is consistent with the high speed of response and also limits the impact of stochastic behavior on the chemical reactions. We therefore assumed that diffusion kinetics were not limiting and used a stirred-cell model in our simulations.
The macroscopic clustering and spatial reorganization of proteins in the immunological synapse is not required for the early rapid signals we are assessing here [20,47], and hence, the correlation of the formation of this organized multiprotein structure with effective T-cell activation [47] is not inconsistent with our assumption. JDesigner software [48] was used to define the biochemical network for TCR signaling (see Figure S6 and Protocol S4 for a complete description of the model and Protocol S3 for a complete description of the kinetic parameters used in our experiments and their origin). The computer modeling of this network involved solving a set of deterministic differential equations with a Rosenbrock formula of order 2, implemented using Matlab (see Protocol S4).
Solving our computer model for different quantities and qualities of ligands (i.e., different lifetimes of the pMHC–TCR complex) shows how the competition between positive and negative feedbacks defines a digital threshold of T-cell activation in terms of the dynamics of this ligand–receptor interaction. Figure 2D presents the simulated ppERK response after 3 min of exposure to different numbers of ligands of different receptor-binding lifetimes. This response is nearly digital with a sharp threshold at a pMHC–TCR lifetime of 3 s, comparable to the experimentally reported threshold in pMHC–TCR complex lifetime for agonist activity, extrapolated to 37 °C (see Protocol S3). Hence our model shows almost absolute discrimination with respect to the quality of pMHC–TCR ligand interaction, while also showing both fast kinetics and sensitivity to a few agonist ligands.
Testing Three Predictions of the Differential Feedback Model of T-Cell Signaling Control
Lengthening of the ppERK response time at low ligand densities
In our model, the MAPK cascade of concatenated kinase phosphorylations amplifies sparse input signals (the output of the kinetic proofreading of pMHC–TCR interaction) to yield a robust ppERK response (Figures S7 and S8): indeed ten ligands were shown to drive the phosphorylation of 100,000 ERK molecules in T cells (see Figure 1A). One hallmark of such a kinase cascade is the digital response of ERK phosphorylation (i.e., large Hill coefficient at the cellular level [see Figure S8C]), which we have confirmed experimentally (see Figure 1B). A key predicted feature of such a response scheme is the nonlinear lengthening of the response time at low ligand densities (see Figure S8B and S8D).
To examine whether this behavior seen in the simulations (Figure 3A) was characteristic of the biological system, we systematically measured the kinetics of ERK phosphorylation in T cells exposed to different numbers of agonist ligands (Figure 3B). In qualitative agreement with the model (see Figure 3A; Protocol S5), the time delay before the digital ERK response increased dramatically as the number of ligands was decreased towards threshold levels (Figure 3C). There was also a second, less-pronounced, slowing-down of the kinetics of ERK phosphorylation at high levels of presentation, an effect that the model indicates arises from a rapid and massive activation of the negative feedback that limits the efficiency of triggering of the MAPK cascade at high (>1 × 104) agonist display. In the case of non-agonist ligands, a similar temporal imbalance in favor of negative feedback, even at modest ligand levels, abrogates the activation of the MAPK cascade and allows the suppressive regime to dominate at all pMHC densities, preventing effective responses. Although the simulation results and experimental data fit well qualitatively, there was a systematic discrepancy of 13 s in response time that could not be resolved by parameter adjustment without losing other predictions of our simulation. More detailed modeling (in particular by taking into account membrane protein pre-clustering) or more accurate measurements of the relevant kinetic and expression parameters at physiologic temperature may help eliminate this modest kinetic discrepancy.
Figure 3 Experimental Test of Two Predictions of the Computer Simulation of the Early Events in TCR Signaling
(A–C) Characteristic response time for ERK phosphorylation. The characteristic time of the ERK-phosphorylation response was derived by computer simulations of TCR signaling for increasing numbers of agonist ligands (whose lifetime of interaction with the TCR is set at 18 s) (A). This timescale diverges in a nonlinear fashion when the number of agonist ligands is decreased. We then systematically measured the kinetics of the ppERK response of naïve OT-1 T cells upon activation with RMA-S APCs presenting different numbers of SIINFEKL-Kb (B) and derived the characteristic time of response using a generic sigmoidal fit. The divergence of this timescale as the number of agonist ligands is decreased (C) is characteristic of kinase cascades acting as digital filters (A).
(D–F) Comparison of antagonism in T-cell activation in computer simulations and experiments.
(D) Computer simulation of antagonism. We simulated the ppERK response of T cells upon activation with increasing numbers of agonist ligands (whose interaction with the TCR has a lifetime of 18 s) in the presence of 30,000 non-agonist ligands (the two putative antagonist ligands being tested have TCR-interaction lifetimes of 1.7 s [weak antagonist] and 3 s [strong antagonist], respectively). The presence of a large number of sub-threshold ligands inhibits the agonist-induced ppERK response of T cells. The inhibition is calculated as the ratio of the ppERK response in T cells activated with agonist and antagonist together as compared to the ppERK response seen using the agonist alone. This hierarchy of antagonism in early T-cell responses is consistent with the graded activation of SHP-1-mediated negative feedback associated with signaling by sub-threshold ligands.
(E) Experimental test of antagonism. Naïve OT-1 T cells were activated with RMA-S APCs pulsed with an increasing amount of agonist SIINFEKL peptide and an excess of EIINFEKL or SIIRFEKL peptides.
(F) Experimental ppERK response of OT-1 T cells upon activation with RMA-S APCs presenting 25 agonist SIINFEKL-Kb ligands with or without 30,000 antagonists (SIIRFEKL-Kb [weak antagonist] or EIINFEKL-Kb [strong antagonist]).
Hierarchy of antagonism in T-cell signaling
Our computer simulation also enabled us to probe the role of SHP-1 negative feedback in setting the threshold of ligand discrimination. Given the “explosive” responsiveness of the MAPK cascade, T cells must activate a negative feedback that is tuned to the strength and quantity of ligands, to blunt spurious activation with large quantities of low-affinity ligands while allowing sensitive responses towards small quantities of more strongly binding complexes. Negative feedback mediated by SHP-1 has previously been shown to be responsible for TCR antagonism [27], a phenomenon in which simultaneous exposure of a T cell to a large quantity of sub-threshold ligands and a small, otherwise stimulatory, number of agonist ligands results in blunting of the expected response [49,50]. The present quantitative model predicts that TCR antagonism has a counter-intuitive characteristic: the more closely the lifetime of a TCR–non-agonist complex approaches the threshold needed for full signaling, the more strongly this ligand will antagonize T-cell activation by agonists. That is, better binders will actually be better inhibitors until a bifurcation point is reached and they become overtly stimulatory ligands themselves.
This hierarchy of antagonism can be seen in the results of Dittel et al. [46], who reported greater TCRζ phosphorylation by more potent antagonists. This effect is shown in Figure 3D, in which we present a computer simulation of the ppERK response of our TCR signaling model for increasing numbers of agonist pMHCs (whose TCR-binding lifetime is 18 s) in conjunction with 3 × 104 antagonist pMHCs (whose TCR-binding lifetimes are either 3 s or 1.7 s). We examined this prediction experimentally by measuring the OT-1 ERK response to RMA-S cells bearing an increasing amount of agonist SIINFEKL-Kb with or without a large number (3 × 104) of non-agonist ligands on the same cell membrane (Figure 3E and 3F). EIINFEKL-Kb antagonized more effectively than SIIRFEKL-Kb, consistent with the fact that the former ligand forms a longer-lived complex with OT-1 TCR than does the latter (10.7 s versus 6.3 s at room temperature [39]), in accord with the expectations of the model. The slight decrease in the ppERK level among cells responding in the presence of antagonist is also seen with cells stimulated with very low densities of agonist alone and appears to arise from the 3-min assay point capturing these cells prior to their achieving maximum ppERK levels. This slower rise to the maximum at low effective ligand densities is predicted by our model (see Figure 3A).
Flexibility in ligand discrimination for T cells undergoing differentiation
A third general prediction of our computer model is that the precise positioning of the kinetic threshold between agonist and non-agonist ligands for a particular TCR is set by the dynamics of the competition between positive- and negative-feedback loops, which in turn is highly sensitive to modest changes in the intracellular concentration of key components such as SHP-1. Thus, we predicted that T cells could alter their discrimination threshold by small changes in the concentration of these molecules during differentiation.
Consistent with this expectation, analysis of the TCR signaling response of activated OT-1 T cells revealed that these cells transiently demonstrated ERK responses to EIINFEKL-Kb 5 or 6 d after the initial activation, whereas naïve OT-1 (OT-1)naïve and activated OT-1 T cells rested for 11 d (OT-1)day 11 were strictly unresponsive towards this ligand (Figures S9 and S10). An assessment of protein concentrations in these cells revealed a substantial (>10-fold) decrease in the concentration of SHP-1 relative to other measured key signaling molecules in (OT-1)day-5 T cells as compared to (OT-1)naive or (OT-1)day-11 T cells (see Figure S9). The model suggested that the acquisition of responsiveness to a ligand showing poor TCR-binding characteristics, as we saw for (OT-1)day-5 T cells, could be accounted for by this relative diminution in SHP-1 levels. Simulations also predicted that the selective decrease in SHP-1 should yield a peculiar dose response to the low-affinity ligand, with a measurable ppERK response at moderate ligand concentrations, followed by a rapid loss of this signaling response as ligand density increases. This is because the decreased SHP-1 levels slow down the functioning of the negative-feedback pathway in response to moderate levels of weak ligand, allowing activation of ERK; at high levels of presentation, the pace of the SHP-1-mediated negative feedback is accelerated and overrides the delayed activation of ERK. The same change in SHP-1 level was predicted to have no detectable effect on the dose response to a strong agonist.
To test these predictions experimentally, agonist-activated OT-1 T cells were infected with a retrovirus encoding EGFP only or encoding both SHP-1 and EGFP as a bicistronic mRNA. Five days after activation, the control infected OT-1 T cells showed the expected selective decrease in SHP-1 concentration, though the decrease was of smaller magnitude than typically seen with uninfected cells and the corresponding gain in reactivity to weak ligands was less pronounced. Infection with the SHP-1-encoding virus restored the SHP-1 concentration to a level similar to that found in naïve cells (Figure 4A). As the model predicted (Figure 4B), (OT-1)day-5 T cells infected with EGFP-expressing retrovirus responded to EIINFEKL-Kb, but only at intermediate ligand densities, while cells infected with SHP-1/IRES/EGFP-expressing mouse stem-cell virus (MSCV) selectively lost the EIINFEKL-Kb response without any associated loss of sensitivity to activation by the full agonist SIINFEKL-Kb (Figure 4C). These data demonstrate that ligand discrimination is not “hard-wired” into the affinity or structural match between a particular TCR and its ligand, but is modulated by differentiation-related changes in the stoichiometry of components of the signaling network downstream of the receptor.
Figure 4 Experimental Verification of the Predicted Role of Small SHP-1 Concentration Changes in Altering Ligand Discrimination by OT-1 T Cells
(A) Concentrations of signaling molecules in OT-1 T cells 5 d after activation and infection with MSCV retrovirus in vitro (day 5). These concentrations are normalized using the corresponding concentrations in the unstimulated naïve state (day 0).
(B) Computer simulation of the responsiveness of T cells at day 5 after activation with the SHP-1 level set to that seen in the naïve state. The agonist pMHC is set to bind TCR with a characteristic time of 18 s and the non-agonist pMHC is set to bind TCR with a characteristic time of 3 s.
(C) Elimination of the response of day-5 activated cells to EIINFEKL/Kb by expression of additional SHP-1. OT-1 T cells were infected with MSCV retrovirus expressing EGFP (control) or SHP-1/IRES/EGFP, and the ppERK response of infected OT-1 T cells to peptide-pulsed RMA-S was tested on day 5 after initial activation.
Discussion
An extensive literature on the intracellular signals triggered by pMHC-ligand engagement of the TCR suggests that the response is rapid, sensitive, and highly discriminatory. In this study, we have documented another key feature, namely the digital, highly amplified ERK response that occurs at short timescales (<3 min) but correlates with functional responses at >1 h post-TCR engagement. This finding raised a fundamental issue: how can T cells trigger such an “explosive” response while maintaining the specificity of ligand discrimination? In an attempt to construct a model that accounted simultaneously for all four key characteristics of TCR signaling in response to ligand engagement, we combined two recently reported opposing feedback modules [20,27] with a core scheme based on kinetic proofreading [10,11,30,43]. Using realistic kinetic parameter sets for computer simulation of the signaling cascade downstream of TCR engagement, our model yielded an output that had the striking characteristic of a sharp transition in ligand agonist functionality at TCR-binding lifetimes corresponding to those measured in several different T-cell systems [21]. We showed that this transition is also consistent with the very large (1 × >104) shift in potency of pMHC ligands that differ by only a few fold in their binding lifetimes. Our modeling suggests that the sharp threshold for pMHC-receptor lifetimes yielding agonist responses originates from the distinct kinetic characteristics of the phosphatase-mediated negative feedback that suppresses signaling by weak ligands and the ERK-mediated positive feedback that is induced effectively only by more avid ligands of the TCR.
We built upon the observations of Stefanova et al. [27] in constructing a model in which SHP-1 mediated inhibition begins to function quickly upon TCR engagement, but scales in an analog way with input. In contrast, the ERK response was modeled as delayed but (as newly documented here) digital in nature. This combination allows TCR activity induced by a large number of weak ligands to be constantly repressed by a proportional negative feedback that has enough time to quench upstream signals before they reach the limit necessary to trigger the ERK response. Both modeling and experiment confirm that the ERK response is increasingly delayed in onset as the duration of pMHC–TCR binding decreases. In contrast, more strongly binding ligands, though also inducing an initial SHP-1 inhibitory response, override the limited nature of this negative feedback early after ligand engagement by quickly triggering the highly amplified ERK digital response. The magnitude of this ERK activation then prevents inhibition of those TCR not yet inactivated by the gradually rising pSHP-1 levels, permitting effective downstream signaling through diverse pathways that impinge on genes involved in T-cell differentiation. The latter expectation of a transient recruitment of pSHP-1 to agonist-engaged TCRs and the generation of an abortive proximal tyrosine-phosphorylation response in T cells exposed to an agonist when the ERK cascade is inhibited have both been observed in biochemical studies [27]. Overall, these observations provide new insight into how control circuits can be organized to suppress noise generated by large numbers of ligands while promoting highly sensitive responses to a few optimal stimuli in the same cellular context.
Our simulations enabled us to make several predictions that were verified by experiment. Most relevant to our understanding of how T cells set the threshold for discriminating between foreign and self-ligands to promote effective responses without fostering autoimmunity, we predicted that modest changes in intracellular enzyme levels would “tune” this agonist threshold during differentiation [51–53]. This prediction was confirmed in studies showing that the decreased amount of SHP-1 in T cells a few days after activation of naïve T cells accounts for a gain in response to a pMHC ligand that is incapable of stimulating naïve or resting primed cells expressing the same TCR. This sensitivity of the response-threshold position to modest alterations in the intracellular concentrations of key components of the network, particularly SHP-1, was a somewhat surprising result. Stochastic noise in the production and degradation of signaling components might be expected to produce fluctuations of a similar magnitude in key molecules [54] and hence to jeopardize accurate self-/non-self-discrimination by T cells in the periphery after the threshold is set during positive and negative selective events in the thymus.
One possible explanation for how this is avoided is that naïve T cells may have a very stable metabolism that enables them to preserve the phenotype selected for in the thymus prior to overt activation by foreign ligand. Alternatively, others have proposed that T cells can respond to tonic exposure to self-ligands by abrogating self-responsiveness while maintaining reactivity to pathogen-derived ligand [51–53,55]. Perhaps this “self-tuning” involves dynamic adjustment of the competition between positive and negative feedbacks. A third possibility is that such fluctuations do result in an occasional T cell producing potential activation signals upon self-recognition; however, in the non-inflamed state, this would lead to tolerance through deletion or anergy [56]. The danger would be if this occurred during an inflammatory response, but indeed it is just such situations that may be inciting events for autoimmunity [57].
In this same regard, the acquisition among activated T cells of overt signaling responses to variant pMHC ligands that do not evoke such responses among naïve or resting primed T cells with the same TCR is an intriguing finding whose physiological relevance is only evident in one circumstance. Hogquist et al. originally identified EIINFEKL as a peptide driving positive selection of OT-1 T cells under organ-culture conditions in which the usual display of self-peptides is limited [18]. EIINFEKL was also the strongest antagonist of OT-1 T-cell activation by SIINFEKL presented in the context of H-2Kb [18]. Hence, EIINFEKL-Kb was a ligand known to induce some positive signaling in the OT-1 thymocytes and antagonistic negative signaling in peripheral OT-1 lymphocytes. Our model and experiments enable us to hypothesize how this divergent signaling capacity of EIINFEKL-Kb may correlate with up/down expression of components of the TCR signaling machinery and, specifically, SHP-1 [42]. Thus, actively keeping SHP-1 levels low during early T-cell differentiation could allow self-ligands to have weak-agonist function and drive the positive selection of the T-cell repertoire, while increased SHP-1 levels would eliminate this response capacity among the mature T cells that populate the periphery [19]. The “bell-shaped” dose response induced by EIINFEKL-Kb using (OT-1)day-5 T cells has been observed in other biological systems [58,59]: our model suggests that such a nonmonotonic dose response is in fact a reflection of the activation of excess negative feedback at a high dose of weak ligands.
Why more mature T cells that have been recently activated should alter SHP-1 levels so as to regain sensitivity to stimulation by weak ligands is not yet clear, but one possibility is that activated cells use this reprogramming of the signaling threshold to take advantage of abundant self-ligands to promote further differentiation once their initial activation has been “validated” by foreign-agonist recognition. Our biological studies and simulations were both conducted in the absence of such potentially active self-pMHCs. However, a very recent study indicates that this synergy can occur in a narrow time window after previous agonist-mediated T-cell activation [36], consistent with the gain in sensitivity to fast off-rate pMHCs that has been shown here to be due to decreased SHP-1 levels in this time frame.
What are some of the potential limitations of the current model? While it has proved very successful in simulating aspects of T-cell biology that can be verified experimentally and even has correctly predicted some behaviors not previously recognized, we do not know the extent to which the simplifications we have introduced to keep the model tractable have compromised its ability to reflect T-cell physiology. First, this model lacks spatial constraints and treats the T cell as a well-stirred vessel for the first 3 min of TCR signaling. We believe this is justified, based on our quantitative analysis of naïve T cells and their contents, which reemphasized the small cytoplasmic volume of these cells and the resulting high concentrations of signaling components. For this reason, most enzymatic reactions involved in T-cell signaling are not diffusion-limited. Moreover, the spatial reorganization of membrane signaling proteins during T-cell activation that results in a mature immunological synapse [60] takes place over a substantially longer timescale than the one considered in our model [61], and initial signaling occurs prior to the large-scale protein clustering involved in the formation of this synaptic structure. This does not mean that local inhomogeneities in protein distribution in the membrane (e.g., “rafts”), or involving scaffolded protein complexes in the cytoplasm, do not influence signaling behavior.
More elaborate modeling tools that preserve spatial information ([62]; M. Meier-Schellersheim et al., unpublished data) will be needed to expand analyses of T-cell signaling. This may be particularly relevant in understanding how self-ligands can synergize with agonist pMHCs in promoting T-cell activation [36] and in better modeling the role of signal spreading among engaged and nonengaged TCR in the action of pSHP-1 and ppERK. Second, we have omitted explicit specification of a number of molecules that are well documented in the literature to affect T-cell signaling responses, such as CD45, Csk, and several adapter proteins [63,64]. However, the influence of these components on signaling was implicitly incorporated in some of the kinetic parameters (e.g., tonic dephosphorylation), and we feel this is justified by the absence of evidence that any of these components has first-order sensitivity to the quality of the pMHC–TCR ligand interaction. Third, we have introduced modifications to the kinetic parameters of pMHC–TCR interaction measured at room temperature to match the model's output to biological experiments conducted at 37 °C. Whether our approximations in this regard are accurate are not yet clear, because evidence for both linear and nonlinear effects of temperature on pMHC interactions with TCRs have been reported [9,39,65]. Finally, we have considered here only the signaling involved under conditions in which the CD8 coreceptor plays a key role in the response. This is not an absolute necessity for all TCR-mediated activation, but it is a common feature of many physiological T-cell responses including that of the OT-1 cells we used for the biological component of the present study.
Although the primary aim of this work has been to better understand how TCR signaling is regulated and contributes to the proper performance of T cells in the immune system, the results we have obtained showing how simple feedback loops operate to suppress biological noise, amplify responses, and allow flexibility of function are likely to be relevant to many other biological systems. The roles of negative and positive feedback are well documented in gene regulatory networks [66] and, in particular, in developmental systems, where they can impose irreversible state changes on the system, providing unidirectionality to differentiation events [67]. It remains to be seen whether the specific features of the opposing feedback pathways modeled here, (rapidly initiating analog negative feedback versus delayed, digital positive feedback) are critical in other biological systems. More generally, the value we document here for quantitative modeling rather than just qualitative cartoon depiction of signaling circuits, and the importance of documenting physiologically relevant signaling dynamics in simulation outputs, indicate that methods and tools for constructing models, conducting simulations, and measuring values will be increasingly critical aspects of experimental biology.
Materials and Methods
Computer modeling of the early events in T-cell activation
The network of the biochemical reactions taking place upon pMHC interaction with a TCR was created using JDesigner (http://www.cds.caltech.edu/˜hsauro/JDesigner.htm) and converted into a Matlab file. The resulting set of deterministic differential equations was solved with a Rosenbrock formula of order 2 implemented. Complete descriptions of the models (for the simple kinetic-proofreading schemes and for the full TCR signaling cascade) are available in Protocol S1 and Figure S6.
Cells, peptides, proteins, and antibodies
Splenocytes and lymphocytes were isolated from H-2b OT-1 TCR transgenic mice (Tacline 175, Taconic [18]) on a Rag-2−/− background [68] and used directly as responding naïve OT-1 T cells. RMA-S TAP-deficient T cell lymphoma cells [35] were used as APCs. The agonist ovalbumin peptide SIINFEKL and its variants SAINFEKL, EIINFEKL, SIIRFEKL (all >95% pure) were obtained through the National Institute of Allergy and Infectious Disease Research Technologies Branch. E10 antibody against ppERK was purchased from Cell Signaling Technology (Beverly, Massachusetts, United States); MR9–4(PE) against Vβ5.2, 53–6.7(PE) against CD8α, H1-2F3(PE) against CD69, and XMG1.2(APC) against IFNγ were from BD Biosciences Pharmingen (San Diego, California, United States); K-23 against ERK1, C-19 against SHP-1, and C-14 against ERK2 were from Santa Cruz Biotechnology (Santa Cruz, California, United States); 3A5 against Lck, purified MEK, purified ppERK, and purified ZAP70 proteins were from Upstate Technologies (http://www.upstate.com). SHP-1 was purified from the lysates of Escherichia coli cells transformed with a SHP-1-GST-encoding plasmid (a gift from D. Nandan, University of British Columbia, Vancouver, Canada) and calibrated against pure GST. The 25D1.16 monoclonal antibody specific for SIINFEKL-Kb [38] was used as a hybridoma supernatant.
Quantitation of peptide presentation
RMA-S cells were pulsed with serial dilutions of stimulating peptides for 1–2 h at 37 °C in serum-supplemented RPMI-1640 medium. Cells were washed and then stained with phycoerythrin-coupled anti-H-2Kb AF6–88.5 antibody, whose fluorescence was calibrated with Quantibrite beads (BD Biosciences Pharmingen) and anti-IgG beads (Bangs Laboratories, http://www.bangslabs.com). We also used the combination of anti-SIINFEKL-Kb 25D1.16 antibody and phycoerythrin-conjugated anti-mouse antibodies (Jackson Immunoresearch, West Grove, PA), calibrated with the anti-H-2Kb staining at high levels of presentation, to achieve better resolution at low concentration of pulsing peptides.
Flow-cytometric measurement of intracellular signaling responses
T cells (5 × 105) were mixed with peptide-pulsed RMA-S (2 × 106), spun at 370 g for 5 s, and placed at 37 °C for various amount of time. T cell–APC conjugates were then separated with ice-cold PBS/2.5 mM EDTA, and fixed with 4% paraformaldehyde for 30 min on ice. Cells were then permeabilized with 90% methanol for 30 min on ice, washed twice with PBS/4% fetal bovine serum (FACS buffer), incubated with 1 μg/ml E10 in FACS buffer, and finally stained with 1 μg/ml of phycoerythrin-labeled anti-mouse immunoglobulin. Staining was immediately measured by flow cytometry (FACSCalibur, BD Biosciences Pharmingen), after gating for small cells based on forward scatter. Calculation of the percentage of ppERK+ cells was performed with FlowJo (Treestar, http://www.treestar.com).
Determination of the characteristic ppERK response time
We defined the characteristic ppERK response time of a T cell as the time yielding 50% of the maximal response for a given level of presentation of SIINFEKL-Kb.
Flow-cytometric analysis of functional T-cell activation
T cells (5 × 105) were mixed with peptide-pulsed RMA-S (2 × 106), spun at 10,000 rpm for 5 s in an Eppendorf microfuge (Hamburg, Germany), and placed at 37 °C for 3 h. T cell–APC complexes were then dissociated with ice-cold PBS/2.5 mM EDTA, and stained for CD69, then analyzed by FACS. The fraction of live APC was also determined to yield a measure of the cytotoxic activity of the T cells. For IFNγ expression, T-cell activation was performed as before, with the addition of 2 μm monensin. Cells were fixed, permeabilized with FACS buffer containing 0.1% saponin, stained for IFNγ, and analyzed by flow cytometry.
Quantitative measurements of intracellular protein levels
Intracellular protein levels were assessed by the lysis of 100,000 naïve OT-1 T cells in 1% NP-40 (Pierce Biotechnology, Rockford, Illinois, United States) with complete protease inhibitor (Boehringer Ingelheim, Ingelheim, Germany). The proteins in these lysates were separated by SDS-PAGE using an 8%–16% gel in parallel with serial dilutions of protein standards for immunoblotting calibration and then transferred to nitrocellulose membranes. After development of the blot with the relevant antibodies, a Kodak Image Station 440 was used to quantitate the bands and yield the protein content per cell. To quantitate the expression of receptors on the surface of T cells, we used standard beads (Quantibrite, BD Biosciences Pharmingen) to calibrate the antibody staining of Vβ5.2 and CD8α.
Fit of the distribution of ppERK
The pattern of ppERK in T cells as measured by flow cytometry was fitted with the sum of two log-normal distributions, representing nonactivated and fully activated T cells. Free parameters were the modes of staining of ppERK− cells and ppERK+ cells, and the percentage of ppERK+ cells. Coefficients of variation (CVs) were set at 55%, corresponding to the measured CV for the distribution of ERKs in naïve T cells.
Overexpression of SHP-1 by retroviral infection of activated OT-1 T cells
To examine the effect of altering SHP-1 levels on signaling in response to various pMHC ligands of the TCR, we used the retroviral vector MSCV expressing either SHP-1/IRES/EGFP or just EGFP [27]. Ecotropic Phoenix packaging cells (a kind gift of G. Nolan, Stanford University, Palo Alto, California, United States) were transfected with DNA corresponding to these two viral constructs, and supernatants were collected for spin-infection of OT-1 T cells undergoing proliferation after activation with OVA-pulsed splenocytes from B6 mice, followed by culture in 7.5% T-STIM (BD Biosciences Pharmingen) [27]. The ppERK response of infected cells to peptide-pulsed RMA-S was measured after gating on EGFP+ cells. Estimate of intracellular levels of expression of signaling components was performed by immunoblotting using lysates corresponding to 100,000 T cells. Correction for the low percentage of infection (typically 15%) was made to estimate the overexpression of SHP-1 in MSCV(SHP-1/IRES/EGFP)–infected T cells compared to MSCV(EGFP)–infected T cells.
Supporting Information
Figure S1 Calibration of the Presentation of SIINFEKL-Kb on the Surface of RMA-S APC
(A) Distribution of presentation of SIINFEKL-Kb on the surface of RMA-S APCs, measured by staining with 25D1.16 antibody, for different concentrations of the agonist peptide SIINFEKL.
(B) Quantitation of agonist pMHC presentation on the surface of RMA-S APCs. Calibration of the fluorescence staining (mean fluorescence intensity) by the 25D1.16 antibody was performed with Quantibrite beads (BD Biosciences Pharmingen).
(C) Fit of the distribution of 25D1.16 antibody fluorescence staining on the surface of RMA-S APC pulsed with 10 nM SIINFEKL peptide. The distribution is log normal with CV = 51% (χ2 = 3.2 for 100 points, and three fitting parameters).
(D) CV of the distributions presented in (A) for different concentrations of SIINFEKL.
(213 KB PDF).
Click here for additional data file.
Figure S2 Estimate of the Absolute Number of ERK Molecules Involved in the pMHC Response of Naïve OT-1 T Cells
To estimate the intrinsic ERK phosphorylation in OT-1 T cells independently of pMHC stimulation and the maximal possible response, we compared the FACS staining with E10 anti-ppERK or anti-mouse IgG1 isotype control for OT-1 cells alone, OT-1 cells activated with unpulsed RMA-S, OT-1 cells activated with SIINFEKL-pulsed RMA-S cells, and phorbol myristic acetate–activated OT-1 cells (as a control for maximal response). Typically, the full ppERK response of naïve T cells to APC stimulation involves 50% of the total pool of ERK. The ppERK response that is independent of pMHC stimulation is negligible within our experimental resolution.
(221 KB PDF).
Click here for additional data file.
Figure S3 Quantitation of Surface and Cytoplasmic Signaling Proteins
(1A and 1B) The numbers of TCR and CD8 molecules on OT-1 T cells were determined using calibrated flow-cytometric measurements.
(2A and 2B) The number of intracellular signaling proteins per cell was determined by immunoblotting using purified proteins as calibration standards. An example of the method as applied to ERK2 is presented.
(405 KB PDF).
Click here for additional data file.
Figure S4 Comparison of ERK Phosphorylation, IFNγ Production, and the Cytotoxic Activity of Rested OT-1 T Cells Following Activation by Peptide-Pulsed RMA-S APCs
(A) ERK phosphorylation, (B) IFNγ production, and (C) cytotoxic activity. Biological responses of OT-1 T cells were measured as described in Materials and Methods for various concentrations of SIINFEKL, EIINFEKL, and SIIRFEKL peptides used to pulse RMA-S APCs. For these experiments, we used OT-1 T cells that had been previously activated 8 d prior to the assay, expanded in medium supplemented with 7.5% T-stim and 10% FCS, and rested for 2 d before restimulation.
(165 KB PDF).
Click here for additional data file.
Figure S5 Measurement of the Cytoplasmic Volume of Naïve OT-1 T Cells
Confocal microscopy was used to determine the cytoplasmic volume of naïve OT-1 T cells. The diameter of naïve T cells is 5.6 ± 0.5 μm (n = 10)—hence a cellular volume of 90 fl. After measuring the volume of the nucleus, the cytoplasmic volume of T cells can be estimated to be 15 ± 3 fl. The cytoplasmic concentrations of ZAP70, MEK1, and ERK2 are 120 μm, 20 μm, and 10 μm, respectively.
(1.4 MB PDF).
Click here for additional data file.
Figure S6 Explicit Description of the Computer Model of the early TCR Signaling Events
Our model was designed using JDesigner software. To view the model itself, download the software from http://www.cds.caltech.edu/˜hsauro/JDesigner.htm. Then download the model and open it using the JDesigner program. Boxes represent molecules or molecular complexes. Arrows represent chemical reactions (either reversible or irreversible). Boxes labeled “Node” represent active intermediates in enzymatic reactions. Dashed boxes are alias nodes (used when the same molecular species appears at many places in the model). It should be noted that two parameters must be specified before running any simulation: first, the number (i.e., concentration) of pMHC, and second, the lifetime of the pMHC–TCR complex.
(A and B) Complete model for TCR signaling, yielding fast and sensitive ligand discrimination.
(C–K) Separate modules of the biochemical network.
(13 MB PDF).
Click here for additional data file.
Figure S7 Three Outputs of the Computer Simulation of the Early Events in Naïve T-Cell Activation
Four different pMHC ligands are tested in these simulations, with TCR-binding lifetimes of 0.3 s (representing a self-ligand involved in thymic selection of mature T cells), 3 s (antagonist), 7 s (weak agonist), and 18 s (strong agonist). These simulations were performed with 1 × 104 pMHC being presented to a naïve T cell on each APC.
(A) Kinetics of phosphorylation of the adapter.
(B) Kinetics of loading of pSHP-1 onto Lck in TCR complexes. SHP-1-decorated TCR complexes cannot drive downstream signaling. Note the transient decoration of TCR by SHP-1 for strong-agonist and weak-agonist ligands: these ligands trigger a ppERK response, which protects TCR-bound Lck from further SHP-1 binding. This transient binding of pSHP-1 to components of the TCR complex has been observed in 5C.C7 T cells activated using agonist ligand [27]
(C) Kinetics of ppERK response. Only weak- and strong-agonist pMHC can trigger a ppERK response. These simulations fit well with the experimental results presented in Figure 3B.
(72 KB PDF).
Click here for additional data file.
Figure S8 Computer Simulation of the MAPK Module
(A) Molecular scheme to simulate the MAPK cascade [33].
(B) Computer simulation of the kinetics of ppERK response in the MAPK module for different numbers of Ras–GTP ligands.
(C) Dose-response curve for the phosphorylation of ERK as a function of the number of Ras–GTP ligands. This curve could be fitted with a Hill coefficient of 13 and an EC50 of 5.3. In our model, the MAPK module acts as digital filter with a low threshold.
(D) Characteristic time of ERK phosphorylation in our computer simulation as a function of input Ras–GTP. Note the divergence of time to ppERK generation when Ras–GTP decreases towards the absolute threshold of response.
(220 KB PDF).
Click here for additional data file.
Figure S9 Modulation of Peptide-Kb Responsiveness of OT-1 T Cells at Three Stages of Differentiation
(A) Comparison of the ppERK response of OT-1 T cells to TCR restimulation at different times after activation, proliferation, and differentiation in vitro. The response was measured after 3 min of contact with RMA-S APCs pulsed with agonist (SIINFEKL) or non-agonist (EIINFEKL) peptides. The response to the agonist is essentially the same at the three stages of differentiation. Non-agonist EIINFEKL peptide does trigger a ppERK response in day-5 T cells, but only for intermediate levels of presentation.
(B) Measurement of the concentrations of different signaling molecules in OT-1 T cells 5 d after activation in vitro (day 5), or 11 d after activation in vitro (day 11). These concentrations were measured by flow cytometry after intracellular staining and are presented after normalization by the corresponding concentrations in the unstimulated naïve state (day 0). Note that SHP-1 is significantly reduced in day-5 cells.
(301 KB PDF).
Click here for additional data file.
Figure S10 Functional Response of OT-1 T Cells (6 d after Initial Activation in vitro)
Examples of the gain of functional responses (CD69 upregulation [A] and cytotoxicity [B]) to a narrow presentation range of EIINFEKL-Kb that parallel with the ppERK responses shown in Figure S9. This experiment is representative of seven experiments. In three other experiments, no acquisition of EIINFEKL-Kb responsiveness was observed on the particular day that was studied after activation. Neither was SHP-1 decreased relative to other signaling components in these particular experiments. This appears to reflect the narrow time window within which the discordance in signaling molecule concentrations occurs in these cultures, and which can vary by a day, or more in individual experiments, such that the phenomenon can be missed when only a single day post-activation is analyzed.
(202 KB PDF).
Click here for additional data file.
Protocol S1 Comparison of Kinetic Proofreading Schemes of TCR Signaling
Protocol S1 presents quantitative arguments to show that classical kinetic-proofreading schemes fail to reconcile the conjoint requirements of speed, sensitivity, and specificity in T-cell activation.
(198 KB PDF).
Click here for additional data file.
Protocol S2 Computation of the ppERK Response of T Cells at the Population Level Based on the ppERK Response at the Individual Cell Level
Protocol S2 presents a quantitative derivation of the ppERK dose response of T cells to presented ligands. The noise in the number of presented ligands and the all-or-none ppERK response of T cells are convolved to yield the final dose response of T cells. This derivation reconciles the all-or-none response of T cells as measured at the individual cell level, with the broad dose response measured at the population level.
(41 KB DOC).
Click here for additional data file.
Protocol S3 Biochemical Kinetic Parameters Used in the Computer Model of TCR Signaling in Naïve OT-1 T Cells
Protocol S3 reports all the biochemical kinetic parameters derived from biophysical and enzymological measurements drawn from the literature. These are the parameters that we used to model the early events in TCR signaling.
(83 KB DOC).
Click here for additional data file.
Protocol S4 Computer Model of the Early TCR Signaling Events
See Figure S6 for viewing instructions.
(257 KB XML).
Click here for additional data file.
Protocol S5 Reduction of the Divergence in Response Times of T Cells at Low Ligand Presentation by the Biological Variation (“Noise”) in Ligand Presentation
To match computer simulation to experiment, the theoretical output for the time required to generate a ppERK response in T cells was convolved taking into account the log-normal distribution of ligand presentation on individual APCs.
(28 KB DOC).
Click here for additional data file.
This research was supported by the Intramural Research Program of the National Institutes of Health, National Institute of Allergy and Infectious Disease. GAB wishes to acknowledge helpful discussions with and/or technical help from Irena Stefanová, Nihal Altan-Bonnet, Jérôme Delon, David Margulies, and Martin Meier-Schellersheim. Special thanks go to Ravi Rao and Herbert Sauro for updating JDesigner to produce the final versions of the network figures that appear in this paper and in the supplementary figures. GAB was supported by the Helen Hay Whitney foundation.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. GAB and RNG conceived and designed the experiments. GAB performed the experiments. GAB and RNG analyzed the data and wrote the paper.
Citation: Altan-Bonnet G, Germain RN (2005) Modeling T cell antigen discrimination based on feedback control of digital ERK responses. PLoS Biol 3(11): e356.
Abbreviations
APCantigen-presenting cell
CVcoefficient of variation
ERKextracellular signal-regulated kinase
IFNγinterferon gamma
MAPKmitogen-activated protein kinase
MSCVmouse stem-cell virus
pMHCpeptide–major histocompatibility complex molecule
ppERKdoubly phosphorylated
SHP-1SH2 domain-containing tyrosine phosphatase
TAPtransporter associated with antigen processing
TCRT cell receptor
==== Refs
References
Janeway CA Travers P Walport M Shlomchik M Immunobiology: The immune system in health and disease 2004 New York Garland Publishing
Paul WE Fundamental immunology 2003 Philadelphia Lippincott-Raven
Germain RN MHC-dependent antigen processing and peptide presentation: Providing ligands for T lymphocyte activation Cell 1994 76 287 299 8293464
Chicz RM Urban RG Lane WS Gorga JC Stern LJ Predominant naturally processed peptides bound to HLA-DR1 are derived from MHC-related molecules and are heterogeneous in size Nature 1992 358 764 768 1380674
Benoist C Mathis D Paul WE T-lymphocyte differentiation and biology Fundamental immunology, 4th ed 1999 Philadelphia Lippincott-Raven 367 409
Werlen G Hausmann B Naeher D Palmer E Signaling life and death in the thymus: Timing is everything Science 2003 299 1859 1863 12649474
Rojo JM Janeway CA The biologic activity of anti-T cell receptor V region monoclonal antibodies is determined by the epitope recognized J Immunol 1988 140 1081 1088 2449495
Gil D Schamel WW Montoya M Sanchez-Madrid F Alarcon B Recruitment of Nck by CD3 epsilon reveals a ligand-induced conformational change essential for T cell receptor signaling and synapse formation Cell 2002 109 901 912 12110186
Krogsgaard M Prado N Adams EJ He XL Chow DC Evidence that structural rearrangements and/or flexibility during TCR binding can contribute to T cell activation Mol Cell 2003 12 1367 1378 14690592
McKeithan TW Kinetic proofreading in T-cell receptor signal transduction Proc Natl Acad Sci U S A 1995 92 5042 5046 7761445
Rabinowitz JD Beeson C Wulfing C Tate K Allen PM Altered T cell receptor ligands trigger a subset of early T cell signals Immunity 1996 5 125 135 8769476
Qi SY Groves JT Chakraborty AK Synaptic pattern formation during cellular recognition Proc Natl Acad Sci U S A 2001 98 6548 6553 11371622
Eisen HN Specificity and degeneracy in antigen recognition: Yin and yang in the immune system Ann Rev Immunol 2001 19 1 21 11244028
Garcia KC Teyton L Wilson IA Structural basis of T cell recognition Ann Rev Immunol 1999 17 369 397 10358763
Aivazian D Stern LJ Phosphorylation of T cell receptor zeta is regulated by a lipid dependent folding transition Nat Struct Biol 2000 7 1023 1026 11062556
Werlen G Palmer E The T-cell receptor signalosome: A dynamic structure with expanding complexity Curr Opin Immunol 2002 14 299 305 11973126
Hogquist KA Tomlinson AJ Kieper WC McGargill MA Hart MC Identification of a naturally occurring ligand for thymic positive selection Immunity 1997 6 389 399 9133418
Hogquist KA Jameson SC Heath WR Howard JL Bevan MJ T cell receptor antagonist peptides induce positive selection Cell 1994 76 17 27 8287475
Lucas B Stefanova I Yasutomo K Dautigny N Germain RN Divergent changes in the sensitivity of maturing T cells to structurally related ligands underlies formation of a useful T cell repertoire Immunity 1999 10 367 376 10204492
Germain RN Stefanova I The dynamics of T cell receptor signaling: Complex orchestration and the key roles of tempo and cooperation Ann Rev Immunol 1999 17 467 522 10358766
Davis MM Boniface JJ Reich Z Lyons D Hampl J Ligand recognition by alpha beta T cell receptors Ann Rev Immunol 1998 16 523 544 9597140
Lord GM Lechler RI George AJ A kinetic differentiation model for the action of altered TCR ligands Immunol Today 1999 20 33 39 10081228
Li QJ Dinner AR Qi S Irvine DJ Huppa JB CD4 enhances T cell sensitivity to antigen by coordinating Lck accumulation at the immunological synapse Nat Immunol 2004 5 791 799 15247914
Matsui K Boniface JJ Steffner P Reay PA Davis MM Kinetics of T-cell receptor binding to peptide/I-Ek complexes: Correlation of the dissociation rate with T-cell responsiveness Proc Natl Acad Sci U S A 1994 91 12862 12866 7809136
Kersh GJ Kersh EN Fremont DH Allen PM High- and low-potency ligands with similar affinities for the TCR: The importance of kinetics in TCR signaling Immunity 1998 9 817 826 9881972
Sykulev Y Joo M Vturina I Tsomides TJ Eisen HN Evidence that a single peptide-MHC complex on a target cell can elicit a cytolytic T cell response Immunity 1996 4 565 571 8673703
Stefanova I Hemmer B Vergelli M Martin R Biddison WE TCR ligand discrimination is enforced by competing ERK positive and SHP-1 negative feedback pathways Nat Immunol 2003 4 248 254 12577055
Delon J Bercovici N Raposo G Liblau R Trautmann A Antigen-dependent and -independent Ca2+ responses triggered in T cells by dendritic cells compared with B cells J Exp Med 1998 188 1473 1484 9782124
Fiering S Northrop JP Nolan GP Mattila PS Crabtree GR Single cell assay of a transcription factor reveals a threshold in transcription activated by signals emanating from the T-cell antigen receptor Genes Dev 1990 4 1823 1834 2123468
Chan C George AJ Stark J Cooperative enhancement of specificity in a lattice of T cell receptors Proc Natl Acad Sci U S A 2001 98 5758 5763 11344310
Clements JL Boerth NJ Lee JR Koretzky GA Integration of T cell receptor-dependent signaling pathways by adapter proteins Ann Rev Immunol 1999 17 89 108 10358754
Alberola-Ila J Forbush KA Seger R Krebs EG Perlmutter RM Selective requirement for MAP kinase activation in thymocyte differentiation Nature 1995 373 620 623 7854419
Huang CY Ferrell JE Ultrasensitivity in the mitogen-activated protein kinase cascade Proc Natl Acad Sci U S A 1996 93 10078 10083 8816754
Xiong W Ferrell JE A positive-feedback-based bistable “memory module” that governs a cell fate decision Nature 2003 426 460 465 14647386
Townsend A Ohlen C Bastin J Ljunggren HG Foster L Association of class I major histocompatibility heavy and light chains induced by viral peptides Nature 1989 340 443 448 2666863
Krogsgaard M Li QJ Sumen C Huppa JB Huse M Agonist/endogenous peptide–MHC heterodimers drive T cell activation and sensitivity Nature 2005 434 238 243 15724150
Sporri R Reis e Sousa C Self peptide/MHC class I complexes have a negligible effect on the response of some CD8+ T cells to foreign antigen Eur J Immunol 2002 32 3161 3170 12555661
Porgador A Yewdell JW Deng Y Bennink JR Germain RN Localization, quantitation, and in situ detection of specific peptide-MHC class I complexes using a monoclonal antibody Immunity 1997 6 715 726 9208844
Alam SM Davies GM Lin CM Zal T Nasholds W Qualitative and quantitative differences in T cell receptor binding of agonist and antagonist ligands Immunity 1999 10 227 237 10072075
Rosette C Werlen G Daniels MA Holman PO Alam SM The impact of duration versus extent of TCR occupancy on T cell activation: A revision of the kinetic proofreading model Immunity 2001 15 59 70 11485738
Bode HW Network analysis and feedback amplifier 1945 New York Van Nostrand
Plas DR Johnson R Pingel JT Matthews RJ Dalton M Direct regulation of ZAP-70 by SHP-1 in T cell antigen receptor signaling Science 1996 272 1173 1176 8638162
Chan C Stark J George AJ Feedback control of T-cell receptor activation Proc R Soc Lond B Biol Sci 2004 271 931 939
Irles C Symons A Michel F Bakker TR van der Merwe PA CD45 ectodomain controls interaction with GEMs and Lck activity for optimal TCR signaling Nat Immunol 2003 4 189 197 12496963
Lorenz U Ravichandran KS Pei D Walsh CT Burakoff SJ Lck-dependent tyrosyl phosphorylation of the phosphotyrosine phosphatase SH-PTP1 in murine T cells Mol Cell Biol 1994 14 1824 1834 8114715
Dittel BN Stefanova I Germain RN Janeway CA Cross-antagonism of a T cell clone expressing two distinct T cell receptors Immunity 1999 11 289 298 10514007
Lee KH Dinner AR Tu C Campi G Raychaudhuri S The immunological synapse balances T cell receptor signaling and degradation Science 2003 302 1218 1222 14512504
Sauro HM Kholodenko BN Quantitative analysis of signaling networks Prog Biophys Mol Biol 2004 86 5 43 15261524
Racioppi L Matarese G D'Oro U De Pascale M Masci AM The role of CD4-Lck in T-cell receptor antagonism: Evidence for negative signaling Proc Natl Acad Sci U S A 1996 93 10360 10365 8816805
Jameson SC Carbone FR Bevan MJ Clone-specific T cell receptor antagonists of major histocompatibility complex class I-restricted cytotoxic T cells J Exp Med 1993 177 1541 1550 8496675
Grossman Z Singer A Tuning of activation thresholds explains flexibility in the selection and development of T cells in the thymus Proc Natl Acad Sci U S A 1996 93 14747 14752 8962126
Grossman Z Paul WE Autoreactivity, dynamic tuning and selectivity Curr Opin Immunol 2001 13 687 698 11677091
Singh NJ Schwartz RH The strength of persistent antigenic stimulation modulates adaptive tolerance in peripheral CD4+ T cells J Exp Med 2003 198 1107 1117 14530379
Cook DL Gerber AN Tapscott SJ Modeling stochastic gene expression: Implications for haploinsufficiency Proc Natl Acad Sci U S A 1998 95 15641 15646 9861023
Smith K Seddon B Purbhoo MA Zamoyska R Fisher AG Sensory adaptation in naive peripheral CD4 T cells J Exp Med 2001 194 1253 1261 11696591
Schwartz RH T cell anergy Ann Rev Immunol 2003 21 305 334 12471050
Lang KS Recher M Junt T Navarini AA Harris N Toll-like receptor engagement converts T-cell autoreactivity into overt autoimmune disease Nat Med 2005 11 138 145 15654326
Ashwell JD Fox BS Schwartz RH Functional analysis of the interaction of the antigen-specific T cell receptor with its ligands J Immunol 1986 136 757 768 2416829
Tang Q Subudhi SK Henriksen KJ Long CG Vives F The Src family kinase Fyn mediates signals induced by TCR antagonists J Immunol 2002 168 4480 4487 11970992
Davis DM Dustin ML What is the importance of the immunological synapse? Trends Immunol 2004 25 323 327 15145322
Grakoui A Bromley SK Sumen C Davis MM Shaw AS The immunological synapse: A molecular machine controlling T cell activation Science 1999 285 221 227 10398592
Slepchenko BM Schaff JC Macara I Loew LM Quantitative cell biology with the Virtual Cell Trends Cell Biol 2003 13 570 576 14573350
Samelson LE Signal transduction mediated by the T cell antigen receptor: The role of adapter proteins Ann Rev Immunol 2002 20 371 394 11861607
Hermiston ML Xu Z Weiss A CD45: A critical regulator of signaling thresholds in immune cells Ann Rev Immunol 2003 21 107 137 12414720
Willcox BE Gao GF Wyer JR Ladbury JE Bell JI TCR binding to peptide-MHC stabilizes a flexible recognition interface Immunity 1999 10 357 365 10204491
Milo R Itzkovitz S Kashtan N Levitt R Shen-Orr S Superfamilies of evolved and designed networks Science 2004 303 1538 1542 15001784
Davidson EH Rast JP Oliveri P Ransick A Calestani C A genomic regulatory network for development Science 2002 295 1669 1678 11872831
Shinkai Y Rathbun G Lam KP Oltz EM Stewart V RAG-2-deficient mice lack mature lymphocytes owing to inability to initiate V(D)J rearrangement Cell 1992 68 855 867 1547487
Stefanova I Dorfman JR Germain RN Self-recognition promotes the foreign antigen sensitivity of naive T lymphocytes Nature 2002 420 429 434 12459785
|
16231973
|
PMC1262625
|
CC BY
|
2021-01-05 08:21:45
|
no
|
PLoS Biol. 2005 Nov 25; 3(11):e356
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030356
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1623197110.1371/journal.pbio.0030368Research ArticleEvolutionDrosophilaDivergent Selection and the Evolution of Signal Traits and Mating Preferences Divergent Selection and Mate PreferencesRundle Howard D [email protected]
1
Chenoweth Stephen F
1
Doughty Paul
1
¤Blows Mark W
1
1School of Integrative Biology, University of Queensland, St. Lucia, Queensland, AustraliaKingsolver Joel Academic EditorThe University of North CarolinaUnited State of America11 2005 25 10 2005 25 10 2005 3 11 e36813 5 2005 27 8 2005 Copyright: © 2005 Rundle et al.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
New Environments Set the Stage for Changing Tastes in Mates
Mating preferences are common in natural populations, and their divergence among populations is considered an important source of reproductive isolation during speciation. Although mechanisms for the divergence of mating preferences have received substantial theoretical treatment, complementary experimental tests are lacking. We conducted a laboratory evolution experiment, using the fruit fly Drosophila serrata, to explore the role of divergent selection between environments in the evolution of female mating preferences. Replicate populations of D. serrata were derived from a common ancestor and propagated in one of three resource environments: two novel environments and the ancestral laboratory environment. Adaptation to both novel environments involved changes in cuticular hydrocarbons, traits that predict mating success in these populations. Furthermore, female mating preferences for these cuticular hydrocarbons also diverged among populations. A component of this divergence occurred among treatment environments, accounting for at least 17.4% of the among-population divergence in linear mating preferences and 17.2% of the among-population divergence in nonlinear mating preferences. The divergence of mating preferences in correlation with environment is consistent with the classic by-product model of speciation in which premating isolation evolves as a side effect of divergent selection adapting populations to their different environments.
Experimentally manipulating the resource environment of Drosophila serrata reveals that mating preferences can evolve, at least in part, as a result of environmentally-based divergent natural selection.
==== Body
Introduction
It is common in natural populations for individuals of one sex (usually females) to prefer certain trait values over others in their choice of mates [1–5]. Such mating preferences have long been thought to be key to speciation, because their divergence among populations will generate premating (behavioral) reproductive isolation. Consistent with this, mating preferences have been observed to vary among populations and closely related species in nature [5–10], and, in several taxa, evidence suggests that the resulting behavioral isolation has been instrumental in initiating speciation [11].
According to various speciation models, divergence among populations in mating preferences can occur in two main ways [12]. First, initial divergence in mating preferences in arbitrary directions may be caused by chance events such as unique mutations and/or the order in which they appear. Although such genetic drift is unlikely to cause substantial preference divergence on its own [11], sexual selection may subsequently amplify this initial divergence to yield a wide array of possible outcomes [13–16]. Speciation by sexual conflict is a popular example of such a model [16–18]. Second, initial divergence in mating preferences may be caused by divergent natural selection between environments, which may also in some [19–21], but not all [22] models, be subsequently amplified by sexual selection.
The roles of genetic drift and divergent natural selection in the diversification of mating preferences are not well understood. A number of comparative studies implicate sexual selection in speciation in a variety of taxa [11], although comparative approaches are unable to provide direct tests of the evolutionary mechanisms responsible for the initial divergence in mating preferences. The role of divergent natural selection in the evolution of premating isolation has been tested experimentally, with results clearly demonstrating the feasibility of this mechanism under some conditions [11,12,23]. Unfortunately, how divergent selection generates premating isolation is typically not known, because the mating preferences responsible are generally not identified in such experiments. A complementary experimental approach to understanding the mechanisms responsible for the diversification of mating preferences has not been developed. Such an approach is important, because the details of signal trait and preference evolution are key to distinguishing various speciation models [11].
Here we present an evolutionary experiment designed to directly test the role of divergent selection in the diversification of mating preferences. A clear expectation of how divergent selection should affect mating preferences is provided by the classic by-product model of allopatric speciation. According to this model, reproductive isolation evolves as a side effect of divergent selection adapting populations to their different environments [12,23–25]. If differences in mating preferences are a key trait contributing to reproductive isolation, independent populations adapted to different environments should diverge in mating preferences, whereas those adapted to similar environments should express the same preference. Laboratory experiments [23,26,27] and studies in nature [12,28–30] have confirmed these predictions for the evolution of premating isolation. Here, we provide an experimental test of whether these same predictions can be verified for the evolution of female mating preferences.
We tested how adaptation to two novel resource environments affected the evolution of female mating preferences in Drosophila serrata, a species in which mate choice has been investigated in a number of genetic and evolutionary experiments. D. serrata uses multiple contact pheromones, composed of nonvolatile cuticular hydrocarbons (CHCs), in both mate choice within populations [31–35] and species recognition [36,37]. Most importantly, male CHCs have been shown to respond rapidly to both natural and sexual selection [37,38], demonstrating that these signal traits readily evolve when selection is manipulated. However, how female mating preferences for male CHCs respond to divergent selection has not been determined.
We do so here by deriving 12 replicate populations from a common ancestor and propagating four of them in each of three separate treatment environments: their ancestral laboratory environment (yeast food) and two novel environments (rice and corn food). We show how the novel environments affected the evolution of CHCs and female mating preferences for them using a three-stage process. First, we demonstrate that CHCs adapted to the novel environments using the classic pattern of parallel evolution. Parallel evolution provides strong evidence that divergent selection between environments is responsible for trait evolution, because other mechanisms of evolution, such as genetic drift, are unlikely to produce similar changes in independent populations in correlation with environment [39,40].
Second, we demonstrate for each population the importance of CHCs in determining male mating success by employing population-level sexual selection gradients to estimate the form and strength of female mating preferences for male CHCs. Like many signaling systems [41,42], mate choice in D. serrata depends on the collective presence of multiple traits; here we consider the nine male CHCs shown by past studies to be associated with male mating success [31–33,35] and species recognition [36,37]. Third, to deal with the complexity generated by estimating 528 separate sexual selection gradients within a single experimental design, we employ a multivariate model fitting approach that uses partial F-tests to partition the effects of linear (directional) and nonlinear (quadratic and correlational) sexual selection within and among treatment environments. A role for divergent selection in preference evolution is demonstrated by consistent changes in preferences in correlation with treatment environment.
Results
Adaptation of Male and Female CHCs
CHCs adapted to the novel food environments, evolving in parallel in correlation with these environments. As indicated by the significant sex × treatment interaction (Table 1), the response to selection differed in males and females. CHCs also varied significantly among populations within the treatment environments (Table 1). Examination of the first canonical variate (CV) of the sex × treatment interaction (CV1, the linear combination of eight logcontrast-transformed CHCs that explains the most variance—85.2% in this case—in the sex × treatment interaction) reveals that, relative to the populations in the ancestral yeast environment, sexual dimorphism in the combination of CHCs that responded to selection tended to increase in populations adapted to rice and decrease in populations adapted to corn (Figure 1). When the sexes were analyzed separately, the treatment effect was significant in females (p = 0.018) but not in males (p = 0.153), indicating greater adaptation to the treatment environments by females than males.
Figure 1 Adaptation of Male and Female CHCs to the Different Treatment Environments
Variation among populations is presented as the first CV of the sex × treatment interaction from a MANOVA of the eight logcontrast CHCs of individuals from the 12 populations. Males are represented by filled symbols, and females by open symbols. The four replicate populations within each treatment are indicated by the different shaped symbols (there is no correspondence among treatment environments of populations represented by the same symbol).
Table 1 MANOVA Testing the Effects of Various Sources of Variation on the Eight Logcontrast Transformed CHCs Measured on Virgin Flies from 12 Experimental Populations of D. serrata
Sexual Selection on Male CHCs Within Populations
Consistent with results from past studies [31–35], female mating preferences generated strong sexual selection on male CHCs in these populations (Table S1). Overall, linear sexual selection on the eight logcontrast CHCs was significant in each of the 12 populations (p < 0.0001 in all cases) and explained 29%–59% (mean 46%) of the variance in male mating success (Table 2). The addition of all nonlinear sexual selection was also highly significant overall in each of the 12 populations (p < 0.0001 in all cases), and the combination of linear and nonlinear selection explained 38%–68% (mean 56%) of the variance in male mating success (Table 2).
Table 2 Proportion of Total Variation in Male Mating Success Accounted for by Sexual Selection on Male CHCs
Variation among Populations and Treatments in Sexual Selection on Male CHCs
Sexual selection on male CHCs generated by female mating preferences varied significantly among populations overall, both for linear and nonlinear selection (partial F-tests, p < 0.0001 in both cases; Table 3). This indicated that female mating preferences diverged among populations.
Table 3 Linear Models and Partial F-Tests on Significance of Variation
Sexual selection on the eight male logcontrast CHCs also varied consistently among treatment environments, indicating that at least a component of the among-population divergence in mating preferences occurred in correlation with environment. This variation approached significance for linear selection (partial F-test, p = 0.054; Table 3) and accounted for at least 17.4% of the among-population variation in linear selection. For nonlinear sexual selection, this among-treatment variation was highly significant (partial F-test, p = 0.0011; Table 3) and accounted for at least 17.2% of the among-population variation in nonlinear selection. Because the overall importance of CHCs in explaining variation in male mating success was similar in all three environments (see Table 2), there is no indication that females in either novel environment became more or less choosy with regard to male CHCs. Rather, it is the combinations of male CHCs associated with mating success that changed.
Variation among treatments in linear sexual selection was greatest for CHC 2-Me-C26, although the univariate interaction of this trait with treatment was not significant (F
2,2517 = 2.47, p = 0.085). The next greatest contribution to among-treatment variation in linear sexual selection was made by CHC 2-Me-C28, although again the univariate interaction of this trait with treatment was non-significant (F
2,2517 = 1.55, p = 0.213). To provide a visual interpretation of how sexual selection varied among treatment environments, we used nonparametric thin-plate splines [43] to explore the bivariate fitness surfaces associated with these two CHCs. In the ancestral yeast environment, the fitness surface resembled a sloping plane, consistent with past experiments in this environment [33]. In the two novel environments, the slope of the fitness surfaces decreased, and combinations of CHCs that were unattractive in the ancestral environment (i.e., high values of 2-Me-C26 and low values of 2-Me-C28) appeared to have increased in attractiveness (Figure 2A).
Figure 2 Thin-Plate Spline Representations of Bivariate Fitness Surfaces for Male CHCs for Which Female Mating Preferences Evolved in Correlation with Treatment Environment
(A) Visualization of the fitness surface of the two male CHCs for which linear sexual selection varied most among treatments.
(B) Visualization of the fitness surface of the two male CHCs for which nonlinear sexual selection varied most among treatments.
The four replicate populations within each treatment were pooled in each case. To aid in comparisons across treatments, a single smoothing parameter (λ) was chosen that gave the lowest generalized cross-validation score in all three treatments [43] separately for (A) (λ = −1.0) and (B) (λ = −0.2).
Among-treatment variation in nonlinear sexual selection was contributed to most by correlational selection between CHCs (Z)-9-C25:1 × 2-Me-C28 and (Z)-9-C25:1 × 2-Me-C26; univariate trait × treatment interactions were significant for each ([Z]-9-C25:1 × 2-Me-C28, F
2,2409 = 7.04, p < 0.001; [Z]-9-C25:1 × 2-Me-C26, F
2,2409 = 4.41, p = 0.012). The nonparametric visualization of the bivariate fitness surface generated by the first pair of CHCs (Figure 2B) indicates how correlational selection between them has changed across treatments. The highest fitness combination in the ancestral yeast environment (i.e., high values of both [Z]-9-C25:1 and 2-Me-C26) has low mating success in the corn environment, changing the curvature along one diagonal of the surface from concave (bowl-shaped) in yeast to convex (humped) in corn.
Discussion
Mating preferences may diverge between populations as a result of genetic drift or because of divergent natural selection between environments [12]. Direct experimental tests of either of these mechanisms, however, are lacking. Here we used an experimental evolution approach, involving an ancestral laboratory and two novel resource environments, to evaluate the role of divergent selection in the evolution of female mating preferences among replicate populations of D. serrata. CHCs, the traits that predict male and female mating success, evolved in response to the new environments. Furthermore, as determined using more than 1,250 independent mate choice trials, female mating preferences for these same CHCs also diverged among populations, with a component of this divergence occurring consistently among treatment environments. This provides a direct experimental demonstration that mating preferences can evolve, at least in part, as a result of environmentally based divergent selection.
An important component of our experimental design was the choice of novel environments to which D. serrata populations were exposed. Mate choice in the D. serrata species complex is based largely on nonvolatile CHCs [31–37]. Divergent selection was applied using different resource environments, because CHC expression in insects depends on the amino acids available in their diet [44], and Drosophila CHCs are known to be affected by larval substrate [45]. It is therefore not surprising that these environments generated divergent natural selection on CHCs, nor that adaptation occurred in every CHC except 2-Me-C26, which alone did not display a significant univariate sex × treatment interaction (p = 0.420). However, male and female CHCs responded very differently to the novel environments. The response to selection of homologous traits will depend on the different selective optima for each sex and on the intersex genetic correlations [46]. Our results demonstrate that the level of sexual dimorphism in this species can evolve as predicted from the generally low intersex genetic correlations for these traits [32].
Under the assumption that the divergence of mating preferences among populations is a key source of premating isolation, our results are consistent with the classic by-product model of allopatric speciation. In this model, new species arise as a side effect of natural selection adapting populations to their different environments [12,23–25]. Although premating isolation has been shown to evolve in parallel in correlation with environment [12,23,26–30], little is known about the role of mating preferences in this process. Here we have used a direct experimental manipulation to show that mating preferences, a trait long thought to be an important source of premating isolation, can evolve, at least in part, in correlation with environment. Linear sexual selection exerted by female mating preferences diverged in the novel environments primarily on two male CHCs: 2-Me-C26 and 2-Me-C28. Correlational sexual selection exerted by female preferences between these two CHCs and a third, (Z)-9-C25:1, was also changed by selection in the novel environments.
There are two classes of mechanisms by which divergent selection between environments could cause consistent mating preference evolution. In the first, CHCs evolve by divergent natural selection, adapting to the different treatment environments. Mating preferences evolve as a correlated response to this, due to either pleiotropy or linkage disequilibrium between the genes involved in adaptation and those affecting mating preferences [19,21]. In the second, mating preferences diverge between environments either due to direct natural selection on the preferences themselves [20,47] or as a plastic response to the different environments. This preference divergence generates divergent sexual selection on signal traits (CHCs), and subsequent CHC divergence may feed back on preferences, furthering their divergence as a correlated response. Determining the mechanism responsible for preference divergence among environments in our populations will thus require additional information about the independent and combined roles of natural and sexual selection in the evolution of CHCs and mating preferences. This will require further experiments that independently manipulate the presence and absence of both forms of selection within a single experimental design [38].
Surprisingly, there was no strong association in our experiment between the CHCs that adapted and those involved in preference changes. Linear preferences changed most among environments for 2-Me-C26. However, as noted earlier, 2-Me-C26 is the only CHC for which there was no evidence of adaptation. Nonlinear preferences changed for correlational selection between CHCs (Z)-9-C25:1 and 2-Me-C26, and (Z)-9-C25:1 and 2-Me-C28. Both (Z)-9-C25:1 and 2-Me-C28 do show strong evidence of adaptation, although primarily in females. That adaptation of CHCs occurred to a greater extent in females than in males may seem surprising, given that models of sexual selection predict the coevolution of signal traits and preferences [13]. This could arise, however, if male CHCs were evolving primarily by sexual selection; it may take time for divergent female preferences to become established and to subsequently affect male trait evolution. Alternatively, if CHCs are condition-dependent and males allocate resources preferentially to them, CHCs themselves might evolve less than other traits competing for the available resources. Whatever the reason, the lack of an association between the evolution of male signal traits and female preference is a clear evolutionary outcome of our experiment and suggests that the evolution of signal traits and preferences in novel environments may be a more complex process than is currently appreciated.
All else being equal, the consistent evolution of mating preferences in correlation with environment should eventually cause the parallel evolution of premating isolation [12,26–30]. In our experiment, however, genetic drift was almost certainly responsible for a component of the among-population divergence in mating preferences, and this may ultimately counter the parallel evolution of preferences and premating isolation. Because we have only a single estimate of the linear and nonlinear mating preferences (sexual selection gradients) for each population, variation among populations within treatments is a consequence of both genetic drift and measurement error associated with each selection gradient. It is therefore not possible to distinguish between drift and measurement error in contributing to preference divergence among the populations in our experiment. Quantifying the importance of genetic drift in future experiments will be difficult because it requires the repeated measurement of population-level selection gradients.
Finally, changes in population-level mate preferences (i.e., sexual selection gradients) are ultimately the product of evolutionary changes in individual preference functions and/or their frequency within populations [48]. A complete understanding of how mating preferences diverge will thus also require additional studies that explore how natural selection and genetic drift affect the evolution of individual preferences. Knowledge of the underlying evolution in individual preferences responsible for changes in population-level mating preferences may have important implications for our understanding of speciation, because the shape and diversity of individual preference functions within populations will affect the evolution of reproductive isolation between populations. Determining how preference evolution generates reproductive isolation remains a central and untested issue in speciation research.
Materials and Methods
Experimental populations
In April of 2002, 12 populations of D. serrata were independently derived from the Forster stock previously described [31,32,37]. These populations were assigned to one of three treatment environments (yeast, rice, or corn), yielding four replicate populations in each environment. The environments varied in the food medium on which the flies were raised. The yeast treatment, representing the ancestral environment, used the same food medium that has been used to maintain this stock since its establishment in the laboratory in January, 1998. This is a standard laboratory medium (Table S2). In the novel environments, the majority of the yeast was replaced with either rice flour (rice environment) or corn starch (corn environment; Table S2). These two new environments were chosen because the expression of CHCs in insects is dependent on the amino acids available [44], and Drosophila CHCs are known to be affected by larval substrate [45]; different substrates have also been shown to affect the strength of reproductive isolation in another species (D. mojavensis) [49,50].
Populations were maintained with nonoverlapping generations by transferring approximately 100 individuals of unknown sex into each of two new bottles every generation. In October 2003, prior to measuring their CHCs and mating preferences, all flies from every population were raised on the ancestral yeast medium for two generations to remove any environmental effects. At this time, the populations on yeast were 37 generations old. Because generation times were slower in the novel environments, the corn and rice populations had each evolved for 29 generations.
Measurements of CHCs
CHCs were extracted from individual flies as previously described [36]. Samples were analyzed using gas chromatography (GC) and flame ionization detection on an Agilent Technologies (Wilmington, Delaware, United States) 6890N gas chromatograph fitted with a HP5 column of 50 m × 0.32 mm internal diameter and a pulsed splitless front inlet. The temperature program began by holding at 57 °C for 1.1 min, then increased to 190 °C at 100 °C min−1, followed by 190–310 °C at 25 °C min−1, then finally holding at 310 °C for 5 min. Individual CHC profiles were determined by integration of the area under nine peaks. These are the same peaks as used in past studies [31–35], identified in order of their retention times as: (Z,Z)-5,9-C24:2; (Z,Z)-5,9-C25:2; (Z)-9-C25:1; (Z)-9-C26:1; 2-Me-C26; (Z,Z)-5,9-C27:2; 2-Me-C28; (Z,Z)-5,9-C29:2; and 2-Me-C30 [51].
Relative amounts of each of these nine CHCs were calculated for each individual by dividing the area under each peak by the total area under all their peaks. In GC analysis, this use of proportional peak areas is favored over absolute values because, even with the use of internal size standards, absolute values are often subject to large experimental error [36,52]. The use of proportions introduces nonindependence among peaks within the CHC profiles of individuals, because the area under any one peak influences the total area and thus the proportional values of other peaks. This unit-sum constraint, characteristic of compositional data, is removed using a logcontrast transformation [36,53]. Logcontrasts were calculated using the proportional area of (Z,Z)-5,9-C24:2 as the divisor, yielding eight logcontrast peak values for every individual. Because logcontrast peak values derive ultimately from proportional data, all analyses described below address changes in the relative abundance of CHCs within individuals. Statistical analyses were performed using SAS v. 8.2 (SAS Institute, Cary, North Carolina, United States).
Adaptation of male and female CHCs
Ten virgin males and ten virgin females were collected from each population at eclosion using light CO2 anesthesia. Flies were held separately by sex in vials of five flies/vial; males were transferred singly into new vials after 2 d. All holding vials included a small amount of live yeast sprinkled on top of the food. At 4 d post-eclosion, CHC profiles of all flies were determined using GC as described above.
Differences in the CHC profiles of males and females from each population were tested using the following linear model:
where Y is the logcontrast peak value for the ith CHC (i = 1 − 8), S
j is the effect of sex, T
a is the effect of treatment environment (yeast, rice, or corn), ST
ja is the effect of the interaction of sex with treatment, P(T)
b(a) is the effect of replicate populations nested within treatment, and ɛ is the error. Population was modeled as a random effect whereas all other factors were fixed. MANOVA was used to test the significance of all terms in this model. Because of the nested design, significance of the treatment effect was tested over the mean square (MS) of P(T). All other terms were tested over MSerror.
This model yielded a significant sex × treatment interaction, so to isolate the combination of CHCs that were responsible for this response to selection, we conducted a multifactorial canonical discriminant analysis. This entailed performing an eigenanalysis on the resultant matrix of E−1H, where E was the sums of squares and cross-product (SSCP) matrix of the ɛ
ijk term, and H was the SSCP hypothesis matrix for the sex × treatment interaction [54]. The first two eigenvectors (CVs) accounted for 100% of the variance at the sex × treatment level.
Measurement of female preferences for male CHCs
Female mating preferences can be measured in two fundamentally different ways [48]. First, preference functions can be determined for individual females [4]. While this approach allows the diversity of female preferences within populations to be visualized, statistical descriptions of these functions can be complex, making comparisons among populations (our goal here) difficult. Alternatively, population-level sexual selection gradients [55] may be used to describe the form and strength of sexual selection (female preference) operating on male traits. We use this approach because comparison of sexual selection gradients among populations is straightforward to accomplish within a well-described modeling framework [56].
Female choice trials were conducted in glass vials containing 8 ml of standard yeast medium and using 4-d-old virgin flies that had been held as described above. In each trial, a single female from one of the experimental populations was placed together with two 4-d-old virgin males from the ancestral laboratory population (Forster). These males were used to provide a standard for comparison of female preferences among populations; their use means that any differences among populations can be attributed unambiguously to the evolution of female preference as opposed to the males from which the females are choosing. Forster males were raised and held using the same protocol as described above for the experimental populations.
An average of 106 trials (range 96–112) were conducted for each of the 12 populations, and mating occurred in nearly all (> 99%) of them. Mating vials were observed, and once intromission had been achieved between the female and one of the two males, all flies were anesthetized with CO2 and the chosen and rejected males had their CHCs extracted for subsequent GC analysis (females were discarded). CHC profiles of each male were integrated and proportional peak areas were logcontrast transformed.
Characterizing sexual selection within populations
Linear sexual selection on the eight male logcontrast CHCs arising from female mating preferences was analyzed separately in each population using the standard first-order polynomial regression model [55]. Although male mating success was binomially distributed in these analyses, parametric significance testing was performed in all cases using standard methods within a linear model framework, because when sample sizes are large and the probability of either outcome is equal (as in the present case), the binomial distribution provides an excellent approximation of the normal distribution. Results of bootstrap analyses from past experiments have confirmed the accuracy of this approximation [32]. Males were treated as independent replicates in these analyses; this has no discernable effect on the significance of individual selection gradients when compared with treating females as replicates (Figure S1 and S2; Protocol S1).
To confirm the presence of sexual selection on male CHCs by female mating preferences in our populations, we estimated, separately for each population, the strength of linear sexual selection on the eight logcontrast CHCs for each male. Similar to past studies [31–35], multicollinearity among these logcontrast CHCs was minimal, so these values were used directly in the analysis. From these regressions, the proportion of total variation in male mating success accounted for by linear sexual selection on all eight logcontrast CHCs was given by the adjusted coefficient of determination (r
2), with significance indicated by the overall fit of the model.
To evaluate the overall significance of nonlinear selection on the eight logcontrast CHCs in our populations, we first conducted a canonical rotation to place all nonlinear selection on the eight eigenvectors of the matrix of quadratic and cross-product terms (i.e., the gamma matrix) in each population, thus eliminating all cross-product terms [57]. Nonlinear sexual selection on these eight eigenvectors was then analyzed using the standard second-order polynomial regression model [55,58]. The overall significance of all nonlinear selection in each population was then evaluated using partial F-tests [35,59] that compared the fit of the models with and without the eight quadratic terms. Conducting such a rotation does not affect the amount of nonlinear selection present in each population, but does increase the likelihood of detecting its significance by reducing the number of nonlinear coefficients from 36 to eight.
Variation among populations in sexual selection
To determine if female mating preferences diverged among populations, we tested for variation among populations in both linear and nonlinear sexual selection on male CHCs. Among-population variation in linear sexual selection was tested using the following model:
where Y is the mating success of the ith male from the bth population (b = 1 − 12). C
l is the effect on male mating success of the lth male logcontrast CHC, representing linear sexual selection on this trait. Variation among populations in linear sexual selection on male logcontrast CHC peak value l would be indicated by a significant C
l
P
b interaction. To evaluate whether linear sexual selection varied among the twelve populations, we used a single partial F-test [35,59] that compared the fit of the above model with one lacking all of the C
l
P
b interactions.
Among-population variation in nonlinear sexual selection was evaluated in an analogous manner using the following model:
where C
l
C
m is the combined effect on male mating success of the lth and mth male logcontrast CHCs, representing nonlinear selection (quadratic: l = m; correlational: l ≠ m) on these traits. Variation among populations in nonlinear sexual selection would be indicated by a significant C
l
C
m
P
b interaction. To evaluate whether nonlinear selection varied among populations overall, we again used a single partial F-test [35,59] that compared the fit of this model with one lacking all of the C
l
C
m
P
b interactions. By excluding from the reduced model the interactions of all nonlinear terms with population, significance of this partial F-test reflects the combined importance of among-population variation in all forms of nonlinear sexual selection.
Variation among treatments in sexual selection
To determine if natural selection adapting populations to their novel treatment environments caused consistent mating preference divergence, we tested for variation among treatments in both linear and nonlinear sexual selection on male CHCs. Among-treatment variation in linear sexual selection was tested using the following model:
where the bth population is nested within the ath treatment environment (yeast, rice, and corn). Treatment was modeled as a fixed effect and population was modeled as a random effect nested within treatment. Variation among treatments in linear sexual selection on male logcontrast CHC peak value l would be indicated by a significant C
l
T
a interaction. To evaluate whether linear sexual selection varied among treatments overall, we used a single partial F-test [35,59] that compared the fit of the above model with one lacking all of the C
l
T
a interactions.
Among-treatment variation in nonlinear sexual selection was evaluated in a similar manner using the following model:
Variation among treatments in nonlinear sexual selection would be indicated by a significant C
l
C
m
T
a interaction in this model. To evaluate whether nonlinear selection varied among treatments overall, we again used a single partial F-test [35,59] that compared the fit of this model with one lacking all of the C
l
C
m
T
a interactions.
The contribution of divergent selection to the among-population diversification of mating preferences was quantified using MS ratios. For linear sexual selection, the total among-population variation is represented by the MS (C
l
P
b) from Equation 2, and the among-treatment variation by the MS (C
l
T
a) from Equation 4. Their ratio, MS (C
1
T
a) / MS (C
l
P
b), is an estimate of the proportion of the total among-population variation in mating preferences that is grouped by treatment environment. For nonlinear sexual selection, this ratio is MS (C
l
C
m
T
a) / MS (C
l
C
m
P
b), obtained from Equations 5 and 3, respectively, and is an estimate of the proportion of the total among-population variation in nonlinear mating preferences that is grouped by treatment environment.
Supporting Information
Figure S1 Quantifying Potential Bias in the Magnitude of Selection Gradients Caused by Pseudoreplication
Mean sexual selection gradients on eight logcontrast male CHCs from three geographic populations are presented. For each population, mean selection gradients were estimated from 1,000 bootstrap replicates each of two subpopulations composed of 128 males (64 chosen, 64 rejected) randomly sampled from the population of 256 males. In subpopulation A, 64 trials were randomly selected, and both the chosen and rejected males were used. In subpopulation B, a single male (chosen or rejected) was sampled from each of the 128 trials. The line is a one-to-one line.
(19 KB CDR).
Click here for additional data file.
Figure S2 Quantifying Potential Bias in the Significance Level of Selection Gradients Caused by Pseudoreplication
Mean significance levels for the sexual selection gradients on eight logcontrast male CHCs from three geographic populations estimated from 1,000 bootstrap replicates of subpopulations A and B as described in Figure S1. The line is a one-to-one line.
(19 KB CDR).
Click here for additional data file.
Table S1 Standardized Linear and Nonlinear Sexual Selection Gradients on the Eight Male CHCs for Each of the 12 Populations
(45 KB WPD).
Click here for additional data file.
Table S2 Media Recipes for the Three Treatment Environments
(19 KB WPD).
Click here for additional data file.
Protocol S1 Pseudoreplication and Multiple-Choice Mating Trials
(24 KB WPD).
Click here for additional data file.
We thank M Higgie, G Joseph, and A Skroblin for help conducting the work, RB O'Hara and M Whitlock for statistical advice, and M Higgie for providing the unpublished data used in the Supporting Information. Comments from four anonymous reviewers also helped improve the manuscript. This research was supported by grants from the Australian Research Council (HDR, SFC, and MWB) and the University of Queensland (PD and HDR).
Competing interests. The authors have declared that no competing interests exist.
Author contributions. HDR, SFC, PD, and MWB conceived and designed the experiments. HDR, SFC, and MWB performed the experiments. HDR and MWB analyzed the data. HDR, SFC, and MWB wrote the paper.
¤ Current address: Western Australian Museum, Perth, Western Australia, Australia
Citation: Rundle HD, Chenoweth SF, Doughty P, Blows MW (2005) Divergent selection and the evolution of signal traits and mating preferences. PLoS Biol 3(11): e368.
Abbreviations
CHCcuticular hydrocarbon
CVcanonical variate
GCgas chromatography
MSmean square
==== Refs
References
Kirkpatrick M Sexual selection by female choice in polygynous animals Annu Rev Ecol Syst 1987 18 43 70
Houde AE Endler JA Correlated evolution of female mating preference and male colour patterns in the guppy Poecilia reticulata
Science 1990 248 1405 1408 17747527
Andersson M Sexual selection 1994 Princeton Princeton University Press 624
Ritchie MG The shape of female mating preferences Proc Natl Acad Sci U S A 1996 93 14628 14631 8962104
Endler JA Houde AE Geographic variation in female preferences for male traits in Poecilia reticulata
Evolution 1995 49 456 468
Wells MM Henry CS Howard DJ Berlocher SH Songs, reproductive isolation, and speciation in cryptic species of insects Endless forms: Species and speciation 1998 Oxford Oxford University Press 217 233
Foster SA Endler JA Geographic variation in behavior: Perspectives on evolutionary mechanisms 1999 Oxford Oxford University Press 336
Simmons LW Zuk M Rotenberry JT Geographic variation in female preference functions and male songs of the field cricket Teleogryllus oceanicus
Evolution 2001 55 1386 1394 11525462
Boughman JW Divergent sexual selection enhances reproductive isolation in sticklebacks Nature 2001 411 944 948 11418857
Kwiatkowski MA Sullivan BK Kwiatkowski MA, Sullivan BK (2002) Geographic variation in sexual selection among populations of an iguanid lizard, Sauromalus obesus (= ater ) Evolution 2002 56 2039 2051 12449491
Coyne JA Orr HA Speciation 2004 Sunderland (Massachusetts) Sinauer Associates 545
Schluter D The ecology of adaptive radiation 2000 Oxford Oxford University Press 288
Lande R Models of speciation by sexual selection on polygenic traits Proc Natl Acad Sci U S A 1981 78 3721 3725 16593036
Kirkpatrick M Sexual selection and the evolution of female choice Evolution 1982 36 1 12
West-Eberhard M Sexual selection, social competition, and speciation Q Rev Biol 1983 58 155 183
Rice WR Howard DJ Berlocher SH Intergenomic conflict, interlocus antagonistic coevolution, and the evolution of reproductive isolation Endless forms: Species and speciation 1998 Oxford Oxford University Press 261 270
Gavrilets S Rapid evolution of reproductive barriers driven by sexual conflict Nature 2000 403 886 889 10706284
Gavrilets S Waxman D Sympatric speciation by sexual conflict Proc Natl Acad Sci U S A 2002 99 10533 10538 12149438
Lande R Rapid origin of sexual isolation and character divergence in a cline Evolution 1982 36 213 223
Endler JA Signals, signal conditions, and the direction of evolution Am Nat 1992 139 S125 S153
Schluter D Price T Honesty, perception and population divergence in sexually selected traits Proc R Soc London Ser B 1993 253 117 122
Ryan MJ Rand AS Species recognition and sexual selection as a unitary problem in animal communication Evolution 1993 47 647 657
Rice WR Hostert EE Laboratory experiments on speciation: What have we learned in 40 years? Evolution 1993 47 1637 1653
Mayr E Systematics and the origin of species 1942 New York Columbia University Press 368
Dobzhansky T Genetics and the origin of species, 3rd Ed 1951 New York Columbia University Press 364
Kilias G Alahiotis SN Pelecanos M A multifactorial genetic investigation of speciation theory using Drosophila melanogaster
Evolution 1980 34 730 737
Dodd DMB Reproductive isolation as a consequence of adaptive divergence in Drosophila pseudoobscura
Evolution 1989 43 1308 1311
Rundle HD Nagel L Boughman JW Schluter D Natural selection and parallel speciation in sympatric sticklebacks Science 2000 287 306 308 10634785
Nosil P Crespi BJ Sandoval CP Host-plant adaptation drives the parallel evolution of reproductive isolation Nature 2002 417 440 443 12024213
McKinnon JS Mori S Blackman B David L Kingsley D Evidence for ecology's role in speciation Nature 2004 429 294 298 15152252
Hine E Lachish S Higgie M Blows MW Positive genetic correlation between female preference and offspring fitness Proc R Soc London Ser B 2002 269 2215 2219
Chenoweth SF Blows MW Signal trait sexual dimorphism and mutual sexual selection in Drosophila serrata
Evolution 2003 57 2326 2334 14628920
Blows MW Chenoweth SF Hine E Orientation of the genetic variance-covariance matrix and the fitness surface for multiple male sexually selected traits Am Nat 2004 163 329 340 15026971
Petfield D Chenoweth SF Rundle HD Blows MW Genetic variance in female condition predicts indirect genetic variance in male sexual display traits Proc Natl Acad Sci U S A 2005 102 6045 6050 15840726
Chenoweth SF Blows MW Contrasting mutual sexual selection on homologous signal traits in Drosophila serrata
Am Nat 2005 165 281 289 15729657
Blows MW Allan RA Levels of mate recognition within and between two Drosophila species and their hybrids Am Nat 1998 152 826 837 18811430
Higgie M Chenoweth S Blows MW Natural selection and the reinforcement of mate recognition Science 2000 290 519 521 11039933
Blows MW Interaction between natural and sexual selection during the evolution of mate recognition Proc R Soc London Ser B 2002 269 1113 1118
Endler JA Natural selection in the wild 1986 Princeton Princeton University Press 336
Schluter D Nagel LM Parallel speciation by natural selection Am Nat 1995 146 292 301
Moore AJ The evolution of social signals: Morphological, functional, and genetic integration of the sex pheromone in Nauphoeta cinerea
Evolution 1997 51 1920 1928
Brooks R Endler JA Direct and indirect sexual selection and quantitative genetics of male traits in guppies (Poecelia reticulata)
Evolution 2001 55 1002 1015 11430637
Blows MW Brooks R Kraft PG Exploring complex fitness surfaces: Multiple ornamentation and polymorphism in male guppies Evolution 2003 57 622 630
Nelson DR Stanley-Samuelson DW Nelson DR Methyl-branched lipids in insects Insect lipids: Chemistry, biochemistry and biology 1993 Lincoln (Nebraska) University of Nebraska Press 271 315
Toolson EC Markow TA Jackson LL Howard RW Epicuticular hydrocarbon composition of wild and laboratory-reared Drosophila mojavensis Patterson and Crow (Diptera: Drosophilidae) Ann Entomol Soc Am 1990 83 1165 1176
Lande R Sexual dimorphism, sexual selection, and adaptation in polygenic characters Evolution 1980 34 292 305
Boughman JW How sensory drive can promote speciation Trends Ecol Evol 2002 17 571 577
Wagner WE Measuring female mating preferences Anim Behav 1998 55 1029 1042 9632487
Etges WJ Premating isolation is determined by larval substrates in cactophilic Drosophila mojavensis
Evolution 1992 46 1945 1950
Stennett MD Etges WJ Premating isolation is determined by larval rearing substrates in cactophilic Drosophila mojavensis . III. Epicuticular hydrocarbon variation is determined by use of different host plants in Drosophila mojavensis and Drosophila arizonae
J Chem Ecol 1997 23 2803 2824
Howard RW Jackson LL Banse H Blows MW Cuticular hydrocarbons of Drosophila birchii and D. serrata Identification and role in mate choice in D. serrata
J Chem Ecol 2003 29 961 976 12775155
Savarit F Ferveur JF Genetic study of the production of sexually dimorphic cuticular hydrocarbons in relation with the sex-determination gene transformer inDrosophila melanogaster
Genet Res Camb 2002 79 23 40
Atchison J The statistical analysis of compositional data 1986 London Chapman and Hall 416
Rencher AC Methods of multivariate analysis 1995 New York Wiley 627
Lande R Arnold SJ The measurement of selection on correlated characters Evolution 1983 37 1210 1226
Draper NR John JA Response-surface designs for quantitative and qualitative variables Technometrics 1988 30 423 428
Blows MW Brooks R Measuring nonlinear selection Am Nat 2003 162 815 820 14737718
Brodie ED Moore AJ Janzen FJ Visualizing and quantifying natural selection Trends Ecol Evol 1995 10 313 318 21237054
Bowerman BL O'Connell RT Linear statistical models: An applied approach 1990 Belmont (California) Duxbury Press 1024
|
16231971
|
PMC1262626
|
CC BY
|
2021-01-05 08:21:28
|
no
|
PLoS Biol. 2005 Nov 25; 3(11):e368
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030368
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1623197210.1371/journal.pbio.0030371Research ArticleEcologyEvolutionInfectious DiseasesEpidemiology/Public HealthVirusesPrimatesWave-Like Spread of Ebola Zaire Wave-Like Spread of Ebola ZaireWalsh Peter D [email protected]
1
Biek Roman
2
Real Leslie A
2
1 Max-Planck-Institute for Evolutionary Primatology, Leipzig, Germany,2 Department of Biology, Emory University, Atlanta, Georgia, United States of AmericaHarvey Paul Academic EditorUniversity of OxfordUnited Kingdom11 2005 25 10 2005 25 10 2005 3 11 e37112 4 2005 31 8 2005 Copyright: © 2005 Walsh et al.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Charting the Path of the Deadly Ebola Virus in Central Africa
In the past decade the Zaire strain of Ebola virus (ZEBOV) has emerged repeatedly into human populations in central Africa and caused massive die-offs of gorillas and chimpanzees. We tested the view that emergence events are independent and caused by ZEBOV variants that have been long resident at each locality. Phylogenetic analyses place the earliest known outbreak at Yambuku, Democratic Republic of Congo, very near to the root of the ZEBOV tree, suggesting that viruses causing all other known outbreaks evolved from a Yambuku-like virus after 1976. The tendency for earlier outbreaks to be directly ancestral to later outbreaks suggests that outbreaks are epidemiologically linked and may have occurred at the front of an advancing wave. While the ladder-like phylogenetic structure could also bear the signature of positive selection, our statistical power is too weak to reach a conclusion in this regard. Distances among outbreaks indicate a spread rate of about 50 km per year that remains consistent across spatial scales. Viral evolution is clocklike, and sequences show a high level of small-scale spatial structure. Genetic similarity decays with distance at roughly the same rate at all spatial scales. Our analyses suggest that ZEBOV has recently spread across the region rather than being long persistent at each outbreak locality. Controlling the impact of Ebola on wild apes and human populations may be more feasible than previously recognized.
Genetic, spatial, and temporal data reveal that the Zaire strain of Ebola virus has spread recently across the region rather than being a long-term resident in the locations where the outbreaks have occurred.
==== Body
Introduction
In the past decade, the highly virulent Zaire strain of Ebola virus (ZEBOV) has repeatedly emerged into rural human populations in Gabon and Republic of Congo [1,2] (Figure 1). Compelling genetic evidence [2] suggests that ZEBOV entered human populations when people handled infected carcasses of western gorillas (Gorilla gorilla) and common chimpanzees (Pan troglodytes) during massive ape die-offs [3,4]. The risk of new ZEBOV outbreaks in the two countries poses a continuing threat to humans as well as to the largest remaining gorilla and chimpanzee populations in the world.
Figure 1 Maps of Ebola Zaire Outbreaks
(A) Human outbreak locations in Gabon and Congo as reported [2]. Also shown are October 2003 human outbreak at Mbandza village and April 2004 ape die-off around Iboundji (Lokoué) Clearing in Odzala National Park. Yellow arrows represent epizootic path suggested by phylogenetic analyses.
(B) Sites of all primary outbreaks of Ebola Zaire in humans documented [2,14] and the epizootic path suggested by the spatio-temporal pattern of outbreaks (yellow arrows). Best fitting origin found through ML search for the spatial location that produced the strongest correlation between outbreak date and geographic distance from the origin. ML search based on the correlation between patristic genetic distance and spatial separation between outbreaks places the epizootic pivot point just southeast of Booué. In both figures, shading of circles is proportional to time after first outbreak in series.
It seems highly unlikely that ZEBOV has caused similarly damaging ape die-offs in Gabon and Congo during the past century. Ape reproductive rates are so low that recovery from population reductions as dramatic as recently caused by ZEBOV would take 75 years or more [4]. Thus, if large ape die-offs had occurred in the past half-century, one would expect to find large zones of low ape density. Extensive surveys conducted during the 1980s and early 1990s [5–10] showed no such evidence. This observation is consistent with the fact that no human outbreaks of ZEBOV were recognized in Gabon or Congo before 1994. The big question is, thus: Why has ZEBOV now emerged so explosively?
There are two contrasting answers to this question. First, many authors have either assumed [11–14,] or concluded [2,15–19] that ZEBOV has long been present in the region and that its emergence is due to an increase in the rate at which human or non-human apes come into contact with some yet-to-be-identified reservoir host. Both habitat disturbance [1,16] and climatic factors [2,20,21] have been proposed as triggers for ZEBOV emergence. The alternative, which so far has received little attention, is that the virus we know as ZEBOV has actually spread only recently to each outbreak site.
While the history of ZEBOV has so far remained elusive, examples from many other viruses show that the spatio-temporal dynamics of a virus are reflected in its phylogenetic structure [22–24]. Viruses that have long been maintained within a single host population, for example, tend to have a high diversity of genetic lineages, especially if they are subject to little or no selection at the host population level. Given sufficient levels of dispersal, related genotypes may be widely distributed and will show little spatial clustering. Hepatitis C virus, for instance, is considered to have a long association with humans, but has a number of strains with worldwide distribution, probably due to inadvertent infections resulting from medical interventions at a global scale [25].
In contrast, restricted genetic diversity and rapid turnover of genotypes are hallmarks of viruses that are either spreading or are subject to continuous positive selection. Fox rabies and influenza A are typical examples, respectively [22,23]. Although, in either case, phylogenies exhibit a characteristic ladder-like pattern, the underlying mechanism (genetic drift in the former, selection in the latter) is fundamentally different and should leave distinguishable signatures in the spatio-temporal distribution and genetic substitution patterns of the virus. Given these general principles, our aim for the present study was to use a combination of genetic, spatial, and temporal data to discriminate between the two hypotheses for ZEBOV emergence (i.e. long term, local persistence versus recent spread).
Genetic information for testing these hypotheses is available from gene sequences sampled from human outbreaks. If ZEBOV has been persistent at localities across the region for hundreds or thousands of years, then the virus should have diverged into a number of distinct genetic lineages whose most recent common ancestor (MRCA) long pre-dates the first recognized ZEBOV outbreak in 1976 at Yambuku, Democratic Republic of Congo (DRC). Furthermore, outbreaks subsequent to Yambuku could equally have been caused by more ancestral or more recently derived lineages. On the other hand, if ZEBOV has spread through the region only recently, all viruses sampled should be descendants of the same genetic lineage with an MRCA close to the 1976 sequence from Yambuku. Along a given spatial trajectory, genotypes involved in more recent outbreaks should be more or less direct descendants of viruses found during previous outbreaks (creating the characteristic ladder-like pattern). Successive outbreaks should also be progressively more divergent from the MRCA. The only reasonable scenario under which such a pattern would be expected from a long-resident virus is one of continuous selection for new virus variants. However, identical selection pressures would have to apply over much of the geographic range of the virus, and movement throughout the range would have to be high for the same selected variants to occur in different localities (e.g., influenza [23]).
A second source of information lies in the spatio-temporal pattern of ZEBOV emergence. Under either hypothesis, local transmission events during a given outbreak might result in new cases appearing further and further away from the outbreak origin. However, if outbreaks are truly independent emergences from a persistent, widely distributed ZEBOV population, no spatial trend should be apparent in the locations of different outbreaks over the entire period since 1976. In contrast, if ZEBOV has spread from a mid-1970s origin near Yambuku, then new outbreaks should move further and further away from Yambuku as time passes. If ZEBOV is transmitted through some sort of local contact process, then the rate of spread should be consistent across spatial scales (Protocol S1). If the spreading wave has made changes in direction, then outbreak date should be correlated with geographic distance along the invasion corridor rather than simply with straight-line distance from Yambuku.
A third class of information lies in the spatial structure of virus genotypes. Outbreaks at the front of a spreading wave should show a correlation between genetic and spatial distance that is detectable at different spatial scales (Protocol S1). Changes in the direction of spread might weaken such isolation by distance at large scales, but correlation strength would remain high if spatial distance was measured from the origin of the wave and along the putative path of spread. The development of strong spatial structuring would also be possible in a long-resident virus, but not necessarily, as high levels of gene flow would tend to spatially randomize genotypes [26]. Thus, the absence of spatial structuring would argue against spread, but not necessarily against local persistence.
We tested this series of predictions regarding local persistence or recent spread of ZEBOV by analyzing data on the spatio-temporal pattern of outbreaks together with glycoprotein (GP) gene sequences collected from human outbreaks. We found the data to be inconsistent with the idea that the ZEBOV outbreaks of the past 30 years are caused by a virus that has been a long-term resident at each site. Instead, all our results are concordant with the hypothesis of a recent ZEBOV wave that spread through the area in a relatively consistent and predictable manner.
Results
Phylogenetic Structure and Selection Patterns
Maximum likelihood (ML) and Bayesian phylogenetic estimation approaches produced highly similar phylogenetic trees in which only one major lineage could be distinguished (Figure 2). All of the major structural features showed high statistical support. Both approaches placed the earliest outbreak (Yambuku,1976) very near the tree root (which estimates the MRCA), implying that ZEBOV sequences obtained at all other localities evolved from a virus very similar to Yambuku sometime after 1976.
Figure 2 ML Tree of Full-Length (> 2,000 bp) ZEBOV-GP Sequences
Tree was found in Paup* and rooted in a separate analysis using ICEBOV as an out-group. The latter analysis excluded a 576-bp variable region for which alignment with ZEBOV was uncertain (see Materials and Methods). Numbers next to branches indicate percent support based on 1,000 bootstrap replicates and posterior probabilities obtained in a molecular clock-based analysis in program BEAST (only values > 70% are shown).
Both trees also exhibited a series of ancestor–descendant relationships between outbreak localities (i.e., Yambuku→Mayibout, Mayibout→Booué, Booué→Mendemba) that closely mirrored the time sequence of ZEBOV outbreaks, with the most recent outbreaks falling furthest from the tree root. The tendency for new outbreaks to be directly descendent from immediately preceding outbreaks implies that outbreaks have occurred only in newly infected areas: either at the front of a narrow, advancing wave or through a series of long jumps in which each outbreak was seeded by the previous one. The rate of nucleotide substitution thereby remains fairly constant through time, as a molecular clock model could not be rejected (Table S1).
Because the observed rapid turnover of viral genotypes could potentially be the consequence of selection for new variants, we tested for positive selection in two ways. First, we examined the ratio of non-synonymous to synonymous substitutions (dN/dS) along tree branches. A model distinguishing different dN/dS for internal and tip branches did not fit any better than a model with one ratio applied to the entire tree (p = 0.395; Table S2), indicating that there was no relative increase in dN associated with branches that gave rise to future lineages. Such an increase would be expected if positive selection were a major driving force behind the rapid turnover of virus genotypes [27]. But, given the small overall number of mutations on the tree, our statistical power to detect such an effect was low, especially if only a small number of sites were subject to positive selection.
In our second analysis, we tested whether individual sites showed evidence of selection. A model of nearly neutral evolution was rejected in favor of a model accounting for sites under positive selection (p = 0.009; Table S2). However, only one codon site (amino acid position 370) had a posterior probability greater than 95% of being under positive selection in a subsequent Bayesian assignment. This site appeared to have undergone three changes back and forth between isoleucine and methionine (Figure S1). A number of sites experienced a single amino acid change, many of which fell on internal branches. While these results are consistent with positive selection playing a role in ZEBOV evolution, they are inconclusive as to whether the specific amino acid replacements we observed are truly due to Darwinian selection or merely represent neutral evolution.
Spatio-Temporal Structure of Outbreaks
ZEBOV outbreaks showed a distinct spatio-temporal pattern, both over the entire period since 1976 and during shorter time intervals. For example, between 2001–2004 in the Gabon-Congo border area, both human outbreaks and animal carcasses that tested positive for Ebola independently showed statistically significant patterns of eastward spread (Figure 3A). Furthermore, the eastward spread rate estimated for the 2001–2004 period (46.1 km per year) changed little if the 1996 human outbreak at Booué and a nearby Ebola-positive chimpanzee carcass were added to the analysis (47.6 km per year; Figure 3B). This rate consistency over a large temporal interval is concordant with a single spread process, from Booué eastward through the Gabon-Congo border area.
Figure 3 Spatial Spread of ZEBOV
(A) Relationship between date and longitude of outbreaks in Gabon-Congo border area. Blue squares, human outbreaks [2] and 2003 outbreak at Mbandza village; red circles, animal carcasses testing positive for Ebola [18]; gray diamonds, ape die-off at Ibounji\Lokoue clearing. Regression line is for pooled data. Analyzed separately, human outbreaks and Ebola+ animal carcasses both show significant correlations between longitude and date (human outbreaks n = 12, R2 = 0.48, p = 0.01; animal carcasses n = 13, R2 = 0.91, p < 0.001).
(B) Added are 1996 human outbreak at Booué and Ebola+ chimpanzee carcass from nearby Lope [1]. The lack of reported human outbreaks between 1996 and 2001 may simply reflect the extremely low village density between Booué and Mendemba (Figure 1).
(C) Time after Yambuku versus straight line distance from Yambuku to all subsequent human outbreaks, including [2,14] and Mbandza village (R2 = 0.42, n = 17, p = 0.005).
(D) Same as (C) but with distance from Yambuku to the recent Gabon-Congo border outbreaks measured as passing through Booué (R2 = 0.97,n =17, p < 0.001). All figures include outbreak sites cited [2,14] for which no ZEBOV-GP sequences were publicly available.
The pattern of spread from west to east does not continue with the outbreaks preceding Booué. However, the ancestor–descendant relationships in the ZEBOV phylogeny (Figure 2) suggest a coherent spread pattern in the period preceding the Booué outbreak. The position of Yambuku near the tree root suggests that the ZEBOV spread originated somewhere near Yambuku in about 1976 and continued both south to Kikwit and west to Booué (Figure 1B). This hypothesis was supported by a ML search for the epizootic origin that maximized the correlation between geographic distance from the origin and time after the origin. This search chose a February 1973 origin just northwest of Yambuku (Figure 1B). The phylogenetic position of the Booué sequence as a direct ancestor to all of the 2001–2003 outbreaks (Figure 2) suggests that the western front then turned eastward toward the Gabon-Congo border. This abrupt change of direction may have been caused by natural features such as rivers, as frequently observed in spreading pathogens [28–30], and was suspected prior to any genetic data becoming available (see Discussion).
The hypothesis of a pivot point at Booué was strongly supported by the observed relationship between geographic distance from the putative origin at Yambuku and time after Yambuku. If geographic distances from all other outbreaks to Yambuku were measured in a straight line, then the relationship was significant but relatively weak (Figure 3C). However, if geographic distances to the Gabon-Congo border outbreaks were routed through Booué, correlation strength increased dramatically (Figure 3D). The great strength of this relationship was not due to a single outlier point, as all of the major legs of the putative epizootic path showed similar spread rates (Yambuku→Kikwit = 51.7 km per year, Yambuku→Mekouka = 56.9 km per year, Booué→Mendemba = 48.5 km per year, Mendemba→Iboundji = 47.9 km per year).
Spatial Structure of Genotypes
At the local as well as the regional scale, spatial structure was evident in the distribution of ZEBOV genotypes. For instance, the 2001–2003 outbreaks on the Gabon-Congo border showed a clear pattern of decreasing genetic similarity with increasing geographic distances (Figure 4A). The tight spatial structuring of genotypes at this relatively small scale fits the notion that ZEBOV transmission is a local contact process involving short movements of a few kilometers or less [26], a conclusion concordant with the observed consistency in the time rate of epizootic spread (Figure 3A and 3B).
Figure 4 Correlation between Geographic Distance and Patristic Genetic Distance
(A) ML genetic distances (substitutions per nucleotide site) plotted as function of geographic distance separating pairs of outbreak sites for the six full-length, georeferenced sequences sampled by Leroy et al. (R2 = 0.70, Mantel test p = 0.002). Makokou and Yembelengoye sequences excluded because of unknown spatial origin of case and partial sequence, respectively (Protocol S3).
(B) Correlation between straight line distance from the initial ZEBOV outbreak site at Yambuku and patristic genetic distance to Yambuku for all available georeferenced, full-length sequences (R2 = 0.38 , n = 11, p = 0.040).
(C) Same as (B) but with geographic distances to the recent Gabon-Congo border outbreaks measured as passing through Booué (R2 = 0.92 , n = 11, p < 0.001).
Geographic structuring of genotypes was also evident at higher spatial scales. The correlation between geographic distance to Yambuku and genetic divergence from Yambuku was only weak. However, as with the spatio-temporal analysis, a striking improvement in model fit was observed when geographic distances were routed through Booué (Figure 4B and 4C). A ML search for an epizootic pivot point that maximized the correlation between geographic distance and genetic divergences chose a pivot point just west of Booué (Figure 5).
Figure 5 Epizootic Pivot Point
Shading of each grid cell indicates the strength of correlation (R2) between geographic distance and patristic genetic distance when that grid cell (rather than Booué) is used as the epizootic pivot point. The position of the best fitting pivot point (shown with a white X) along the Ogooue River (blue line) is consistent with a river crossing near Booué, with subsequent movement east toward the Mendemba area. It is not consistent with gene flow directly between the other mid-1990s outbreak localities (Mekouka and Mayibout) and Mendemba.
Discussion
Our results clearly challenge the belief that ZEBOV has been persistently present for a long time at the outbreak sites in Gabon and Congo. First, our phylogenetic results imply that all known ZEBOV emergences occurring after Yambuku in 1976 were caused by direct and closely related descendents of a Yambuku-like virus. The descent of all known ZEBOV viruses from a very recent common ancestor is clearly inconsistent with the notion that they have long been evolving independently and in situ.
Second, a similar ancestor–descendent relationship connects the outbreaks of the mid-1990s to those of 2001–2004 (Figure 2). This replacement of virus over time and space by closely related but progressively more divergent genotypes is typically observed under spatial spread or continuous positive selection. Although our analyses were consistent with some codon sites being under positive selection, statistical power was too weak to reliably identify such sites. Fit of the molecular clock suggests that many of the observed substitutions are effectively neutral, so that the number of positively selected sites may be small. More importantly, spread and adaptation are not mutually exclusive. In fact, the very recent descent of all ZEBOV variants from a Yambuku-like common ancestor necessarily implies relatively rapid spread of variants across a large range. Thus, whether or not the ladder-like structure of the ZEBOV tree bears a signature of positive selection, it is much more consistent with recent spread than with independent evolution at each outbreak locality.
Third, we found a general correlation between when new ZEBOV cases were observed and their geographical distance to previous cases. Importantly, this relationship and the corresponding rate of spread of about 50 km per year remained consistent over multiple spatial scales. The low p values in our correlation analyses indicate that observing such a pattern by chance, as the hypothesis of long-term presence of ZEBOV in the area would require, would be highly unlikely. A recent ZEBOV outbreak in May 2005 at Etoumbi village, which occurred after this paper was submitted for review, further provided an opportunity to test our model. Reassuringly, the spread rate from 2001–2003 did an excellent job of predicting the Etoumbi outbreak, given its distance from the 2001 outbreak at Mendemba (Figure 6).
Figure 6 Spatial Spread from 2001–2005
Distance of each human outbreak site from the initial outbreak at Mendemba village, Gabon, plotted as a function of time after the Mendemba outbreak. Dashed regression line uses only outbreaks from 2001–2003 (R2 = 0.43, p = 0.04). Solid regression line includes May 2005 outbreak at Etoumbi village (R2 = 0.73, p < 0.001).
Fourth, we identified a pattern of constantly increasing genetic divergence among virus genotypes with increasing geographic distance. As pointed out initially, a roughly linear relationship between genetic divergence and geographic separation is an expected outcome under spatial spread, particularly if spread has occurred along a relatively narrow front (Protocol S1). Although isolation by distance itself could also be found among locally resident viruses, such a scenario would be inconsistent with any of the previous results, including the dramatically improved fit of the spatial-genetic correlation when routing distances through Booué.
Taken together, our results clearly point to the conclusion that ZEBOV has gradually spread across central Africa from an origin near Yambuku in the mid-1970s. Under this scenario, the distinct phylogenetic tree structure, the strong correlation between outbreak date and distance from Yambuku, and the correlation between genetic and geographic distances can be interpreted as the outcome of a consistently moving wave of ZEBOV infection.
The large-scale spatial correlations we identified were particularly strong under the assumption that the ZEBOV wave changed direction at Booué. This hypothesis may seem ad hoc but was actually posed by one of the authors (PDW) in a paper published a year before genetic data from the Gabon-Congo border region became available [4]. Transect surveys and numerous reports from local villagers had suggested that the second largest river system in equatorial Africa (the Ogooue-Ivindo-Ayina) had largely contained the 1994–1996 outbreaks in the Minkebe region of northern Gabon [3,4]. An Ebola-positive chimpanzee was then found south of the river near Booué in 1996 [1], and subsequent surveys revealed suspiciously low ape densities southeast of Booué. Thus, all of the genetic correlation analyses we report here represent independent confirmation of an a priori hypothesis of spread from Booué, posed before genetic data were available. Likewise, the phylogenetic analysis, which identified the Booué virus sequence as the direct ancestor of all viruses observed later near the Gabon-Congo border, also indicates that Booué forms an epidemiologic link between previous and subsequent outbreaks. The effect of major rivers in channeling spread is well documented for other diseases in natural populations [28–30].
Whether ZEBOV was resident (but undetected) in the central African forest block before the mid-1970s, or is an invader from outside the region remains unclear. Blood samples taken from both human [11] and non-human primates [17] suggest that some filovirus was already present in western equatorial Africa before the mid-1990s ape die-offs. Unfortunately, the serological tests employed were not specific to ZEBOV [31]. Therefore, it is impossible to tell whether these positive results were caused by a virus with a very recent common ancestor of the lineage we know as ZEBOV or by some more distantly related virus that is cross-reactive. The co-occurrence of both moderately high seropositivity and high ape densities at some sampling localities argues that the assayed virus was not highly virulent. ZEBOV has caused such high mortality rates in recent ape outbreaks that by the time these populations recover to high density (if they recover), individuals born after the outbreaks will greatly outnumber seropositive survivors (if any are still alive). Thus, moderate to high levels of seropositivity in ape populations are not consistent with high virulence. The absence of large human outbreaks in western equatorial Africa before the mid-1990s is consistent with a non-virulent virus, although the possibility that smaller outbreaks occurred but were not recognized cannot be excluded.
The high rate of positive results in past serological surveys may explain why previous authors appear not to have seriously considered the possibility of recent ZEBOV spread. Apart from the serological results, the other major argument for long-term persistence at each locality has involved the mutational stability of ZEBOV-GP. The absence of mutations within several closely monitored human transmission chains has been used to argue that ZEBOV-GP evolves too slowly for a wildlife epizootic lasting only a few years to have generated the sequence variation observed in the recent Gabon-Congo border outbreaks [2,18]. However, a formal statistical power analysis shows that the number of human cases involved in the cited transmission chains was far too small to reach this conclusion (Protocol S2). In fact, our molecular clock analyses showed that ZEBOV evolves at a rate comparable to other RNA viruses, about 8 × 10−4 substitutions per site per year (Table S3) and that the MRCA of the Gabon-Congo border outbreaks occurred in 1999 (CI = 1998–2000), well after the 1996 Booué outbreak. Thus, the genetic stability noted between ZEBOV outbreaks appears to be the consequence of short time separation rather than slow evolution.
Although our results strongly support the hypothesis that ZEBOV spread recently to the outbreak sites in Gabon and Congo, it is still unclear through which reservoir host(s) ZEBOV spread occurred. Spread might have taken place through transmission within some wildlife reservoir endemic to the region or through the wave-like invasion of an infected reservoir. Whatever the reservoir species or group of species, the striking constancy in the rates of ZEBOV spread and evolution suggests either that its distribution and abundance are fairly uniform throughout the affected area, or that its range has been expanding at a uniform rate. At the same time, we found that the large-scale pattern of spread is well represented as one-dimensional, which contradicts expectations for a radiating wave in a uniformly distributed host. As pointed out before, we suspect that the channeling effect of rivers may be responsible for this pattern. The large time gaps between human outbreaks may simply reflect the fact that much of the proposed epizootic path is very lightly inhabited by humans. For example, the 1996 outbreak site at Booué and the 2001 outbreak site at Mendemba are separated by 250 km of forest crossed by a single road, along which lie only a handful of small villages.
The conclusion that ZEBOV has recently spread further begs the question of whether this spread was triggered by some ecological change (perhaps anthropogenic in origin), by some change in the virus itself (for example, a mutation to higher virulence), or simply by some stochastic event. Answering these questions remains a challenge for future research. To what extent ZEBOV transmission between apes plays a role in either ZEBOV spatial spread or ape die-offs also remain open questions.
Our results also warrant a re-evaluation of the potential for Ebola control. The consistent rate of Ebola spread suggests that control efforts may not need to encompass the entire region, but could be concentrated directly ahead of the advancing wave. Knowledge of the future path of spread could be used to strategically allocate the delivery of an Ebola vaccine [32–34] (cf. rabies [35]) when a successful vaccine is developed.
If the past spread rate of about 50 km per year continues in the current direction, Ebola Zaire should hit the populated areas north and east of Odzala National Park within the next one to two years. Most of the handful of parks still containing populations of gorillas large enough to be viable in the long term might be reached within three to six years. Saving these viable ape populations should be a top priority.
Materials and Methods
Phylogenetic estimation
Our phylogenetic analyses included 13 of the 14 published ZEBOV-GP gene sequences, including sequences from the Gabon-Congo border outbreaks of 2001–2003. A. Sanchez (Centers for Disease Control and Prevention) kindly provided us with a sequence from the 1996 outbreak at Mayibout, Gabon. As the out-group for our analyses, we used the sequence from the 1994 Ebola outbreak at Tai Forest, Cote d'Ivoire (ICEBOV, the strain genetically closest to ZEBOV).
Much of the information for estimating time rates of divergence from the ZEBOV root lies in a hyper-variable region in the middle of GP [36]. This region is highly divergent from the closest out-group, ICEBOV [37], causing problems of saturation and precluding accurate alignments even at the amino acid level. Therefore, we initially trimmed the hyper-variable region (sites 925–1,500) and used ICEBOV as the out-group to estimate the topological root of ZEBOV based on the remaining 1,473 base pairs (bp). A ML tree using all available ZEBOV sequences greater than 2,000 bp was found in a heuristic search under a GTR+G model in Paup*4.0b10 [38]. Keeping the resulting tree root, we then excluded ICEBOV and used the full-length sequences (2,049 bp) to re-estimate tree topology (which remained virtually the same) and branch lengths for ZEBOV only. Selection of the evolutionary models and model parameters was done in Modeltest 3.6 [39] using recommendations by Posada and Buckley [40]. In parallel, we used the dated full-length sequences in a Bayesian coalescent analysis that simultaneously estimated tree topology and the rate of clocklike divergence from the root but did not require an out-group (see next paragraph).
We estimated the evolutionary rate of ZEBOV by taking advantage of the temporal spread of sequence sampling, spanning almost three decades. Based on the rooted ML tree obtained, program TipDate [41] was used to estimate the ML rate of evolution under the “single rate dated tips” model. Fit of the “single rate dated tips” model was assessed against a model with unconstrained branch lengths (i.e., different rates for each branch) and a single rate model that did not take sampling dates into account. In addition to the ML estimates, Bayesian coalescent-based estimates of the evolutionary rate and tree topology of ZEBOV under a “single rate dated tips” model were obtained in the program BEAST [42]. Two independent runs were performed with 20 million states each, of which the first 2 million were removed as burn-in. Along with the two parameters of interest (rate and topology), the program yielded estimates of the transition/transversion ratio and proportion of invariant sites. From the states visited, 9,000 trees (one every 4,000 states) were used to compute posterior probability frequencies of support for individual nodes. Besides strengthening support for most of the nodes identified in the ML analysis, this analysis independently placed the Yambuku sequences at the root of the ZEBOV tree. Varying prior assumptions about the demographic history of ZEBOV (exponential growth and decline instead of constant population size) had virtually no effect on the results (unpublished data). Further information on methods and our choice of sequences can be found in Protocol S3.
Testing for positive selection
Only 29 inferred dS and 39 inferred dN were found in the ZEBOV-GP phylogeny. Although the small number of changes clearly limited our statistical power, we attempted to characterize selection patterns for the GP gene using the program CodeML [43]. First, we tested for evidence of positive selection on particular branches. If replacement of viral lineages over space and time were due to positive selection, non-synonymous changes should be found particularly on the internal branches of the tree because these branches represent lineages that have spread successfully, potentially due to a fitness advantage [44,45]. By the same argument, relative rates of dN to dS should be higher in the case of positive selection for internal branches compared with external branches, which left no descendents within our sample. Therefore, we estimated different dN/dS ratios for internal branches and branches connected to tree tips and compared this with a model with a single ratio for all branches. Secondly, we tested for the presence of particular codon sites under positive selection [46], followed by the identification of such sites using an empirical Bayes procedure [47]. For all tests, relative fit of nested models was determined by using likelihood ratio testing. Finally, we used ancestral reconstruction to map putative amino acid substitutions onto the phylogeny.
Estimating epizootic origin
To find a putative epizootic origin, we posited that an epizootic wave started somewhere in central Africa sometime before the 1976 Yambuku outbreak, spread outward at a constant rate, then made a change of direction at Booué. Under these assumptions, the spatial distance (di) of outbreak location i from a putative point of origin should be a linear function of time (ti) since the outbreak at the putative origin
with distances to the Gabon-Congo border outbreaks measured through Booué. We estimated the origin by searching for a date, location, and value of the slope parameter a that minimized the squared differences between the predicted distance and the observed distance summed over all outbreak locations (including Yambuku). The best fitting date and location was in the forest–savannah transition zone just to the north and west of Yambuku (4.0°N, 21.6°E) in February 1973 (Figure 1B). The coefficient of determination for this best fitting origin (R2 = 0.99; n = 18; p < 0.001) was only marginally better than the coefficient of determination achieved using Yambuku as the origin (Figure 3C).
Estimating epizootic pivot point
To search for the best fitting pivot point for the epizootic, we performed a Pearson product-moment correlation between the geographic distance of each outbreak site along the putative epizootic path from Yambuku and the patristic genetic distance of that outbreak site from Yambuku. However, instead of routing geographic distances from Yambuku to the recent Mendemba series of outbreaks through Booué (as in our other analyses), we cut the region into a 0.1° (about 11 km) grid and routed distances through the midpoint of each cell on the grid. The cell that produced the strongest correlation between geographic and genetic distance from Yambuku (plotted in Figure 5 as the coefficient of determination, R2) was taken to be the ML pivot point. We excluded Kikwit from this analysis because our phylogenetic analyses suggested that it was a distinct lineage that had diverged from the Gabon-Congo lineage at some unknown location south and west of their common ancestor Yambuku. Because we had no a priori hypothesis about where in space this divergence occurred, we had no means of calculating the spatial distances between the Gabon-Congo outbreaks and Kikwit.
Supporting Information
Protocol S1 Simulating Spatial Spread
(333 KB PDF).
Click here for additional data file.
Protocol S2 Statistical Power Analysis
(173 KB PDF).
Click here for additional data file.
Protocol S3 Phylogenetic Estimation
(228 KB PDF).
Click here for additional data file.
Figure S1 Selection Analysis
Distribution of inferred amino acid changes on the ZEBOV phylogeny. Substitutions were estimated using ancestral reconstruction in program CodeML. Codon site 370, the only site identified as being under positive selection, is shown in blue.
(123 KB DOC)
Click here for additional data file.
Table S1 Support For the Molecular Clock in ZEBOV
(100 KB PDF).
Click here for additional data file.
Table S2 Assessing Fit of Selection Models [46] to the ZEBOV Glycoprotein Data Using Likelihood Ratio Testing
(100 KB PDF).
Click here for additional data file.
Table S3 Evolutionary Rate Estimates for ZEBOV Obtained under ML in TipDate and under a Bayesian Framework Using Markov Chain Monte Carlo Integration in BEAST
(89 KB PDF).
Click here for additional data file.
Accession Numbers
The GenBank (http://www.ncbi.nlm.nih.gov/GenBank) accession numbers for the gene sequences discussed in this paper are: the Booué Gabon outbreak of 1996 (AY058898); the Gabon-Congo border outbreaks of 2001–2003 (Ektakangaye, AY526100; Entsiami, AY526102; Makoukou, AY526101; Mendemba A, AY526105; Mvoula, Ay526104; Olloba, AY526099); the Kikwit, DRC outbreak of 1995 (U28077.1); the Mekouka Gabon outbreak of 1994 (U77384.1); the Tai Forest, Cote d' Ivoire outbreak of 1996 (U28006); and the Yambuku, DRC outbreak of 1976 (Eckron 76 [Yambuku-E], U81161.1; Mayinga [Yambuka-M], U231887.1).
We thank C. Henderson and J. Snaman for help with preparation of phylogenetic trees. J. Chave, J. Duschoff, D. Purves, and L. Waller made useful comments on statistical analyses. We thank Stuart Nichol and Pierre Rollin for their comments on the manuscript and helpful discussion of the ideas presented here. We thank B. Karesh for initially putting the authors in contact. This research was supported by the National Science Foundation (DEB 0213001 to PDW) and the National Institutes of Health (RO1 AI047498 to LAR).
Competing interests. The authors have declared that no competing interests exist.
Author contributions. PDW, RB, and LAR analyzed the data and wrote the paper.
Citation: Walsh PD, Biek R, Real LA (2005) Wave-like spread of Ebola Zaire. PLoS Biol 3(11): e371.
Abbreviations
bpbase pair
dNnon-synonymous substitution
dSsynonymous substitution
GPglycoprotein
ICEBOVEbola Cote D' Ivoire
MLmaximum likelihood
MRCAmost recent common ancestor
ZEBOVZaire strain of Ebola virus
==== Refs
References
Georges AJ Leroy EM Renaut AA Benissan CT Nabias RJ Ebola hemorrhagic fever outbreaks in Gabon, 1994–1997: Epidemiologic and health control issues J Infect Dis 1999 179 S65 S75 9988167
Leroy EM Rouquet P Formenty P Souquiere S Kilbourne A Multiple Ebola virus transmission events and rapid decline of central African wildlife Science 2004 303 387 390 14726594
Huijbregts B De Wachter P Obiang LSN Akou ME Ebola and the decline of gorilla Gorilla gorilla and chimpanzee Pan troglodytes populations in Minkebe Forest, north-eastern Gabon Oryx 2003 37 437 443
Walsh PD Abernethy KA Bermejo M Beyers R De Wachter P Catastrophic ape decline in western equatorial Africa Nature 2003 422 611 614 12679788
Tutin CEG Fernandez M Nationwide census of gorilla (Gorilla gorilla) and chimpanzee (Pan troglodytes) populations in Gabon Am J Primatol 1984 6 313 336
Carroll RW Relative density, range extension, and conservation potential of the lowland gorilla (Gorilla gorilla) in the Dzanga-Sangha region of southwestern Central African Republic Mammalia 1988 52 309 323
Fay M Agnagna M Census of gorillas in northern Republic of Congo Am J Primatol 1992 27 275 284
Stromayer AK Ekobo A The distribution and number of forest dwelling elephants in extreme southeastern Cameroon Pachyderm 1992 15 9 14
Williamson E Usongo L Gorilla survey in the Dja Reserve, Cameroun Gor Conserv News 1996 10 11 14
Bermejo M Status and conservation of primates in Odzala National Park, Democratic Republic of Congo Oryx 1999 33 323 331
Monath TP Ecology of Marburg and Ebola viruses: Speculations and directions for future research J Infect Dis 1999 179 S127 S138 9988176
Rodriguez LL De Roo A Guimard Y Trappier SG Sanchez A Persistence and genetic stability of Ebola virus during the outbreak in Kikwit, Democratic Republic of Congo, 1995 J Infect Dis 1999 179 S170 176 9988181
Peterson AT Bauer JT Mills JN Ecologic and geographic distribution of filovirus disease Emerg Infect Dis 2004 10 40 47 15078595
Peterson AT Carroll DS Mills JN Johnson KM Potential mammalian filovirus reservoirs Emerg Infect Dis 2004 10 2073 2081 15663841
Morvan JM Deubel V Gounon P Nakoune E Barriere P Identification of Ebola virus sequences present as RNA or DNA in organs of terrestrial small mammals of the Central African Republic Microbes Infect 1999 1 1193 1201 10580275
Gonzalez JP Nakoune E Slenczka W Vidal P Morvan JM Ebola and Marburg virus antibody prevalence in selected populations of the Central African Republic Microbes Infect 2000 2 39 44 10717539
Leroy EM Telfer P Kumulungui B Yaba P Rouquet P A serological survey of Ebola virus infection in central African nonhuman primates J Infect Dis 2004 190 1895 1899 15529251
Rouquet P Froment JM Bermejo M Kilbourn A Karesh W Wild animal mortality monitoring and human Ebola outbreaks, Gabon and Democratic Republic of Congo, 2001–2003 Emerg Infect Dis 2005 11 283 290 15752448
Allela L Bourry O Pouillot R Delicat A Yaba P Ebola virus antibody prevalence in dogs and human risk Emerg Infect Dis 2005 11 385 390 15757552
Tucker CJ Wilson JM Mahoney R Anyamba A Linthicum K Climatic and ecological context of the 1994–1996 Ebola outbreaks Photogramm Eng Rem S 2002 68 147 152
Pinzon JE Wilson JM Tucker CJ Arthur R Jahrling PB Trigger events: Enviroclimatic coupling of Ebola hemorrhagic fever outbreaks Am J Trop Med Hyg 2004 71 664 674 15569802
Holmes EC The phylogeography of human viruses Mol Ecol 2004 13 745 756 15012753
Grenfell BT Pybus OG Gog JR Wood JL Daly JM Unifying the epidemiological and evolutionary dynamics of pathogens Science 2004 303 327 332 14726583
Real LA Henderson JC Biek R Snaman J Lambert JT Unifying the spatial population dynamics and molecular evolution of epidemic rabies virus Proc Natl Acad Sci U S A 2005 102 12107 12111 16103358
Smith DB Pathirana S Davidson F Lawlor E Power J The origin of hepatitis C virus genotypes J Gen Virol 1997 78 321 328 9018053
Epperson BK Geographical genetics 2003 Princeton (New Jersey) Princeton University Press 376
Yang Z Bielawski JP Statistical methods for detecting molecular adaptation Trends Ecol Evol 2000 15 496 503 11114436
Bourhy H Kissi B Audry L Smreczak M Sadkowska-Todys M Ecology and evolution of rabies virus in Europe J Gen Virol 1999 80 2545 2557 10573146
Mondet B Yellow fever epidemiology in Brazil—New considerations Bull Soc Pathol Exot 2001 94 260 267 11681224
Smith DL Lucey B Waller LA Childs JE Real LA Predicting the spatial dynamics of rabies epidemics on heterogeneous landscapes Proc Natl Acad Sci U S A 2002 99 3668 3672 11904426
McCormick JB Ebola virus ecology J Infect Dis 2004 190 1893 1894 15529250
Sullivan NJ Geisbert TW Geisbert JB Xu L Yang ZY Accelerated vaccination for Ebola virus haemorrhagic fever in nonhuman primates Nature 2003 424 681 684 12904795
Garbutt M Liebscher R Wahl-Jensen V Jones S Moller P Properties of replication-competent vesicular stomatitis virus vectors expressing glycoproteins of filoviruses and arenaviruses J Virol 2004 78 5458 5465 15113924
Jones SM Feldmann H Stroher U Geisbert JB Fernando L Live attenuated recombinant vaccine protects nonhuman primates against Ebola and Marburg viruses Nat Med 2005 11 786 790 15937495
Brochier B Costy F Pastoret PP Elimination of fox rabies from Belgium using a recombinant vaccinia-rabies vaccine—An update Vet Microbiol 1995 46 269 279 8545965
Sanchez A Trappier SG Mahy BWJ Peters CJ Nichol S The virion glycoproteins of Ebola viruses are encoded in two reading frames and expressed through transcriptional editing Proc Natl Acad Sci U S A 1996 93 3602 3607 8622982
Suzuki Y Gojobori T The origin and evolution of Ebola and Marburg viruses Mol Biol Evol 1997 14 800 806 9254917
Swofford DL PAUP: Phylogenetic analysis using parsimony, version 4 [computer program] 2002 Sunderland (Massachusetts) Sinauer Associates
Posada D Crandall KA MODELTEST: Testing the model of DNA substitution Bioinformatics 1998 14 817 818 9918953
Posada D Buckley T Model selection and model averaging in phylogenetics: Advantages of Akaike information criterion and Bayesian approaches over likelihood ratio tests Syst Biol 2004 53 793 808 15545256
Rambaut A Estimating the rate of molecular evolution: Incorporating non-contemporaneous sequences into maximum likelihood phylogenies Bioinformatics 2000 16 395 399 10869038
Drummond AJ Rambaut A BEAST, version 1.03 [computer program] 2003 Available: http://evolve.zoo.ox.ac.uk/beast/ . Accessed 16 September 2005
Yang Z PAML: A program package for phylogenetic analysis by maximum likelihood Comput Appl Biosci 1997 13 555 556 9367129
Sheridan I Pybus OG Holmes EC Klenerman P High-resolution analysis of hepatitis C virus adaptation and its relationship to disease progression J Virol 2004 78 3447 3454 15016867
Shackelton LA Parrish CR Truyen U Holmes EC High rate of viral evolution associated with the emergence of carnivore parvovirus Proc Natl Acad Sci U S A 2005 102 379 384 15626758
Yang Z Nielsen R Goldman N Pedersen AM Codon-substitution models for heterogeneous selection pressure at amino acid sites Genetics 2000 155 431 449 10790415
Yang ZH Wong WSW Nielsen R Bayes empirical Bayes inference of amino acid sites under positive selection Mol Biol Evol 2005 22 1107 1118 15689528
|
16231972
|
PMC1262627
|
CC BY
|
2021-01-05 08:21:51
|
no
|
PLoS Biol. 2005 Nov 25; 3(11):e371
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030371
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 1623197410.1371/journal.pbio.0030374Research ArticleCell BiologyMolecular Biology/Structural BiologyEukaryotesDifferential Recruitment of Pre-mRNA Splicing Factors to Alternatively Spliced Transcripts In Vivo Recruitment of Splicing FactorsMabon Stephen A
1
Misteli Tom [email protected]
1
1National Cancer Institute, National Institutes of Health, Bethesda, Maryland, United States of AmericaWickens Marv Academic EditorUniversity of WisconsinUnited States of America11 2005 25 10 2005 25 10 2005 3 11 e37420 1 2005 7 9 2005 Copyright: © 2005 Mabon and Misteli.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Alternative mRNA Splicing: Control by Combination
Alternative splicing in mammalian cells has been suggested to be largely controlled by combinatorial binding of basal splicing factors to pre-mRNA templates. This model predicts that distinct sets of pre-mRNA splicing factors are associated with alternatively spliced transcripts. However, no experimental evidence for differential recruitment of splicing factors to transcripts with distinct splicing fates is available. Here we have used quantitative single-cell imaging to test this key prediction in vivo. We show that distinct combinations of splicing factors are recruited to sites of alternatively spliced transcripts in intact cells. While a subset of serine/arginine protein splicing factors, including SF2/ASF, SC35, and SRp20, is efficiently recruited to the tau gene when exon 10 is included, these factors are less frequently associated with tau transcription sites when exon 10 is excluded. In contrast, the frequency of recruitment of several other splicing factors is independent of splicing outcome. Mutation analysis of SF2/ASF shows that both protein–protein as well as protein–RNA interactions are required for differential recruitment. The differential behavior of the various splicing factors provides the basis for combinatorial occupancy at pre-mRNAs. These observations represent the first in vivo evidence for differential association of pre-mRNA splicing factors with alternatively spliced transcripts. They confirm a key prediction of a stochastic model of alternative splicing, in which distinct combinatorial sets of generic pre-mRNA splicing factors contribute to splicing outcome.
Quantitative single-cell imaging reveals distinct combinations of splicing factors recruited to sites of alternatively spliced transcripts in intact cells.
==== Body
Introduction
Pre-mRNA splicing is one of the key steps in the maturation of newly synthesized transcripts. During the splicing process, non-coding intron sequences are removed and the coding exon sequences are joined in a multi-step reaction, carried out by the spliceosome, a large protein complex consisting of small ribonucleoprotein particles (snRNP), and associated non-snRNP proteins [1,2]. The major class of non-snRNPs are the serine-arginine rich SR-proteins, characterized by the presence of one or two RNA-binding domains and a C-terminal SR-rich domain (RS-domain)[3,4]. The spliceosome carries out the splicing reaction in a series of complex spatial and temporal conformational rearrangements leading to the alignment of the reaction substrates and formation of the spliced mRNA [1,2].
While many introns are constitutively excised during splicing, other introns are subject to alternative splicing. During the alternative splicing reaction multiple mRNA species are generated by combinatorial inclusion or exclusion of exons [5–9]. Alternative splicing is rapidly emerging as a ubiquitous cellular event with estimates suggesting that 60%–90% of human genes are alternatively spliced [10]. Alternative splicing thus contributes significantly to protein diversity beyond that encoded in the genome sequence. Defects in alternative splicing and generation of aberrant ratios of multiple mRNA isoforms from a single gene are now also recognized as major contributors to human disease [11–13]. Amongst many others, alternative splicing defects have been identified as the cause of numerous diseases including β-thalassemia, cystic fibrosis, and premature aging [11].
The molecular basis for how alternative splice sites are selected is still largely unclear [7]. Although a handful of, often tissue-specific, dedicated splice site selection factors have been identified [13–16], the predominant model for splice site choice envisions that selective usage of an exon is due to the differential association of distinct sets of generic splicing factors [8,17]. This model is indirectly supported by the observation that in vivo and in vitro titration of multiple factors against each other modulates alternative splice site switching, presumably by antagonistic competition for splice sites [18–21]. A key prediction from this model is that alternatively spliced substrates should have different sets of splicing factors associated with them. Biochemical approaches to isolate spliceosomes with differential splicing factor composition have so far been unsuccessful, likely due to the heterogeneity of in vitro reactions and/or the transient nature of the association of splicing factors with their templates [2].
To circumvent these methodological limitations, we applied quantitative single cell microscopy to ask whether differential sets of splicing factors associate with alternatively spliced transcripts. This approach is based on the fact that pre-mRNA splicing factors are non-randomly distributed within the mammalian cell nucleus. The vast majority of pre-mRNA splicing components are enriched in distinct nuclear sub-compartments termed speckles or splicing factor compartments [22,23]. Despite the local concentration of splicing factors in these nuclear domains, they are most likely not active sites of pre-mRNA splicing but represent storage/assembly sites for splicing components [24–26]. Activation of genes near splicing factor compartments results in the physical recruitment of splicing factors from speckles and their accumulation at sites of active transcription and splicing [24]. This accumulation can be detected by quantitative fluorescence microscopy and used as an in vivo assay for recruitment of factors to pre-mRNA [26–28]. The observed recruitment is dependent on the functional interaction of splicing factors with the pre-mRNA since transcripts with no introns or mutant pre-mRNAs that cannot be spliced do not recruit splicing factors [28]. In addition, mutant splicing factors do not accumulate efficiently at transcription sites [27]. Using this quantitative in vivo single cell microscopy assay, we here provide evidence for differential recruitment of several splicing factors to alternatively spliced transcripts. Our findings confirm a key prediction of a stochastic, combinatorial recruitment model for alternative splicing.
Results
An In Vivo Assay for Recruitment of Splicing Factors to Alternatively Spliced Transcripts
We sought to establish an experimental system to study the recruitment of splicing factors to alternatively spliced transcripts in vivo. As a model system we used the tau gene, which encodes a microtubule binding protein and whose aberrant alternative splicing is the cause of frontotemporal dementia and parkinsonism linked to chromosome 17 (FTDP-17), a parkinsonism-like neurological disorder [29]. In healthy individuals, exon 10 of the tau gene is included or excluded from the mRNA with roughly equal probability during pre-mRNA splicing [29]. In contrast, among patients with FTDP-17 mutations near the 5′ splice site of exon 10 result in predominant inclusion of exon 10 [29–31]. A minigene system consisting of tau exon 10 flanked by partial tau intron sequences and HIV TAT exons has previously been used to mimic the alternative splicing behavior of the full-length wild-type (wt) and mutant tau gene in COS-7 cells (Figure 1A) [30,31].
Figure 1 Characterization of Stable Cell Lines Expressing Alternatively Spliced tau Minigenes
(A) Schematic representation of the tau exon 10 minigene system. Alternative inclusion or exclusion of exon 10 as observed for endogenous tau is recapitulated in a minigene containing flanking HIV-TAT exons.
(B) Preferential inclusion of tau exon 10 detected by gel electrophoresis. The wttau exon 10 (tau10wt) is evenly included or excluded from the mRNA, whereas a tau exon 10 containing a T>C mutation at position −1 of exon 10 (tau10−1) is predominantly included in the same context. Exon 10 inclusion is detected as a 246-bp by RT-PCR using primers (arrows) in the TAT exons. Exon 10 exclusion is detected by a 153-bp band.
(C) Quantitative real-time PCR analysis to determine exon 10 inclusion. Results are averages ± standard deviations from at least three experiments.
(D) RNA-FISH using nick-translated probes against the tau minigene to detect sites of tau transcription. The vast majority of cells in a stable cell population contained one or two tau minigene insertion sites. Arrowheads indicate tau transcription sites. Scale bar = 3.5 μm.
We used this well-characterized minigene to establish an experimental system to study potential differences in recruitment of splicing factors to alternatively spliced transcripts in vivo. To this end, monoclonal stable COS-1 cell lines expressing either tau10wt or tau10−1 were generated (see Materials and Methods for details). In COS-1 cells stably expressing tau10wt, exon 10 was predominantly excluded, whereas in tau10−1 stable cell lines, exon 10 was predominantly included (Figure 1B). This pattern is identical to that observed for endogenous tau in patients and in the transiently expressed minigene system [29,30]. Quantitation by real-time PCR analysis demonstrated a ∼ 5-fold increase in exon 10-containing mRNA in tau10−1 cells compared to tau10wt cells (Figure 1C). The transcription site of the stably integrated minigene was readily detectable by RNA fluorescence in situ hybridization (FISH) using nick-translated probes to full-length tau mRNA (Figure 1D). Cell clones for both tau10wt and tau10−1 were isolated containing the transgene in a single genomic integration site (Figure 1D).
Differential Recruitment of SR-Proteins to Alternatively Spliced Transcripts
We used this experimental system to ask whether splicing factors associate differentially with alternatively spliced transcripts in vivo. To this end we took advantage of a previously characterized quantitative single cells splicing factor recruitment assay [27]. In this assay recruitment of a splicing factor to a specific transcription site is detected by simultaneous FISH to visualize the site of transcription and indirect immunofluorescence (IF) using specific antibodies to detect endogenous pre-mRNA splicing factors. As previously described, positive recruitment of splicing factors is defined as > 2-fold accumulation of splicing factor signal at the site of transcription above the global nucleoplasmic signal as determined by quantitative linescan analysis of single transcription sites (Figure 2A; see Materials and Methods for details) [27]. The obligatory U2 associated factor U2-B′′ served as a positive control and to determine the detection sensitivity of the recruitment assay. As expected U2-B′′ was detected at sites of transcription with high frequency both in cells expressing tau exon 10wt or exon 10−1 (Figures 2A and 2 B). Recruitment was detected at the site of tau transcription in 78 ± 6% of exon 10wt cells and 77 ± 14% of exon 10−1 cells (Figure 2A and 2B). Somewhat lower accumulations of typically between 1.3–1.9 fold over the nucleoplasmic signal were found in most other cells, and no tau transcription sites were detected that were entirely devoid of U2-B” signal (unpublished data). In contrast to the obligatory splicing factor U2-B”, the SR protein SF2/ASF showed differential recruitment between tau10wt and tau10−1 cells (Figure 2A and 2B). Endogenous SF2/ASF was recruited to transcription sites in more than 68 ± 6% of tau10−1 cells, but accumulated at transcription sites in only 31 ± 2% of tau10wt cells (Figure 2B). This difference in accumulation was significant at the p < 0.005 level in a Student t-test. The SR protein SC35 behaved differently from SF2/ASF in that it was only poorly recruited to either transcript (Figures 2A and 2B). SC35 was recruited to tau transcription sites in 38 ± 13% of tau10wt cells, and to 48 ± 2% of sites in tau10−1 cells (Figure 2B). SRp40 in contrast was recruited to the tau transcript in more than 80% of cells, regardless of the nature of the transcript (Figure 2B). Similar observations were made in several clonal cell lines (unpublished data). The observations that SF2/ASF was differentially recruited to the two tau transcripts, while SC35 was recruited to neither and SRp40 was recruited to both indicate that distinct SR proteins are differentially recruited to sites of alternatively spliced transcripts in vivo.
Figure 2 Differential Recruitment of Endogenous Pre-RNA Splicing Factors to Alternatively Spliced Transcripts
(A) Recruitment of splicing factors to tau transcription sites detected by combined RNA-FISH using specific probes against the tau minigene (green) and IF microscopy with specific anti-splicing factor antibodies (red). Arrowheads indicate tau transcription sites. All associations were confirmed by linescan analysis. Scale bar = 2.5 μm. (Inset: higher magnifications of the transcription site) Lines indicate the location of the linescans.
(B) Quantitation of percentage of cells with colocalization of tau RNA-FISH and splicing factor signals. Values represent averages from at least 100 transcription sites from at least three experiments ± SEM.
To further validate the correlation between recruitment and exon 10 inclusion, we used a FISH probe complementary to exon 10, thus specifically detecting mRNA containing exon 10 (Figure 3). As expected, splicing factors accumulated at sites of transcription of tau exon 10-containing pre-mRNA (Figure 3). This probe allowed us to test two predictions of differential recruitment. If the extent of recruitment indeed is related to splicing outcome, splicing factors should also be recruited to the subpopulation of transcription sites in tau10wt cells, which generate detectable levels of exon 10-containing tau pre-mRNA. To test this prediction we used the exon10-specific FISH probe to visualize tau10wt cells, which generate detectable levels of exon 10-containing mRNA and determined whether SF2/ASF and SC35 are recruited to those sites (Figure 3). As predicted, SF2/ASF and SC35 were recruited to the subpopulation of transcription sites in wt cells, at which exon 10-containing RNA could be detected (Figure 3A). Although fewer than 30% of all transcription sites in wt cells had splicing factors associated with them when visualized with a full-length tau probe, more than 90% of the subpopulation of wt cells in which significant levels of exon 10-containing RNA were detected with the exon 10-specific probe recruited SF2/ASF and SC35 (Figure 3B). As expected, no increase in the percentage of splicing factor recruitment was found in the tau10−1 cell line when the exon 10-specific probe was used (Figure 3B). This observation suggests that the recruitment behavior of splicing factors is primarily influenced by the predominant splicing outcome at a transcription site.
Figure 3 Recruitment of Endogenous Pre-RNA Splicing Factors to Transcription Sites Containing tau Exon 10 mRNA
(A) Recruitment of splicing factors to tau transcription sites detected by combined RNA-FISH using a specific probe against tau exon 10 (green) and IF microscopy with specific anti-splicing factor antibodies (red). All associations were confirmed by linescan analysis. Scale bar = 2.5 μm.
(B) Quantitation of percentage of cells with colocalization of tau RNA-FISH and splicing factor signals. Values represent averages from at least 100 transcription sites from at least three experiments ± SEM. Recruitment to the subpopulation of transcription sites containing predominantly tau exon10 mRNA in tau10wt cells was as efficient as recruitment in tau10−1 cells.
A second prediction from a recruitment model is that the percentage of cells that generate detectable amounts of mRNA containing exon 10 should be similar to the percentage of cells in which recruitment of SF2/ASF and SC35 is detected when using the full-length probe. Using the exon 10-specific FISH probe, we indeed found that the percentage of cells containing an mRNA-FISH signal for exon 10 was virtually identical to the extent of recruitment observed using a tau full-length probe. In tau10wt cells, 38 ± 8% of cells were positive for exon 10, comparable to the 31 ± 2 % and 38 ± 13% of cells with recruitment of SF2/ASF and SC35, respectively. These observations demonstrate that recruitment of splicing factors is a reflection of the production of exon10-containing RNA at a site of transcription.
Recruitment Is Determined by Splicing Outcome
Tau10wt and the tau10 −1 mutant differ in a single residue at position −1 in the 5′ splice site [30]. Although sequence analysis using manual approaches and EsEFINDER (http://www.rulai.cshl.org/tools/ESE) did not reveal any strong consensus binding sites for SF2/ASF or SC35 in this region (unpublished data), we asked whether the specific mutation in −1 was responsible for the differential association of SF2/ASF or whether differential recruitment of this factor was correlated with splicing outcome per se. To this end, we generated stable cell lines expressing tau minigenes containing a G >A mutation at position +3 (tau10+3) or a C > U mutation at position +14 (tau10+14). These mutations have previously been shown to result in preferential inclusion of exon 10, similar to tau10−1 in patient cells and in transient transfection assays [30]. As expected, stable cell lines expressing either tau10+3 or tau10+14 preferentially included exon 10 (Figure 4A). Quantitative real-time PCR analysis indicated that inclusion in these cell lines was 3- to 4-fold above that observed in tau10wt cells (Figure 4B). Similar to tau10wt and tau10−1 cell lines, typically a single integration site was detected in these cells by RNA-FISH (unpublished data). As observed for tau10−1, U2-B” and SRp40 showed no differential recruitment and these proteins were equally efficiently recruited to tau10wt as to tau10+3 or tau10+14 transcripts, with more than 80% of cells showing an accumulation in either cell type (Figure 4C). Similar to the situation in tau10−1 cells, SF2/ASF was strongly recruited to transcription sites in tau10+3 and tau10+14 cells, with more than 80% of cells showing co-localization (Figure 4C), but was only recruited to 31 ± 2% of transcription sites in tau10wt cells (Figure 4C). SC35 was recruited to 38 ± 5% of tau10wt transcription sites, but to more than 90% of transcription sites in tau10+3 or tau10+14. We conclude that recruitment of these proteins is determined by the splicing outcome and not by the nature of the mutation near the splice site. Taken together, these results suggest that SF2/ASF and SC35 are differentially recruited to alternatively spliced transcripts and that this recruitment correlates with splicing outcome.
Figure 4 Mutation-Independent Differential Recruitment of Endogenous Pre-RNA Splicing Factors
(A) Stable cell lines expressing minigenes containing a G>A mutation at +3 (tau10+3) or a C>U mutation at +14 (tau10+14). Both mutations give preferential inclusion of exon 10 as previously reported for transient expression.
(B) Quantitative real-time PCR analysis demonstrating the preferential inclusion of exon 10 in tau10+3 and tau10+14 compared to tau10wt. Values represent averages ± standard deviations from at least three experiments. Values of included exon 10 are normalized to total minigene RNA.
(C) Quantitative analysis of recruitment of endogenous splicing factors to tau10+3 or tau10+14. Recruitment of splicing factors to tau transcription sites detected by combined RNA-FISH using specific probes against the tau minigene and IF microscopy with specific anti-splicing factor antibodies. Percentage of cells with colocalization of tau RNA-FISH and splicing factor signals is indicated. Values represent averages from at least 100 transcription sites from three experiments ± SEM.
Recruitment of Exogenous SR Proteins
The analysis of endogenous splicing factor recruitment is restricted by the availability of only a limited set of specific antibodies to SR proteins and simultaneous detection of multiple splicing factors is prevented by species cross-reactivity of the available antibodies. To expand our analysis we introduced epitope-tagged SR proteins into tau-expressing cells using transient transfection. Since transient transfection of SR proteins has previously been shown to be able to affect alternative splicing outcome [19], we first determined in a pilot experiment whether expression of SR proteins alters alternative splice site choice in the stably expressed tau minigene. Upon transfection of characterized T7-tagged SRp20, SRp30s, SRp55, 9G8, or SF2/ASF [32,33], a modest increase in inclusion was observed in tau10wt cells and no effect was evident in tau10−1 cells (Figure 5A). The absence of an effect was not due to sub-optimal transfection efficiency since a majority of cells expressed the T7-tagged SR proteins as confirmed by fluorescence microscopy staining (unpublished data). When analyzed in the recruitment assay, as a control and as observed for endogenous SF2/ASF, transiently transfected T7-SF2/ASF was preferentially recruited to tau10−1 (77 ± 5%) compared to tau10wt transcripts (54 ± 9%; Figure 4B), further confirming the transient expression approach. The percentage of recruitment-positive tau10wt cells appeared somewhat higher upon transient transfection compared to that observed for the endogenous protein (54 ± 9% for transient; 31 ± 2% for endogenous), most likely due to the generally more diffuse distribution of the overexpressed splicing factor [32]. Exogenous SRp20 behaved similarly to SF2/ASF and was differentially recruited in tau10−1 cells compared to tau10wt cells (Figure 5B). While T7-SRp20 was recruited in 56 ± 9% of tau10wt cells, the protein was found at tau-1 transcription sites in 81 ± 4% of cells (p < 0.005). No statistically significant differences were observed for the recruitment of 9G8, SRp30s and SRp55 with all proteins being recruited to more than 70% in either tau10wt or tau10−1 cells (Figure 5B). These observations suggest that SRp20 is similarly differentially recruited to alternatively spliced tau as SF2/ASF.
Figure 5 Differential Recruitment of Exogenous Pre-RNA Splicing Factors to tau Minigenes
(A) Semi-quantitative RT-PCR analysis of tau alternative splicing upon expression of T7-tagged SR proteins in COS-7 cells expressing either tau10wt or tau10−1. Transient transfection of splicing factors does not affect the alternative splicing pattern of the tau minigene in stable cell lines.
(B) Quantitative analysis of recruitment of transfected SR proteins to tau10wt or tau10−1. Recruitment of splicing factors to tau transcription sites detected by combined RNA-FISH using specific probes against the tau minigene and IF microscopy with anti-T7 antibody. Percentage of cells with colocalization of tau RNA-FISH and splicing factor signals is indicated. Values represent averages from at least 100 transcription sites from three experiments ± SEM.
Mapping of Differential Recruitment Domains in SF2/ASF
Our results on endogenous and exogenous SF2/ASF demonstrate that this protein is preferentially recruited to the tau minigene when exon 10 is included. We sought to determine what parts of SF2/ASF were responsible for recruitment to tau10−1. SF2/ASF contains two RNA-binding domains in its N-terminal half and a characteristic SR-rich protein–protein interaction domain at its C-terminus (Figure 6A). To ask whether recruitment involved RNA and/or protein–protein interactions, we expressed a set of previously characterized domain deletion mutations of SF2/ASF in tau10−1 cells [32] (Figure 6A). As observed for the wt SF2/ASF protein, expression of the deletion mutants had no significant effect on splicing outcome of tau10−1 (Figure 6B). When the ability of these proteins to be recruited to tau10−1 was assessed, we found that deletion of any one domain resulted in reduction of recruitment from 48 ± 10% for the wt control to ∼ 30–40% for any of the deletion mutants, suggesting that no single domain alone is responsible for recruitment (Figure 6C and 6 D). In contrast, deletion of any combination of two domains resulted in almost complete loss of recruitment (Figure 6C and 6D). None of these latter mutants were recruited to transcription sites in more than 15% of cells (Figure 6C and 6D). Note that expression of the RS-domain alone results in cell toxicity and was not analyzed [32]. These observations suggest that all domains contribute to recruitment of SF2/ASF to tau10−1 and that both RNA binding as well as protein–protein interactions are required for efficient recruitment to alternatively spliced tau. The involvement of multiple domains in the efficient recruitment to alternatively spliced transcripts is similar to the requirements for recruitment to constitutively spliced transcripts [27].
Figure 6 Deletion Mapping of Sf2/Asf Protein Domains Involved in Differential Recruitment
(A) Schematic representation of T7-tagged mutants of SF2/ASF.
(B) Semi-quantitative RT-PCR analysis of tau10−1 splicing upon expression of mutant SF2/ASF. Overexpression of the mutant proteins does not affect tau splicing.
(C) Recruitment of splicing factors to tau transcription sites detected by combined RNA-FISH using specific probes against the tau minigene (green) and IF microscopy with anti-T7 antibody (red). Arrowheads indicate tau transcription sites. Scale bar = 2.5 μm.
(D) Quantitation of percentage of cells with colocalization of tau RNA-FISH and splicing factor signals. Values represent averages from at least 50 transcription sites from three experiments ± SEM.
Discussion
We have used quantitative single cell analysis to address the long-standing question of whether differentially spliced transcripts recruit distinct sets of basal pre-mRNA splicing factors. Biochemical methods have been unable to resolve this issue most likely due to the heterogeneity of isolated transcripts and spliceosomes and the dynamic nature of the spliceosome [2]. To circumvent these limitations, we have applied a validated microscopy recruitment assay allowing us to quantitatively analyze the association of splicing factors with alternatively spliced transcripts in vivo. Based on analysis of endogenous and exogenous SR proteins, we find that the accumulation of a subset of SR proteins at transcription sites correlates with the inclusion of a alternatively spliced exon. While all tested SR proteins accumulated at the transcription sites in cells where exon 10 of the tau gene is predominantly included, several SR proteins, including SF2/ASF, SC35, and SRp20, were found with reduced frequency at transcription sites in cells where tau exon 10 is excluded. The differentiation recruitment of a subset of splicing factors gives rise to combinatorial occupancy on the pre-mRNA.
Two general models for control of alternative splicing have been suggested. A first model envisions that alternative splicing events are controlled by specific alternative splicing factors. Clear examples of this model include sex-lethal in Drosophila melanogaster and NOVA-1, a human neuron specific alternative splicing factor [14,15]. However, use of dedicated alternative splicing regulators might be the exception to the rule. Considering that the majority of the approximately 25,000 human genes are alternatively spliced, many of them at multiple sites, it is hard to image the existence of specific splicing factors for all of these events. An alternative model suggests that the outcome of alternative splicing is largely determined by the combinatorial association of a commonly used set of splicing factors [8]. Evidence for this model is the fact that both in vivo and in vitro alternative splice site selection can be influenced by antagonistic titration of several splicing components [18–21]. In addition, this model is in line with the observation that several antagonistic splicing factors show tissue-specific abundance possibly contributing to tissue specific splicing patterns [34]. Our observations of differential recruitment of constitutive splicing factors verify the key prediction of this model that distinct sets of splicing factors associate with alternatively spliced transcripts. Furthermore, as expected for this model, recruitment of SF2/ASF involves the same protein regions as recruitment to a constitutively spliced transcript suggesting that recruitment to alternatively spliced introns is similar in its molecular mechanisms as recruitment to constitutive introns [27].
Given the differential recruitment of a subset of SR proteins, these factors might be expected to contribute to determining the splicing fate of tau exon 10. These proteins have indeed been shown to slightly favor inclusion of exon 10 in transient transfections in COS-7 cells [31]. However, in our hands only a moderate increase in exon 10 inclusion in the stably integrated tau minigenes was detected upon overexpression of SF2/ASF or SRp20. The limited ability of these differentially recruited proteins to shift the splicing outcome suggests that their mere presence at the template is not sufficient to control alternative splice site selection and that they are not the sole determinants of splice site choice, but that they cooperate with other spliceosome components. In the absence of information on the stoichiometry of SR proteins in spliceosomes involved in alternative splice site choice, it is difficult to predict the effect of their overexpression on a given template.
Our findings are in line with a dynamic, largely stochastic model for formation of the spliceosome on nascent RNA [8]. It seems likely that pre-mRNA splicing factors, which are known to roam the nucleus by diffusion [35,36], are temporarily captured near nascent pre-mRNAs by interaction with the C-terminal domain of the largest subunit of RNA polymerase [37,38]. Our observation of differential accumulation of SR proteins suggests that their putative interactions with the CTD are only transient, since stable interaction would result in equal recruitment of SR proteins to sites of transcription irrespective of splicing outcome. Since binding events are dynamic and probabilistic, not all assembled spliceosomes are identical and multiple types of spliceosomes containing different combinations of sets of splicing factors are formed. The nature of the formed spliceosomes is likely determined by the cellular and local abundance of factors, their post-translational modifications, and their interaction with specific regulatory splicing enhancers and repressor elements in the pre-mRNA. The various types of spliceosomes likely have distinct functional properties and recognize competing splice sites with different efficiencies, thus generating a mixed population of spliced transcripts. This model implies that alternative splice site choice is not absolute and that multiple isoforms of alternatively spliced transcripts should be generated for most transcripts. This is indeed frequently observed for many alternatively spliced pre-mRNAs, including endogenous tau [29,39–41]. A striking example of the stochastic, combinatorial nature of alternative splicing is the presence of different sets of isoforms of the axon guidance protein Dscam in D. melanogaster, where individual neurons express distinct repertoires of Dscam isoform combinations providing a potential mechanism to establish cell identity [42].
A role for combinatorial dynamics in the assembly of the spliceosome is fully consistent with the well-established dynamic changes the spliceosome undergoes during the splicing process where components are exchanged as the splicing reaction progresses [2]. The advantage of a combinatorial mechanism for spliceosome formation is that it offers the opportunity for cells to adjust splicing outcome rapidly in response to intracellular or extracellular signals without the need to synthesize a specific alternative splicing factor, but merely by modulating the binding of the existing splicing machinery by post-translational modifications or change in cellular distribution. This provides an efficient regulatory link that allows rapid response to physiological cues. Precedents for rapid, signal-mediated control of alternative splicing by these mechanisms exist. Signaling-mediated modulation of alternative splicing has been demonstrated for CD44 via phosphorylation of splicing factor Sam68 through the ERK pathway [43]. Similarly, splicing of an adenoviral E1A splicing reporter is affected by alterations of subcellular localization of hnRNPA1 controlled through the MKK(3/6)-p38 pathway [44]. Based on the sum of these considerations, we suggest that combinatorial recruitment of constitutive splicing factors is an essential property of spliceosomes that significantly contributes to generating protein diversity by alternative splicing and plays a critical role in regulation of splicing in vivo.
Materials and Methods
Generation of stable tau cell lines
To generate a plasmid suitable for generation of stable cell lines containing tau minigenes, the characterized tau10wt, −1, +3, and +14 sequences were modified by insertion of a zeocin selection cassette [30]. The tau minigene constructs in pSPL3b were first cut to completion with SphI and heat-deactivated. The entire digest was then end-filled using Klenow fragment and heat deactivated. The DNA was digested with AvrII to completion and the plasmid gel purified. The zeocin cassette was removed from pSV40 (Invitrogen, Carlsbad, California, United States) via double digest with NheI and PvuII and was ligated into the tau minigene plasmid in a 6:1 insert to vector ratio using T4 DNA ligase overnight at 16 °C. This yielded pSPLZtau10.
To generate stable cell lines COS-1 cells were grown on a 100-mm petri dish to 80% confluence in Dulbecco's modified essential medium (DMEM), 10% FBS, 0.3 mg/ml L-glutamine, and 0.1 mg/ml penicillin/streptomycin. Cells were trypsinized, resuspended in 10 ml of DMEM, and pelleted at low speed. The supernatant was removed and cells were resuspended in 200 μl of DMEM containing 4 μg of linearized pSPLZtau10 plasmid, and 16 μg of salmon sperm DNA in a total volume of 30 μl. Cells were electroporated (2-mm gap cuvette, four pulses at 1-ms pulse, and 150-V) in a BTX electroporator ECM 830 (BTX, Holliston, Massachusetts, United States). Cells were placed on a 150-mm petri dish in DMEM overnight. For selection, cells were incubated in DMEM containing 0.9 mg/ml zeocin (Invitrogen). After 1–2 wk, single colonies were selected and expanded in DMEM containing 0.6 μg/ml zeocin. Clones were screened for expression of the tau minigene by RT-PCR using primers located in the HIV-TAT exons. Secondary screening was conducted by RNA-FISH. Stable lines were routinely grown in DMEM with 0.4 μg/ml zeocin.
RT-PCR
Total RNA was extracted using either the RNAWhiz reagent (Ambion, Austin, Texas, United States) or the RNAqueous kit (Ambion) according to the manufacturer's instructions. Conventional RT-PCR was carried out as described in [45] except that Amplitaq Gold (Applied Biosystems, Foster City, California, United States). was used and the PCR reaction was initiated with a 10′ hot start at 95 °C. Quantitative real-time PCR, was performed using iQ SYBR Green Supermix (Biorad, Hercules, California, United States) following the manufacturer's instructions and products amplified, detected, and quantitated using a MyiQ (Biorad) single-color real-time PCR detection system on an iCycler (BioRad). The PCR routine consisted of 3 min at 94 °C, followed by 60 cycles of 20 s at 95 °C, and 1 min at 58 °C. For detection of tau mRNA containing exon 10, the following primers (forward, 5′-TGC AGA TAA TTA ATA AGA AGC TGG AT-3′; and reverse, 5,′- CCG GGA CGT GTT TGA TAT T-3′) located within exon 10 were used. This reaction yields an 81-nucleotide product. To measure the total amount of tau minigene mRNA, the following primers (forward, 5′-GTC ACC TGG ACA ACC TCA AA-3′; and reverse, 5′-CAG CTT GTC ACA GTG CAG TT-3′) located in the TAT donor exon upstream of exon 10 were used. This reaction yields a 62-nucleotide product. Amplification efficiencies were above 90% and production of single bands was confirmed by gel electrophoresis and melting curve analysis in the same run as measurements were done. RNA amounts were determined using calibrated standard curves using a dilution series over five orders of magnitude. Amounts of exon 10-containing mRNA were normalized to total tau minigene RNA. Values represent averages from at least two experiments ± standard deviations.
To determine the effect of overexpression of SR proteins on tau minigene splicing, COS-1 cells stably expressing tau10wt or tau10−1 were transiently transfected as described above with T7-tagged SR protein expression constructs in pCG-T7 as described in [32]. Microscopy was performed 16 h post-transfection.
FISH and IF microscopy
Stable cell lines were treated for FISH exactly as described in [27] except that 0.1% Triton was used during the initial fixation instead of 0.5%. IF was performed exactly as described in [46]. For detection of endogenous splicing factors, FISH was carried out prior to IF without fixation or permeabilization steps between the two procedures. Cells transiently transfected with T7-tagged splicing factors were first subjected to IF, postfixed in 2% paraformaldehyde for 10 min followed by three rinses in PBS of 5 min each. RNA-FISH was then performed overnight. Tau was detected using probes against the entire minigene or a probe spanning the entire exon 10 generated by nick-translation as previously described [27].
Microscopy and quantitative recruitment assay
Images were collected on a Leica TCS-SP confocal microscope. Single optical sections using 8× averaging were acquired by simultaneous scanning to avoid artifactual shift between the two optical channels. The 488-nm laser line of an Ar laser was used for detection of FITC RNA-FISH signals and the 568-nm line of a Kr laser for detection of TexasRed IF signals. Images were analyzed using Metamorph imaging software (Universal Imaging, Downington, Pennsylvania, United States). Recruitment was detected as accumulation of the IF signal at the tau site of transcription as previously described [27]. Positive recruitment was defined by two simultaneous criteria: a) the splicing factor signal was accumulated at the transcription site more than 2-fold above nucleoplasmic signal and b) at least half the splicing factor intensity peak overlapped with at least half of the RNA-FISH intensity peak. All quantitations were performed on unprocessed 8-bit grayscale images with no saturated pixels. At least 100 foci from at least three experiments were quantitated for each splicing factor per cell line. Statistical analysis was performed using a standard Student t-test.
For SF2/ASF mutants that do not give a distinct speckled pattern, quantitation was performed by measuring the average pixel intensity of the TexasRed signal at the transcription site and comparing it to the average TexasRed signal in the rest of the nucleus excluding nucleoli. Enrichment of the splicing factor signal of > 1.4-fold above its average nuclear intensity was defined as recruitment. This threshold was determined empirically by comparing the degree of recruitment of T-7 tagged SF2/ASF using the two quantitation methods. At least 50 foci from at least three experiments were analyzed for each SF2/ASF mutant.
We thank Javier Caceres for critical comments on the manuscript. Fluorescence imaging was performed at the NCI Fluorescence Imaging Facility. This research was supported by the Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research. TM is a Fellow of the Keith R. Porter Endowment for Cell Biology.
Competing interests. The authors have declared that no competing interests exist.
Author contributions. TM conceived and designed the experiments. SAM performed the experiments and analyzed the data. SAM and TM wrote the paper.
Citation: Mabon SA, Misteli T (2005) Differential recruitment of pre-mRNA splicing factors to alternatively spliced transcripts in vivo. PLoS Biol 3(11): e374.
Abbreviations
DMEMDulbecco's modified essential medium
FISHfluorescence in situ hybridization
snRNPsmall nuclear ribonucleoprotein particles
wtwild-type
==== Refs
References
Sanford JR Caceres JF Pre-mRNA splicing: Life at the centre of the central dogma J Cell Sci 2004 117 6261 6263 15591240
Staley JP Guthrie C Mechanical devices of the spliceosome: Motors, clocks, springs, and things Cell 1998 92 315 326 9476892
Fu X-D The superfamily of arginine/serine-rich splicing factors RNA 1995 1 663 680 7585252
Graveley BR Sorting out the complexity of SR protein functions RNA 2000 6 1197 1211 10999598
Black DL Mechanisms of alternative pre-messenger RNA splicing Annu Rev Biochem 2003 72 291 336 12626338
Graveley BR Alternative splicing: Increasing diversity in the proteomic world Trends Genet 2001 17 100 107 11173120
Caceres JF Kornblihtt AR Alternative splicing: Multiple control mechanisms and involvement in human disease Trends Genet 2002 18 186 193 11932019
Smith CW Valcarcel J Alternative pre-mRNA splicing: The logic of combinatorial control Trends Biochem Sci 2000 25 381 388 10916158
Black DL Protein diversity from alternative splicing: A challenge for bioinformatics and post-genome biology Cell 2000 103 367 370 11081623
Boue S Letunic I Bork P Alternative splicing and evolution Bioessays 2003 25 1031 1034 14579243
Faustino NA Cooper TA Pre-mRNA splicing and human disease Genes Dev 2003 17 419 437 12600935
Cartegni L Chew SL Krainer AR Listening to silence and understanding nonsense: Exonic mutations that affect splicing Nat Rev Genet 2002 3 285 298 11967553
Dredge BK Polydorides AD Darnell RB The splice of life: Alternative splicing and neurological disease Nat Rev Neurosci 2001 2 43 50 11253358
Nagoshi RN McKeown M Burtis KC Belote JM Baker BS The control of alternative splicing at genes regulating sexual differentiation in D. melanogaster
Cell 1988 53 229 236 3129196
Jensen KB Dredge BK Stefani G Zhong R Buckanovich RJ Nova-1 regulates neuron-specific alternative splicing and is essential for neuronal viability Neuron 2000 25 359 371 10719891
Grabowski PJ Black DL Alternative RNA splicing in the nervous system Prog Neurobiol 2001 65 289 308 11473790
Horowitz DS Krainer AR Mechanisms for selecting 5′ splice sites in mammalian pre-mRNA splicing Trends Genet 1994 10 100 106 8178363
Mayeda A Krainer AR Regulation of alternative pre-mRNA splicing by hnRNP A1 and splicing factor SF2 Cell 1992 68 365 375 1531115
Cáceres JF Stamm S Helfman DM Krainer AR Regulation of alternative splicing in vivo by overexpression of antagonistic splicing factors Science 1994 265 1706 1709 8085156
Yang X Bani MR Lu SJ Rowan S Ben-David Y The A1 and A1B proteins of heterogeneous nuclear ribonucleoparticles modulate 5′ splice site selection in vivo Proc Natl Acad Sci U S A 1994 91 6924 6928 8041722
Wang J Manley JL Overexpression of the SR proteins ASF/SF2 and SC35 influences alternative splicing in vivo in diverse ways RNA 1995 1 335 346 7489505
Misteli T Cell Biology of transcription and pre-mRNA splicing: Nuclear architecture meets nuclear function J Cell Sci 2000 113 1841 1849 10806095
Lamond AI Spector DL Nuclear speckles: A model for nuclear organelles Nat Rev Mol Cell Biol 2003 4 605 612 12923522
Misteli T Caceres JF Spector DL The dynamics of a pre-mRNA splicing factor in living cells Nature 1997 387 523 527 9168118
Jiménez-García LF Spector DL In vivo evidence that transcription and splicing are coordinated by a recruiting mechanism Cell 1993 73 47 59 8462102
Huang S Spector DL Intron-dependent recruitment of pre-mRNA splicing factors to sites of transcription J Cell Biol 1996 131 719 732
Misteli T Caceres JF Clement JQ Krainer AR WIlkinson MF Serine phosphorylation of SR proteins is required for their recruitment to sites of transcription in vivo J Cell Biol 1998 143 297 307 9786943
Custodio N Carvalho C Condado I Antoniou M Blencowe BJ In vivo recruitment of exon junction complex proteins to transcription sites in mammalian cell nuclei RNA 2004 10 622 633 15037772
Hutton M Lendon CL Rizzu P Baker M Froelich S Association of missense and 5'-splice-site mutations in tau with the inherited dementia FTDP-17 Nature 1998 393 702 705 9641683
Grover A Houlden H Baker M Adamson J Lewis J 5′ splice site mutations in tau associated with the inherited dementia FTDP-17 affect a stem-loop structure that regulates alternative splicing of exon 10 J Biol Chem 1999 274 15134 15143 10329720
Wang J Gao QS Wang Y Lafyatis R Stamm S Tau exon 10, whose missplicing causes frontotemporal dementia, is regulated by an intricate interplay of cis elements and trans factors J Neurochem 2004 88 1078 1090 15009664
Cáceres JF Misteli T Screaton G Spector DL Krainer AR Role of the modular domains of SR-proteins in subnuclear localization and alternative splicing specificity J Cell Biol 1997 138 225 238 9230067
Cáceres JF Screaton GR Krainer AR A specific subset of SR proteins shuttles continuously between the nucleus and the cytoplasm Genes Dev 1998 12 55 66 9420331
Hanamura A Caceres JF Mayeda A Franza BR Krainer AR Regulated tissue-specific expression of antagonistic pre-mRNA splicing factors Rna 1998 4 430 444 9630249
Kruhlak MJ Lever MA Fischle W Verdin E Bazett-Jones DP Reduced mobility of the alternate splicing factor (ASF) through the nucleoplasm and steady state speckle compartments J Cell Biol 2000 150 41 51 10893255
Phair RD Misteli T High mobility of proteins in the mammalian cell nucleus Nature 2000 404 604 609 10766243
Neugebauer KM On the importance of being co-transcriptional J Cell Sci 2002 115 3865 3871 12244124
Bentley D The mRNA assembly line: Transcription and processing machines in the same factory Curr Opin Cell Biol 2002 14 336 342 12067656
Relogio A Ben-Dov C Baum M Ruggiu M Gemund C Alternative splicing microarrays reveal functional expression of neuron-specific regulators in Hodgkin lymphoma cells J Biol Chem 2004 280 4779 4784 15546866
Pan Q Shai O Misquitta C Zhang W Saltzman AL Revealing global regulatory features of mammalian alternative splicing using a quantitative microarray platform Mol Cell 2004 16 929 941 15610736
Yeakley JM Fan JB Doucet D Luo L Wickham E Profiling alternative splicing on fiber-optic arrays Nat Biotechnol 2002 20 353 358 11923840
Neves G Zucker J Daly M Chess A Stochastic yet biased expression of multiple Dscam splice variants by individual cells Nat Genet 2004 36 240 246 14758360
Matter N Herrlich P Konig H Signal-dependent regulation of splicing via phosphorylation of Sam68 Nature 2002 420 691 695 12478298
van der Houven van Oort W Diaz-Meco MT Lozano J Krainer AR The MKK3/6 -p38 signaling cascade alters the subcellular distribution of hnRNP A1 and modulates alternative splicing regulation J Cell Biol 2000 149 307 316 10769024
Kalbfuss B Mabon SA Misteli T Correction of alternative splicing of tau in frontotemporal dementia and parkinsonism linked to chromosome 17 J Biol Chem 2001 276 42986 42993 11560926
Misteli T Spector DL Serine/threonine phosphatase 1 modulates the subnuclear distribution of pre-mRNA splicing factors Mol Biol Cell 1996 7 1559 1572 8898362
|
16231974
|
PMC1262628
|
CC BY
|
2021-01-05 08:21:45
|
no
|
PLoS Biol. 2005 Nov 25; 3(11):e374
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030374
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0030389SynopsisEvolutionDrosophilaNew Environments Set the Stage for Changing Tastes in Mates Synopsis11 2005 25 10 2005 25 10 2005 3 11 e389Copyright: © 2005 Public Library of Science.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Divergent Selection and the Evolution of Signal Traits and Mating Preferences
Experimentally manipulating the resource environment of Drosophila serrata reveals that mating preferences can evolve, at least in part, as a result of environmentally-based divergent natural selection.
==== Body
In the evolutionary war of the sexes, females choose their mates while males fight for the right to inseminate. Darwin explained this widely observed phenomenon in terms of energy expenditure: whichever sex invests more to produce and rear offspring gets to choose. That lot typically falls to females, whose mating preferences have driven the evolution of secondary sex characteristics as diverse as the peacock's extravagant tail and the fiddler crab's outsized claw. Such preferences may also influence speciation by causing reproductive isolation, acting as a behavioral barrier to gene flow between populations in much the same way mountain ranges act as physical barriers. In both cases, isolated populations that once interbred can diverge into separate species.
The effect of sexual selection on speciation has been demonstrated in many different organisms, but it's not so clear which evolutionary mechanisms—genetic drift or natural selection—account for the initial shift in mating preferences that generate divergent sexual selection. Different mating preferences could arise as a by-product of chance events related to unique mutations (genetic drift) that produce arbitrary traits later modified by sexual selection, or as a side effect of changes in traits that arise as populations adapt differently to their local environments (divergent natural selection). If divergent selection affects female mating preferences (assuming that mating preference differences contribute to reproductive isolation), then separate populations that adapt to different environments should also diverge in mating preferences, while populations adapted to similar environments should not.
An evolutionary experiment that manipulated the resource environment of Australian fruitflies (shown mating on a strawberry) showed that female mating preference can evolve, at least in part, in correlation with the environment. (Image: A. O'Toole, University of Queensland)
Working with the Australian fruit fly Drosophila serrata, Howard Rundle, Mark Blows, and their colleagues at the University of Queensland in Australia investigated the role of divergent selection in the evolution of female mating preferences. Mate choice in D. serrata is mediated by nonvolatile pheromones in the insect's outer cuticle, called cuticular hydrocarbons (CHCs). In past experiments, male CHCs had been shown to evolve rapidly in response to changes in selection (which is not surprising since they protect the fly against environmental vagaries), but the consequence for female mating preferences was not known.
To address the effect of divergent selection on the evolution of female mating preferences, the authors created different environments in the lab by raising four duplicate fly populations on three different food resources—with yeast representing the ancestral lab environment (these flies have eaten yeast since the stock was established in 1998), and rice and corn representing two novel environments. Flies were raised for 22 months, then fed yeast for two generations to control for environmental effects, before both CHCs and female mating preferences were estimated for each of the 12 populations.
To estimate female mating preferences, a single female from one of the experimental populations was placed in a vial with two males from the ancestral stock population, providing standard males for comparison of preferences among the populations. An average of 106 trials were conducted for each of the populations. After females had mated with one of the two males, CHCs from the chosen and rejected males were extracted for analysis. Female mating preferences were then determined for each population by calculating sexual selection gradients that related the mating success of the males with their CHCs.
CHC profiles for all the flies revealed that nearly every CHC molecule had adapted to the novel environments, although CHC evolution was greater in females than in males. Surprisingly, however, the mating trials showed that female mating preferences had also diverged consistently among populations in correlation with their environment (preferences were similar among populations from the same environment, but differed among populations from different environments). This so-called parallel evolution, the authors argue, implicates divergent selection over drift in preference evolution because genetic drift is unlikely to produce a pattern of preference evolution that is predictable by environment.
Altogether, the authors conclude, this evolutionary experiment shows that mating preferences “can evolve at least in part in correlation with the environment.” This result is consistent with the classic by-product model of speciation, in which new species arise as a side effect of divergent selection; in this case, mating preferences act as a premating isolation mechanism that arises along with the divergent environments. Interestingly, the authors found no correlation between the CHCs that adapted most and those for which female preferences changed. Teasing apart the relative contributions of natural and sexual selection in the evolution of CHCs and mating preferences may help shed light on the complicated relationship between trait and preference evolution in general—and on the role that preference plays in the emergence of new species. —Liza Gross
|
0
|
PMC1262629
|
CC BY
|
2021-01-05 08:21:28
|
no
|
PLoS Biol. 2005 Nov 25; 3(11):e389
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030389
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0030391SynopsisBioinformatics/Computational BiologyCell BiologyImmunologySystems BiologyBiochemistryMus (Mouse)MammalsVertebratesAnimalsEukaryotesExperimentally Validated Model Accounts for T Cells' Discriminating Ways Synopsis11 2005 25 10 2005 25 10 2005 3 11 e391Copyright: © 2005 Public Library of Science.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Modeling T Cell Antigen Discrimination Based on Feedback Control of Digital ERK Responses
.
==== Body
When a pathogen slips into the body, it might infect a cell or get eaten by a specialized white blood cell. Either way, its proteins get chopped into peptide fragments, loaded onto molecules encoded by genes in the major histocompatibility complex (MHC), and sent to the cell surface as a peptide-loaded MHC molecule (pMHC) for immune surveillance. The immune system can rally against billions of pathogens, in part because every T cell expresses a unique receptor (TCR), acquired during development, that recognizes specific pMHCs. Developing T cells undergo a selection process that weeds out more than 98% of cells, leaving only those cells whose receptors react to “self” pMHCs enough to signal but not enough to fully activate the T cells. Because antigen-presenting cells (APCs) in an infected host bear both self and pathogen-derived pMHCs, proper immune function depends on the discriminatory capacity of TCRs: they must allow a full T cell response to foreign antigens and avoid reacting to self-peptides on the same cell surface.
When a T cell homes in on a pathogenic antigen, the response is quick, sensitive (just a few molecules can set them off), and digital (all or nothing), engaging signaling pathways necessary for T cell activation, proliferation, and survival. (Molecules called agonist ligands trigger TCR signaling.) Signals relay through the cell as kinase enzymes activate proteins by triggering chemical reactions that add one or more phosphates (a process called phosphorylation); removing phosphates (called dephosphorylation) deactivates the proteins and the signal decays.
One explanation for this rapid, sensitive response is that pMHCs induce unique conformational changes in the receptor to initiate signaling. But a TCR can react differently to the same pMHCs at distinct points in its life history, and binding pairs show no conformations specific to agonists versus non-agonists. Alternately, in the kinetic proofreading concept, a kinetic threshold related to the TCR–pMHC binding properties accounts for ligand discrimination—though in simple forms of this hypothesis discrimination occurs at the expense of a rapid, sensitive, digital response.
Hoping to resolve these discrepancies, Grégoire Altan-Bonnet and Ronald Germain combined computer modeling with experimental results to develop a quantitative model of early TCR signaling. The authors revealed a novel aspect of TCR signaling—the explosive digital response of a key enzyme in the pathway—and identified competing feedback systems that may explain how T cells combine selectivity with a rapid, sensitive response.
To construct a predictive model, Altan-Bonnet and Germain measured kinetic aspects of biochemical responses to TCR–pMHC pairing by focusing on extracellular signal-related kinase (ERK), a key player in the TCR signaling cascade. Using T cells harvested from transgenic mice engineered to produce identical TCRs and APCs engineered to express very few self-pMHCs, the authors studied the conditions required for ERK activation. In individual T cells, the ERK pathway was activated by as few as ten foreign ligands. These experiments also revealed a “previously unappreciated aspect” of ERK signaling in T cells: it either occurred or it didn't, but when it did, the result was 100,000 phosphorylated ERK enzymes.
T cells that showed this explosive response could also distinguish between pMHCs with minor differences in TCR binding time, reinforcing the importance of the SHP-1 negative feedback system in preventing “sneak-through activation” by non-agonists. (SHP-1 is a dephosphorylating enzyme.) To test this hypothesis in a quantitative manner, Altan-Bonnet and Germain constructed a TCR signaling model in which kinetic proofreading of TCR–ligand interactions produces a quick initial negative feedback response (mediated by SHP-1) and a delayed, explosive digital ERK positive feedback system that, once activated, overrides the negative pathway and allows productive signaling. The authors then ran signaling simulations using different ligand numbers with different TCR binding lifetimes. Their model “shows almost absolute discrimination” between closely related pMHCs while preserving a fast, sensitive response to just a few agonist ligands.
Confocal microscopy was used to determine the cytoplasmic volume of T cells to infer concentrations of signaling molecules in the cytoplasm
The model also yielded predictions that the authors validated experimentally: the ERK response slows down dramatically at low ligand densities; negative feedback adjusts to ligand strength and quantity to prevent signaling by high concentrations of low-affinity ligands, and allows sensitive responses to low concentrations of high-affinity ligands; differential activation of the negative feedback explains the existence and hierarchy of antagonism in T cell activation; and mature differentiating T cells permit signaling with different levels of ligand discrimination, depending on intracellular concentrations of molecules such as SHP-1.
Altogether, these results suggest that ligand discrimination is not “hard-wired” into TCR–ligand structural affinities. Rather, the threshold that permits TCR signaling varies as concentrations and dynamics of intracellular molecules vary during T cell development and after antigen activation. The model described here, though a necessarily simplified version of TCR signaling, highlights the effectiveness of simple feedback loops in helping cells filter out unwanted signals while effecting quick, sensitive responses—properties crucial for most other regulatory networks. The model also reinforces the importance of detailed probing of cell signaling dynamics to better understand the functions of a living system. —Liza Gross
|
0
|
PMC1262630
|
CC BY
|
2021-01-05 08:28:16
|
no
|
PLoS Biol. 2005 Nov 25; 3(11):e391
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030391
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0030403SynopsisEcologyEvolutionInfectious DiseasesEpidemiology/Public HealthVirusesPrimatesCharting the Path of the Deadly Ebola Virus in Central Africa Synopsis11 2005 25 10 2005 25 10 2005 3 11 e403Copyright: © 2005 Public Library of Science.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Wave-Like Spread of Ebola Zaire
==== Body
Thanks to sensationalized accounts of patients with liquefying flesh and spouting blood, the Ebola virus may well be the most feared disease on the planet. But the reality of the virus, which strikes humans and other primates, is grim enough, with patients experiencing sudden onset of fever, headache, intense weakness, and muscle pain, followed by diarrhea, vomiting, severe rash, organ failure, and massive hemorrhaging, sometimes external, within two to 21 days of exposure. The first human Ebola outbreaks occurred between 1976 and 1979 in Sudan and Zaire (now the Democratic Republic of Congo), where 88% of the 318 infected persons died—a typical mortality rate for this strain, called the Zaire strain of Ebola virus (ZEBOV). It's thought that humans acquired the virus after handling infected gorilla and chimp carcasses.
Over the past ten years, separate outbreaks of the deadly Zaire strain have killed hundreds of humans and tens of thousands of great apes in Gabon and the Republic of Congo—which harbor roughly 80% of the last remaining wild gorilla and chimpanzee populations. Between 1983 and 2000, poaching and logging precipitated catastrophic declines in these great apes, but scientists fear that Ebola may pose an equally deadly threat. Any efforts to contain the next epidemic depend on understanding the dynamics of the virus's spread.
Repeated outbreaks of the Zaire strain of the Ebola virus (Ebola virions pictured above) in central Africa were caused by a recent spread of the virus, rather than by a long-persistent strain at each site
In a new study, Peter Walsh, Roman Biek, and Leslie Real combined genetic data with information on the timing and location of past ZEBOV outbreaks to determine the merits of two competing hypotheses to explain the emergence and spread of the virus. In the prevailing view, ZEBOV arose from long-persistent local strains after increased contact between humans or great apes and an unidentified reservoir host. But Walsh et al. found support for the alternative hypothesis: that ZEBOV had recently spread to the outbreak regions. This is good news because a virus that spreads at a predictable rate in a predictable direction is far easier to control than one that emerges by chance or at the hands of an unknown trigger.
The authors modeled the virus's spread based on assumptions of a long-persistent virus versus a recently emerged virus, and tested the predictions of these competing hypotheses using genetic data—gathered from gene sequences taken from human samples at the different outbreak sites—and information on the spatiotemporal dynamics of the outbreaks. Charting the evolutionary relationships of the viral genotypes identified one major lineage with a most recent common ancestor consistent with the 1976 outbreak. Comparing the path of descent with outbreak localities mirrored the timing of the outbreaks, with new outbreaks directly descending from those preceding.
Analyzing the spatiotemporal pattern of outbreaks revealed a spread at the rate of about 50 kilometers/year—a predictable path not likely for a persistent virus—with the first epidemic in Yambuku, then spreading south to Kikwit and west to Booué, Gabon. Plotting the geographic distribution of genotypes revealed a clear spatial structure at both local and regional scales: the genotypes from the 2001–2003 Gabon/Congo outbreaks, for example, decreased in genetic similarity as distance increased. Again, this finding is consistent with the recently emerged hypothesis, which predicts a correlation between genotype and geography at different distances. Simulations of viral evolution in a spreading epidemic also showed a consistent spread pattern and a strong correlation between genetic divergence and spatial separation. Though the authors can't say how the virus was transmitted, the simulations showed that a “simple nearest neighbor contact process” could produce the linear, uniform spread rates found here.
Though the strength of the individual lines of evidence—timing of origin, spatial spread, and genetic/distance ratio—is not conclusive when considered separately, taken together, they support the hypothesis that a “consistently moving wave of ZEBOV infection” recently spread to outbreak sites in Gabon and Congo. Following its current course, ZEBOV may hit populated areas east of Odzala National Park within 1–2 years and reach most parks containing large populations of western gorillas in 3–6 years. Two Ebola outbreaks have already hit human populations west of Odzala, and over the past two years, the largest gorilla and chimp populations in the world, found in Odzala, have been devastated—the disease is spreading to the last unaffected sector of the park right now. These findings suggest that strategies to protect villagers and some of the last remaining wild apes from future outbreaks would do best to concentrate efforts at the front of the advancing wave—and start acting now. —Liza Gross
|
0
|
PMC1262631
|
CC BY
|
2021-01-05 08:21:28
|
no
|
PLoS Biol. 2005 Nov 25; 3(11):e403
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030403
|
oa_comm
|
==== Front
PLoS BiolPLoS BiolpbioplosbiolPLoS Biology1544-91731545-7885Public Library of Science San Francisco, USA 10.1371/journal.pbio.0030406SynopsisCell BiologyMolecular Biology/Structural BiologyEukaryotesAlternative mRNA Splicing: Control by Combination Synopsis11 2005 25 10 2005 25 10 2005 3 11 e406Copyright: © 2005 Public Library of Science.2005This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are properly credited.
Differential Recruitment of Pre-mRNA Splicing Factors to Alternatively Spliced Transcripts In Vivo
==== Body
In 1977, a flurry of papers ushered in a radical new concept in molecular biology—the idea of RNA splicing. It had been known for some years that the information for building organisms is stored as DNA sequences, which are transcribed into messenger RNAs (mRNAs) before translation into proteins. Although it had been established that the DNA and mRNA sequences line up exactly in bacteria, molecular biologists began to suspect in the mid-1970s that the genomes of eukaryotes (organisms with nuclei) are organized somewhat differently. Eukaryotic genes, it turns out, are encoded in small sections scattered over enormous distances of DNA. To make proteins from these “split genes,” the whole length of DNA is transcribed into pre-mRNA and then converted into mRNA by spliceosomes—molecular machines that remove the non-coding pieces of RNA (the introns) and splice together the protein-coding pieces (the exons).
One important consequence of RNA splicing is that one gene can produce several different mRNA variations, or isoforms, simply by stitching together different combinations of exons. For example, a single gene in vertebrates encodes calcitonin (a thyroid hormone that controls calcium levels) and calcitonin-gene-related peptide (a neuropeptide). Alternative splicing also contributes to human disease—for instance, the selection of different splice sites generates aberrant ratios of mRNA isoforms in several neurological diseases.
But how are these alternative splice sites selected? One popular model proposes that alternative splicing in mammalian cells is largely controlled by binding of general splicing factors to pre-mRNA molecules during the formation of the spliceosome. The spliceosome contains many of these factors, including a class of proteins called SR proteins, which contain one or two RNA-binding domains and a protein–protein interaction domain that is rich in serine and arginine amino acids. An important prediction of the combinatorial model for control of alternative splicing is that alternatively spliced transcripts will recruit different combinations of pre-mRNA splicing factors in vivo. New data from Mabon and Misteli support this prediction.
Pre-mRNA splicing factors accumulate at sites of active transcription, and splicing and can be detected and quantified in individual living cells by tagging the splicing factors with fluorescently labeled antibodies. So, to see whether different factors accumulate at alternatively spliced transcripts, the researchers developed stable cell lines carrying versions of the gene encoding a protein called tau designed to splice in different ways. In healthy people, exon 10 of the tau gene is included or excluded from tau mRNA with roughly equal probability during pre-mRNA splicing; in people with a rare Parkinsonism-like neurological disorder, mutations near one end of exon 10 result in its predominant inclusion. The researchers, therefore, examined cell lines carrying tau genes with and without a mutation of this type to change the ratios of mRNA transcripts including or excluding exon 10. Their results show that a subset of SR protein splicing factors is efficiently recruited to tau transcription sites that produce exon 10–containing mRNA, but less efficiently recruited to transcription sites where exon 10 is excluded.
These results provide the first in vivo evidence for the differential association of pre-mRNA splicing factors with alternatively spliced transcripts, and support a combinatorial mechanism for spliceosome formation. Exactly which splicing factors are recruited to each spliceosome would depend on both the concentration of each factor in individual cell types and the regulatory elements present in each pre-mRNA. The combinatorial mechanism for the control of alternative splicing, the authors suggest, could allow cells to adjust splicing outcome (and consequently which proteins they express) rapidly in response to intracellular or extracellular cues, as well as contributing to the generation of protein diversity. —Jane Bradbury
Pre-mRNA splicing factors (red) are recruited to sites of transcription and splicing (green) in the nucleus of mammalian cells
|
0
|
PMC1262632
|
CC BY
|
2021-01-05 08:21:29
|
no
|
PLoS Biol. 2005 Nov 25; 3(11):e406
|
utf-8
|
PLoS Biol
| 2,005 |
10.1371/journal.pbio.0030406
|
oa_comm
|
==== Front
Ann Gen PsychiatryAnnals of General Psychiatry1744-859XBioMed Central London 1744-859X-4-171620212210.1186/1744-859X-4-17Case ReportNo aggression in a 4-year-old boy with an androgen-producing tumour: Case Report De la Marche Wouter [email protected] Karin [email protected] Annemieke M [email protected] Robert F [email protected] Department of Child and Adolescent Psychiatry, Erasmus Medical Center Rotterdam/Sophia Children's Hospital, The Netherlands2 Department of Paediatrics, Erasmus Medical Center Rotterdam/Sophia Children's Hospital, The Netherlands2005 3 10 2005 4 17 17 25 5 2005 3 10 2005 Copyright © 2005 De la Marche et al; licensee BioMed Central Ltd.2005De la Marche et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The androgen testosterone plays a critical role in many aspects of sexual differentiation. Also, it is thought to induce aggressive behaviours or to play a role in social dominance.
Case presentation
In this case report a 4-year-old boy is described whose testosterone and dehydroepiandrosterone sulphate (DHEA-S) levels were raised to pubertal levels due to a testosterone producing testis tumour. This provided the unique opportunity to examine the effects of elevated levels of androgens on levels of aggression or on social dominance before the onset of puberty.
Conclusion
The present case report does not support the hypothesis of a causal relationship between testosterone and aggression or between testosterone and social dominance in young children.
androgenstestosteroneaggressionchildren
==== Body
Background
The androgen testosterone plays a critical role in many aspects of sexual differentiation. Also, it is thought to induce aggressive behaviours. Results of animal studies have indicated that, in most species, androgens, including testosterone, facilitate aggressive behaviour [1-5]. However, studies that investigated associations between androgens and aggression in humans have yielded inconclusive results and are very difficult to compare due to methodological differences. Moreover, most studies have compared hormone levels in individuals with high versus low aggression levels, or have yielded correlations between testosterone levels and the extent of aggression, which makes inferences about a causal relationship impossible. Furthermore, little is known about the relation between testosterone levels and the aggressive behaviour in pre-adolescent children. Because in normal development, testosterone levels are very low from the age of 6 months until pre-puberty, most studies have used samples that consisted of pubertal and post pubertal males. More recent studies hypothesised there might be an association between testosterone and social dominance instead [6-8]. Because ethically it is impossible to set up experiments to respond to the question of causal relationships, we should rely on natural experiments.
In the present report we describe the case of a 4-year-old boy whose testosterone and dehydroepiandrosterone sulphate (DHEA-S) levels were raised to pubertal levels due to pseudo pubertas praecox based on a testosterone producing testis tumour. This provided the unique opportunity to examine the effects of elevated levels of testosterone and DHEA-S on levels of aggression in a young child. The fact that the androgen levels in this boy were raised to about 30 times the normal level for his age makes this case a unique natural experiment.
Case presentation
In June 2002, a boy aged 4 years and three months, was referred to the outpatient clinic of oncology of the Erasmus Medical Center of Rotterdam. One year before, penis growth had started to accelerate rapidly and 2 months prior to admission, his height had started to increase tremendously, pubic hair had started to grow, his voice had started breaking, and sweat production had increased. Furthermore, he often had erections during the day, got interested in older girls, and started reading magazines that contained pictures of naked women.
At clinical examination, his length was 120 cm (more than 2.5 SD above average) and he weighed 24.6 kg (weight/height ratio +1 SD). His sexual development had reached Tanner stages P3, G3 [9], with no underarm hair growth. Examination of his genitals showed a left testis of 2 ml and a right testis of 6 ml, firm and painless. Further clinical examination was negative, except for glue ears. X-rays of his hand showed that his skeletal age was 5 years ahead of his calendar age. His testosterone level (10.6 nmol/l) was raised to a pubertal level, as well as his DHEA-S level (1.30 μmol/l). Tumour markers carcino-embryonal antigen (CEA), alpha-fetoprotein, and beta human choriogonadotropin (β-HCG) were negative. Ultrasound of the testes showed a testicular tumour. Neither abdominal ultrasound nor abdominal CT or full bonescan (209MBq TC99M) showed signs of metastases.
Pseudo pubertas praecox based on a testicular tumour was diagnosed. After unilateral orchidectomy, pathological examination showed a Leydig Cell Tumour. One week after orchidectomy, hormone levels had normalized (testosterone < 0.10 nmol/l, DHEA-S < 0.20 μmol/l), he stopped having erections, and sweat production decreased, as did his sexual interest. His mother reported he had become more of a child again.
Because of anxiety, some behavioural problems, and delayed speech development, the boy was referred to our outpatient unit for child and adolescent psychiatry. At psychiatric examination, January 2003, we saw a boy who was anxious and withdrawn, had problems in social interaction, and a delayed speech development. Parents reported he had always been very quiet, however, since the beginning of 2002 he had started withdrawing himself more and more from social interactions, both at home and in school. To obtain standardized ratings of psychopathology, both parents, as well as the schoolteacher, were asked to fill out the Child Behaviour Checklist for ages 1 1/2–5 (CBCL 1 1/2–5) and the Caregiver-Teacher Report Form for ages 1 1/2–5 (C-TRF 1 1/2–5) [10]. The mother and the father scored the patient in the clinical range of the Withdrawn Behaviour, Anxious/Depressed Behaviour, and Somatic Complaint scales, but not on the Emotionally Reactive Behaviour, Sleep Problems, Attention Problems, and Aggressive Behaviour scales. The clinical cut-off corresponds with the 98th percentile score in a general population sample [10]. Although the schoolteacher described the boy as quiet and withdrawn, and mentioned delayed speech development, C-TRF syndrome scores were not in the clinical range. No symptoms of externalizing behaviour (aggression, hyperactive behaviour or attention problems) were reported by the parents or the teacher, nor seen during psychiatric assessment. Further psychological testing showed that the boy's intellectual functioning was below average (Wechsler Preschool and Primary Scale of Intelligence – Revised [11]: Total IQ = 84; Verbal IQ = 82; Performance IQ = 92). DSM-IV [12] chronic adjustment disorder with anxiety was diagnosed.
Discussion
In this case study, we report about a 4-year-old boy with an androgen producing testis tumour, in whom testosterone and DHEA-S levels rose to pubertal concentrations. It is a normal biological reaction that, when the testes begin to secrete androgens, sex drive and the motivation to seek sexual contact become stronger and are overtly expressed more and more [13,14]. It is remarkable however that, parallel to the sexual development and increased sexual drive, the boy we describe did not show increased levels of aggression. On the contrary, the boy was described as anxious and withdrawn.
The fact that the raised androgen levels are not accompanied by aggression is not consistent with the results of previous studies that found an association between androgen levels (testosterone and/or DHEA-S) and aggression [15-18]. For instance, in a study in which 8- to 12-year-old children (n = 15) with a Conduct Disorder were compared to normal controls (n = 25), van Goozen et al. [18] found significant correlations between DHEA-S levels and scores on the CBCL and TRF subscales Aggressive Behaviour (r = .46 and .48 respectively) and Delinquent Behaviour (r = .33 and .39 respectively). A possible explanation for the inconsistency between these studies and our 4-year-old boy can be that raised androgen levels are not the cause but a consequence of aggression.
Tremblay et al. concluded that testosterone at age 12 makes an independent contribution to explain social dominance at age 13. They consider as well that the high levels of testosterone could be as well the product as much as the cause of social dominance. They further suggest that the testosterone-dominance link should be present from infancy onwards, if it exists [8]. The boy we described in this case report had primarily high levels of testosterone, that didn't lead to social dominance. In contrast, he was anxious and withdrawn, which might have several causes. First, even without the occurrence of a hormone-producing tumour, the boy might have been at risk for these symptoms; family history was negative for anxiety but mother had suffered from a depressive episode before. Second, due to his extreme height, parents and other adults or children may have made age-inappropriate demands, which may have caused anxiety. Third, extreme changes in hormone levels may have directly resulted in withdrawal from social contacts. Sánchez-Martín et al. [19] examined 28 boys and 20 girls with a mean age of 4 years by videotaping them every morning during free play in the classroom during 4 months. They found that higher levels of testosterone were associated with decreased direct interaction with peers. The last explanation is in contradiction with the social dominance hypothesis.
Most of the studies that showed a relation between androgens and aggression were carried out with adults or adolescents. Previous studies with young children in prepuberty did not find much evidence for this relation [15,20]. For instance, Constantino et al. [20] studied 18 very aggressive 4- to 10-year-olds. They did not find an association between serum testosterone levels and the scores on the Aggressive Behaviour scale of the CBCL. To explain this lack of association, it can be argued that concentrations of androgens in young children are too low to find effects of testosterone on aggression anyway. In the boy in the present case report, however, serum testosterone levels were about 30 times the regular levels for his age and signs of aggression were still not found (figure 1).
Figure 1 Age at the time of the picture: 4 years
Conclusion
Although we know that one case cannot proof nor refute any hypothesis about causal relationships or associations between hormone levels and aggression or social dominance, natural experiments like this can help with formulating hypotheses for further research.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Wouter De la Marche, MD: Mane author of the article.
Karin Prinsen, MSc, has made substantial contributions in writing the background and the psychiatric examination, as well as in revising the discussion.
Annemieke M. Boot, MD, PhD has carried out the somatic examination and supervised the technical examinations. She helped us formulating this part of the article. Robert F. Ferdinand, MD, PhD has been involved in revising this article critically for important intellectual content and has given final approval of this version of the article to be published.
Acknowledgements
Written consent was obtained from the patient's parents for publication of this case report.
==== Refs
Bouissou M Androgens, aggressive behaviour and social relationships in higher mammals Horm Res 1983 18 43 61 6350146
Brain PF Brown K, Cooper SJ Effects of the hormones of the pituitary-gonadal axis on behavior Chemical Influence on Behavior 1979 New York: Academic Press 255 329
Bronson FH Desjardins C Eleftheriou BE, Scott JP Steroid hormones and aggressive behavior in mammals The Physiology of Aggression and Defeat 1971 New York: Plenum 43 63
Moyer KE The Psychobiology of Aggression 1976 New York: Harper & Row
Rose RM Bernstein IS Holaday JW Plasma testosterone dominance rank, and aggressive behavior in a group of male rhesus monkeys Nature 1971 231 366 368 4996062 10.1038/231366a0
Rowe R Maughan B Worthman CM Costello EJ Angold A Testosterone, Antisocial Behavior, and Social Dominance in Boys: Pubertal Development and Biosocial Interaction Biol Psychiatry 2004 55 546 552 15023584 10.1016/j.biopsych.2003.10.010
Schaal B Tremblay TE Soussignan R Susman EJ Male Testosterone Linked to High Social Dominance but Low Physical Aggression in Early Adolescence J Am Acad Child Adolesc Psychiatry 1996 35 1322 1330 8885586 10.1097/00004583-199610000-00018
Tremblay RE Schaal B Boulerice B Arsenault L Soussigna RG Paquette D Laurent D Testosterone, Physical Aggression, Dominance, Physical Development in Early Adolescence Int J Behav Dev 1998 22 753 777 10.1080/016502598384153
Tanner JM Growth at Adolescence 1962 Oxford: Blackwell
Achenbach TM Rescorla LA Manual for ASEBA Preschool Forms & Profiles 2000 Burlington VT: University of Vermont, Research Center for Children, Youth, & Families
Wechsler D Wechsler Preschool and Primary Scale of Intelligence – Revised (WPPSI-R) 1989 San Antonio: Psychological Corporation
American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders, (DSM-IV) 1994 4 Washington DC: American Psychiatric Association
Albert DJ Walsh ML Jonik RH Aggression in humans: what is its biological foundation? Neurosci Biobehav Rev 1993 17 405 425 8309650
Christiansen K Behavioural effects of androgen in men and women J Endocrinol 2001 170 39 48 11431135 10.1677/joe.0.1700039
Finkelstein JW Susman EJ Chinchilli VM Kunselman SJ D'Arcangelo MR Schwab J Demers LM Liben LS Lookingbill G Kulin HE Estrogen or testosterone increases self-reported aggressive behaviors in hypogonadal adolescents J Clin Endocrinol Metab 1997 82 2433 2438 9253313 10.1210/jc.82.8.2433
Scerbo AS Kolko DJ Salivary testosterone and cortisol in disruptive children: relationship to aggressive, hyperactive, and internalizing behaviors J Am Acad Child Adolesc Psychiatry 1994 33 1174 1184 7982868
van Goozen SHM Matthys W Cohen-Kettenis PT Thijssen JHH van Engeland H Adrenal androgens and aggression in conduct disorder prepubertal boys and normal controls Biol Psychiatry 1998 43 156 158 9474448 10.1016/S0006-3223(98)00360-6
van Goozen SHM van den Ban E Matthys W Cohen-Kettenis PT Thijssen JHH van Engeland H Increased adrenal androgen functioning in children with oppositional defiant disorder: a comparison with psychiatric and normal controls J Am Acad Child Adolesc Psychiatry 2000 39 1446 1451 11068901 10.1097/00004583-200011000-00020
Sánchez-Martín JR Fano E Ahedo L Cardas J Brain PF Azpíroz A Relating testosterone levels and free play social behavior in male and female preschool children Psychoneuroendocrenology 2000 25 773 783 10.1016/S0306-4530(00)00025-1
Constantino JN Grosz D Saenger P Chandler DW Nandi R Earls FJ Testosterone and aggression in children J Am Acad Child Adolesc Psychiatry 1993 32 1217 1222 8282667
|
16202122
|
PMC1262688
|
CC BY
|
2021-01-04 16:39:07
|
no
|
Ann Gen Psychiatry. 2005 Oct 3; 4:17
|
utf-8
|
Ann Gen Psychiatry
| 2,005 |
10.1186/1744-859X-4-17
|
oa_comm
|
==== Front
Behav Brain FunctBehavioral and brain functions : BBF1744-9081BioMed Central London 1744-9081-1-161620214010.1186/1744-9081-1-16ResearchEfficacy of atomoxetine in adult attention-Deficit/Hyperactivity Disorder: a drug-placebo response curve analysis Faraone Stephen V [email protected] Joseph [email protected] Thomas [email protected] David [email protected] Lenard [email protected] Fred [email protected] Stephen J [email protected] Department of Psychiatry, SUNY Upstate Medical University, Syracuse, NY 13210, USA2 Department of Psychiatry, Harvard Medical School, Massachusetts General Hospital, Boston, MA 01880, USA3 Lilly Research Laboratories, Indianapolis, IN 46285, USA4 New York University School of Medicine, New York, NY 10016, USA5 Mood Disorders Clinic, Department of Psychiatry, University of Utah Health Sciences Center, Salt Lake City, UT 84132, USA6 Institute of Behavioral Genomics, Department of Psychiatry, University of California, San Diego, La Jolla, CA 92093-0603, USA2005 3 10 2005 1 16 16 22 8 2005 3 10 2005 Copyright © 2005 Faraone et al; licensee BioMed Central Ltd.2005Faraone et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The objective of this study was to evaluate the efficacy of atomoxetine, a new and highly selective inhibitor of the norepinephrine transporter, in reducing symptoms of attention-deficit/hyperactivity disorder (ADHD) among adults by using drug-placebo response curve methods.
Methods
We analyzed data from two double-blind, placebo-controlled, parallel design studies of adult patients (Study I, N = 280; Study II, N = 256) with DSM-IV-defined ADHD who were recruited by referral and advertising. Subjects were randomized to 10 weeks of treatment with atomoxetine or placebo, and were assessed with the Conners Adult ADHD Rating Scales and the Clinical Global Impression of ADHD Severity scale before and after treatment.
Results
Those treated with atomoxetine were more likely to show a reduction in ADHD symptoms than those receiving placebo. Across all measures, the likelihood that an atomoxetine-treated subject improved to a greater extent than a placebo-treated subject was approximately 0.60. Furthermore, atomoxetine prevented worsening of most symptom classes.
Conclusion
From these findings, we conclude that atomoxetine is an effective treatment for ADHD among adults when evaluated using several criteria.
==== Body
Introduction
Several compounds are now recognized as effective treatments for the major symptoms of attention-deficit/hyperactivity disorder (ADHD) in adulthood. The most effective of these include methylphenidate and dextroamphetamine (or mixed dextro- and levoamphetamine); however, the use of other agents, such as bupropion and desipramine, has also received some support. In addition to these, atomoxetine, a highly selective noradrenergic reuptake inhibitor with little affinity for other neurotransmitter systems [1], has been shown to be well tolerated and effective in reducing the symptoms of ADHD in adulthood. In fact, the benefits of atomoxetine for adults with ADHD have now been demonstrated in three studies of adult patients [2,3], with each report establishing the superiority of atomoxetine over placebo in reducing inattentive, hyperactive, and impulsive symptoms of the illness [3]. As a result of its demonstrated efficacy and low occurrence of clinically meaningful side effects [4], atomoxetine recently became the first non-stimulant medication approved for use in the United States for the treatment of ADHD in adults.
Thus, the ability of atomoxetine to reduce symptoms of ADHD among adults has been sufficiently established; however, several key questions about its clinical utility remain unresolved. For example, although the initial studies of the efficacy of atomoxetine provided useful information for clinicians treating adults with ADHD, such as the average magnitude of the decrease in ADHD symptoms associated with drug treatment and the reliability of this effect, the standard methods of data presentation in these reports do not provide information about the full range of effects of this compound. To further characterize the clinical performance of atomoxetine, we completed a drug-placebo response curve analysis of the data initially reported by Michelson et al. [3] This method, described by Faraone et al. [5], is a generalization of receiver operating characteristic (ROC) analysis [6], which has been widely applied to assessing the accuracy of diagnostic tests [7-9]. The goal of this method is to identify additional characteristics of drug-placebo differences that have already been shown to be statistically significant, including: 1) the size of the effect using different response criteria; 2) the nature of individual responses; and 3) the portion of the drug's effect that is due to symptom improvement, the prevention of symptom worsening, or both.
Results
Standard Analyses
To provide a foundation for interpreting the results of drug-placebo response curve analysis of the effects of atomoxetine, we have displayed in Table 1 a summary of results of the standard analyses of these data, which were originally presented by Michelson et al. [3]. Although data on clinical global impression (CGI) endpoints or Conners Adult ADHD Rating Scale (CAARS) ADHD index scores were not presented in the initial report, it is clear that atomoxetine had significant efficacy relative to placebo on all other measures derived from the CAARS or CGI assessments. The most reliable reductions in ADHD symptoms elicited by atomoxetine were seen for clinician-rated global impressions of ADHD severity, and investigator- and self-rated total ADHD symptoms. Atomoxetine more robustly reduced inattentiveness than hyperactivity and impulsivity, as assessed by both investigators and subjects. Overall, these data provided strong and conclusive evidence that atomoxetine was superior to placebo in reducing the symptoms of ADHD in adulthood, warranting further analysis by drug-placebo response curve methods.
Table 1 Summary of Effects of Atomoxetine and Placebo on ADHD Symptoms in Adults*
Study I Study II
Placebo (n = 134) Atomoxetine (n = 133) p Placebo (n = 124) Atomoxetine (n = 124) p
CGI Change Score -0.4 ± 1.0 -0.8 ± 1.2 0.010 -0.5 ± 1.0 -0.9 ± 1.2 0.002
Investigator-Rated CAARS
Total ADHD Symptom Scores -6.0 ± 9.3 -9.5 ± 10.1 0.005 -6.7 ± 9.3 -10.5 ± 10.9 0.002
Inattentive Score -3.1 ± 5.8 -5.0 ± 5.7 0.010 -3.5 ± 5.3 -5.8 ± 6.5 0.001
Hyperactive/Impulsive Score -2.9 ± 4.9 -4.5 ± 5.1 0.017 -3.2 ± 4.7 -4.7 ± 5.3 0.013
Self-Rated CAARS
Total ADHD Symptom Score -9.3 ± 14.0 -16.0 ± 16.2 0.002 -11.6 ± 16.1 -17.3 ± 17.6 0.008
Inattentive Score -8.6 ± 13.8 -15.9 ± 16.3 0.001 -11.3 ± 16.6 -17.1 ± 17.9 0.012
Hyperactive/Impulsive Score -7.5 ± 12.1 -11.9 ± 13.5 0.013 -8.8 ± 13.4 -12.5 ± 14.1 0.025
*Data adapted from Michelson et al., 2003
Data are presented as mean ± standard deviation
CAARS Investigator Ratings
Figures 1, 2, 3, 4 show the results for investigator ratings on the CAARS. Figure 1 compares the effects of atomoxetine on total ADHD score with those of placebo. For different definitions of responsiveness (i.e., different CAARS cutting scores), the points on the curve illustrate the results of two calculations: the rate of response to drug and the rate of response to placebo. For example, the point in Figure 1 labeled -7 is located at coordinates [0.4, 0.54], which indicates that 40% of those treated with placebo achieved a change of -7 in total investigator-rated ADHD symptoms on the CAARS, whereas 54% of those treated with atomoxetine attained the same level of responsiveness. Sequential points further down the curve (i.e., toward the origin) specify increasingly stringent thresholds for defining improvement (i.e., larger decreases in investigator-rated total ADHD score), while points further up the curve denote the proportions of each treatment group that responded to treatment as determined by increasingly lenient criteria for improvement. Thus, these points and the curve that joins them illustrate how the drug- and placebo-response rates change as the cutting score used to define improvement is incrementally changed. From the curve, it is clear that if response criteria between total ADHD symptom change scores of -1 and -14 are used as the cutting score, a greater proportion of atomoxetine-treated individuals than placebo-treated patients will attain that level of symptom improvement. In other words, over this range of cutting scores, the majority of individuals judged as responsive to treatment will have received atomoxetine rather than placebo. Near the most extreme cutting scores (in this case, the points labeled 15 and -46), those treated with placebo were as likely as those treated with atomoxetine to reach response criteria.
Figure 1 CAARS Investigator-Rated Total ADHD Score.
Figure 2 CAARS Investigator-Rated Inattention Subscale.
Figure 3 CAARS Investigator-Rated Hyperactive Subscale.
Figure 4 CAARS Investigator-Rated ADHD Index.
Figure 1 (and each figure) also shows the diagonal line of no effect, which allows us to visualize the size of the drug effect as the degree to which the drug-placebo response curve rises above it. If outcome on drug were worse than outcome on placebo, then the drug-placebo response curve would fall below the line of no effect; however, as Figure 1 shows, atomoxetine produced a drug-placebo response curve that was always above the diagonal line of no effect. This indicates that atomoxetine outperformed placebo throughout the full range of outcome scores. The area under the curve (AUC) is 0.60, which means that atomoxetine outperformed placebo 60 percent of the time, regardless of cutting score.
In addition to this information, and unlike a traditional statistical analysis, Figure 1 also allows us to determine if the effects of atomoxetine were due to its ability to improve the symptoms of ADHD, prevent their worsening, or both. For example, for CAARS total ADHD symptom change scores, a value of 0 indicates no change, and this point on the drug-placebo response curve is labeled. At this point on the curve, we see that 82% of subjects receiving atomoxetine had a score of 0 or greater, i.e., only 18% of atomoxetine-treated patients experienced a worsening of their symptoms. In contrast, 71% of placebo-treated subjects had a score of 0 or greater, which indicates that symptoms worsened in 29% of these patients. Thus, the response curve clearly demonstrates that atomoxetine not only reduced symptoms, but prevented their worsening as well.
Figure 2 illustrates the effects of atomoxetine on investigator-rated inattention on the CAARS. As with total ADHD symptoms, inattentive symptoms seem to be more effectively treated by atomoxetine than placebo, as the drug-placebo response curve was above the diagonal line of no effect and the AUC was 0.61. Relative to total ADHD symptoms, there exists for inattention a narrower range of cutting scores over which the greatest difference in responsiveness was seen between atomoxetine- and placebo-treated subjects. At cutting scores of -4 through -8, approximately 50% more atomoxetine-treated subjects reached the response criterion than did placebo-treated subjects.
Figure 3 shows the effects of atomoxetine and placebo on investigator-rated hyperactivity on the CAARS. Relative to Figure 2 (inattention), the drug-placebo response curve in Figure 3 did not rise as far above the diagonal line of no effect and, consequently, the AUC for hyperactivity (0.58) was lower than that for inattention (0.61). These results indicate that atomoxetine was less effective in reducing (and preventing the worsening of) hyperactivity than in improving attention. However, it is still clear that, relative to placebo, atomoxetine was an effective treatment for hyperactivity when any symptom change score criteria of less than 0 was used as the response criterion.
In Figure 4, the effects of atomoxetine and placebo on investigator-rated ADHD index scores are plotted, and the more restricted rise in this curve is immediately apparent relative to that seen in earlier figures. The AUC for this curve (0.59) was significant (p < 0.001), indicating that atomoxetine was more likely than placebo to reduce the ADHD index score. However, there is a distinct peak in this curve at a response criterion of approximately -8, where atomoxetine-treated subjects were approximately twice as likely as placebo-treated subjects to attain this level of symptom improvement; at other cutting scores (e.g., -3), the benefits of atomoxetine were much more modest.
CAARS Self Ratings
Figures 5, 6, 7, 8, illustrate the results for self-ratings on the CAARS. In general, the effects of atomoxetine on self-ratings on the CAARS mirrored its effects on investigator ratings. From Figure 5 it is clear that, as with investigator-ratings of this measure, self-rated total ADHD symptoms were more likely to be reduced in atomoxetine-treated subjects than in those receiving placebo. In fact, the AUC for this measure (0.61) was virtually identical to that observed on investigator ratings of this measure, and beneficial effects of atomoxetine were observed over roughly the same range of cutting scores. More than 60% of subjects treated with atomoxetine attained the median response score (-6), while only approximately 40% of those treated with placebo saw this level of improvement; thus, when the median response score was used as the cutting score for defining responsiveness, atomoxetine had greater efficacy on self reported total ADHD symptomatology than on investigator ratings of this measure (cf, Figure 1).
Figure 5 CAARS Self-Rated Total ADHD Score.
Figure 6 CAARS Self-Rated Inattention Subscale.
Figure 7 CAARS Self-Rated Hyperactive Subscale.
Figure 8 CAARS Self-Rated ADHD Index.
The shape and position of the drug-placebo response curve for self-rated CAARS inattention change scores (Figure 6) was quite similar to the curve for investigator ratings of this measure (Figure 2). Specifically, the curve was above the diagonal line of no effect and had a significant AUC, indicating that atomoxetine administration was more effective than placebo in reducing inattentive symptoms. Of note, the drug-response curve for self-ratings of hyperactivity very closely matched that for investigator-ratings of this measure (Figure 7), indicating an appreciable amount of divergence in subjects' perceptions of the effects of treatment on their constituent symptom clusters. As expected, self-rated ADHD index change scores mirrored investigator ratings in showing a sizeable benefit of atomoxetine over placebo, especially when symptom change scores in the range of -4 to -9 were used as response criteria (Figure 8).
CGI Clinician Ratings
Figure 9 depicts the effects of atomoxetine and placebo on ADHD symptom severity change scores for the CGI. As with the CAARS measures, the drug-placebo response curve occupied the space above the diagonal line of no effect and the AUC approximated 0.60, indicating an advantage of atomoxetine over placebo in reducing clinician-rated ADHD severity. Also in accord with the CAARS measures, a small protection from symptom worsening was afforded by atomoxetine, as approximately 10% of subjects who received the drug deteriorated clinically, while approximately twice as many placebo-treated subjects did so. The median change in CGI in the combined group of atomoxetine- and placebo-treated subjects was -1, a response criterion attained by almost 55% of drug-treated subjects but by only approximately 40% of those receiving placebo. As expected, these greater rates of improvement (and protection from deterioration) led to the attainment of lower CGI endpoint scores in the atomoxetine-treated group (Figure 10).
Figure 9 CGI ADHD Severity (Change Scores).
Figure 10 CGI ADHD Severity (Endpoint Scores).
Discussion
The results of two large randomized, double-blind, placebo-controlled trials of atomoxetine for the treatment of ADHD in adults were initially reported by Michelson et al. [3], who documented the superiority of this compound relative to placebo in reducing total ADHD symptoms, as well as inattentive and hyperactive symptoms of the illness. Due to the ample size and rigorous design of these studies, as well as the strong statistical significance of their results, the efficacy of atomoxetine has been firmly established. However, the simple knowledge that, on average, atomoxetine is efficacious does not tell clinicians much about its full range of effect.
Drug-placebo response curves provide an easily interpretable format for further evaluating clinically informative characteristics of a compound with proven efficacy. Because atomoxetine has demonstrable efficacy, drug-placebo response curve analysis of its performance against placebo was warranted. Collectively, the drug-placebo response curves presented here for each of the different reporters and the various dependent measures paint a consistent picture of the benefits of atomoxetine. First, it is clear that atomoxetine is superior to placebo in reducing total ADHD symptoms as well as individual symptom clusters, such as inattention and hyperactivity. For each of these measures, the drug-placebo response curve was always situated above the line of no effect, indicating that subjects were more likely to respond to atomoxetine than to placebo over the entire range of possible criteria of responsiveness. In addition, it is clear that atomoxetine targeted the core features of ADHD rather than only one of its most conspicuous features of inattention and hyperactivity, as AUCs across total, inattention, and hyperactivity change scores were quite similar. Second, responsiveness to atomoxetine was reliably assessed by clinicians, investigators, and patients, as the AUCs for the various dependent measures varied little (0.58–0.61) across reporters. Third, atomoxetine not only reduced the symptoms of ADHD, but prevented the worsening of these symptoms as well, a finding that has been seen for drug-placebo response curve analyses of other medications [5,10,11]. In contrast however, these prior drug-placebo response curve analyses have also revealed stronger effects of other medications on clinician-rated ADHD symptomatology, as evidenced by AUCs of 0.86 for Adderall [5,10,11], 0.89 for methylphenidate [5,10,11], and 0.93 for desipramine [5,10,11], as compared to the AUC of approximately 0.60 presently observed for atomoxetine.
In conclusion, we have extended the statistical results of Michelson et al. [3] by using drug-placebo response curves to describe the clinical significance of the efficacy of atomoxetine in the treatment of ADHD among adults. Our method of data presentation provides readers and clinicians with a means of understanding the nature of the effects of this drug, and the degree to which they are clinically relevant. Rather than collapsing individual responses into means or single rates of response, the drug-placebo response curve illustrates clinically meaningful details that often are lost in a standard analysis, such as the ability of atomoxetine to improve outcome and prevent worsening throughout the full range of outcome scores. The present drug-placebo response analysis provided strong support for the efficacy of atomoxetine relative to placebo for reducing inattention, hyperactivity, and total ADHD symptoms assessed by a variety of reporters, and for preventing the worsening of these symptoms. The finding that atomoxetine is efficacious through the full range of outcome further emphasizes the clinical value of treating ADHD adults with this medication.
Methods
Subjects
Two identical randomized, double-blind, placebo-controlled studies were conducted concurrently at 17 (Study I) and 14 (Study II) outpatient sites in North America. Each site's institutional review board evaluated and approved the study protocol, and written informed consent was obtained from each patient. Adults who met DSM-IV criteria for ADHD as assessed by clinical interview and confirmed by the Conners' Adult ADHD Diagnostic Interview for DSM-IV were recruited from clinics and by advertisement. Patients were required to have at least moderate symptom severity, and the diagnosis had to be corroborated by a second reporter for either current symptoms (by a significant other) or childhood symptoms (by a parent or older sibling). Patients who met diagnostic criteria for any other Axis-I disorder using the Structured Clinical Interview for DSM-IV were excluded, as were patients with serious medical illness or habitual substance abuse.
Atomoxetine and Placebo Administration
Following an initial one-week medication washout and evaluation period, patients entered a two-week placebo lead-in phase. Patients who maintained the initial severity criteria required for study entry were randomized to receive atomoxetine or placebo for a 10-week period. Atomoxetine was administered in evenly divided doses in the morning and late afternoon/early evening beginning at a total daily dose of 60 mg. Patients with residual symptoms received higher doses of up to 90 mg/day after two weeks and 120 mg/day after four weeks. If patients developed problems tolerating this regimen, the dose could be decreased to the last tolerated dose or an increase in dosage could be omitted. Across both studies, 270 subjects received atomoxetine, while 263 subjects received placebo. Of these, 197 completed acute treatment with atomoxetine, while 211 placebo-treated subjects completed the trial, a difference that was not significant.
Outcome Measures
The outcome measures examined in this study were derived from the CAARS and the CGI. A clinician completed the CGI before and after the treatment regimen, while both the subject and an investigator completed the CAARS before and after treatment. The three groups of primary dependent measures of this study included: 1.) clinician-rated CGI ADHD Severity change scores and endpoint scores; 2.) investigator-rated inattention, hyperactivity/impulsivity, total symptoms, and ADHD index scores on the CAARS; and 3.) self-rated inattention, hyperactivity/impulsivity, total symptoms, and ADHD index scores on the CAARS.
Drug-Placebo Response Curve Analysis
The rationale and methodology for drug-placebo response curve analysis methods are described in detail by Faraone et al. [5] The goal of response curve analysis is not to demonstrate statistically significant group differences; rather, this method provides an alternative means of displaying differences that have already been demonstrated to be statistically significant. Thus, it does not replace a standard statistical analysis, but augments that analysis by showing the clinical significance of drug effects. For the present study, the use of drug-placebo response curve analysis is warranted, as the statistically significant effects of atomoxetine on reducing symptoms of ADHD in adults have been documented previously.
The drug-placebo response curve is constructed in the following six steps: 1.) Choose an outcome variable, for example the change in CAARS Inattention score from baseline to the end of the study; 2.) At each observed score, calculate separately for the drug and placebo groups the proportion of subjects having that score or a better score. For CAARS change scores, therapeutic change is indicated by negative numbers, i.e., a decrease in the symptom score; 3.) For each observed score, plot these proportions for the drug group on the vertical axis against the proportions computed for the placebo group on the horizontal axis; 4.) Connect the plotted points and label those that correspond to the best response, the 25th percentile of response, the median response, the 75th percentile of response and the worst response; 5.) If the outcome variable is a change score, also label the point corresponding to no change; 6.) Plot the line of no effect, which is the diagonal line from the [0, 0] point to the [1,1] point. Each point along a curve represents an observed outcome score on that measure, and the points on each plot are then connected by line segments. The line of no effect comprises all points for which the proportion of subjects who respond to drug is the same as the proportion who respond to placebo.
The drug-placebo response curve is a graphical method of describing results from a clinical trial, not a statistical test. It is most sensibly used to describe an effect that has been demonstrated with appropriate statistical tools. Nevertheless, the drug-placebo response curve's roots in (ROC) analysis motivate the computation of one statistic, the AUC, which is computed through integration. The area under the drug-placebo response curve ranges from 0.5 (when the drug effect equals the placebo effect) to 1.0 (when the drug is completely effective and the placebo has no effect). The AUC is a useful index of clinical significance because it equals the probability that a randomly selected member of the drug group will have a better result than a randomly selected member of the placebo group [12,13], i.e., the probability that drug will outperform placebo.
In summary, the placebo-response curve provides four pieces of clinically relevant data not typically available from traditional statistical analyses of outcomes data. First, the effect size of a drug on an outcome measure can be determined as the distance between the curve and the line of no effect at any given cut-point. Second, the ratio of drug responders to placebo responders across the range of outcomes can be determined as the area under the curve. Third, the likelihood of a drug to elicit a specific outcome (e.g., a clinically meaningful cut-point) can be determined as the proportion of drug-responders to placebo-responders at any given cut-point. Fourth, the ability of a drug to improve functioning vs. prevent worsening of functioning can be determined as the proportion of drug-responders to placebo-responders at the outcome score representing no change.
Competing interests
Stephen V. Faraone, PhD.- Stephen Faraone receives research funding from Lilly, McNeil and Shire.
Joseph Biederman, MD.- Joseph Biederman receives research support from the following sources: Shire Laboratories, Inc and Eli Lilly & Company, Pfizer Pharmaceutical, Cephalon Pharmaceutical,, Janssen Pharaceutical, Neurosearch. Pharmaceuticals, Stanley Medical Institute, Lilly Foundation, Prechter Foundation, NIMH, NICHD and NIDA
Dr. Joseph Biederman is a speaker for the following speaker's bureaus: Eli Lilly & Company, Pfizer Pharmaceutical, Novartis Pharmaceutical, Wyeth Ayerst, Shire Laboratories Inc, McNeil Pharmaceutical, and Cephalon Pharmaceutical
Dr. Joseph Biederman is on the advisory board for the following pharmaceutical companies: Eli Lilly & Company, CellTech, Shire Laboratories Inc, Novartis Pharmaceutical, Noven Pharmaceutical, McNeil Pharmaceuticals, Janssen, Johnson & Johnson, Pfizer, and Cephalon Pharmaceuticals
Thomas Spencer, MD- Dr. Thomas Spencer receives research support from the following sources: Shire Laboratories, Inc and Eli Lilly & Company, Glaxo-Smith Kline, Pfizer Pharmaceutical, McNeil Pharmaceutical, Novartis Pharmaceutical, and NIMH
Dr. Thomas Spencer is a speaker for the following speaker's bureaus: Glaxo-Smith Kline, Eli Lilly & Company, Novartis Pharmaceutical, Wyeth Ayerst, Shire Laboratories Inc, McNeil Pharmaceutical
Dr. Thomas Spencer is on the advisory board for the following pharmaceutical companies: Shire Laboratories, Inc and Eli Lilly & Company, Glaxo-Smith Kline, Pfizer Pharmaceutical, McNeil Pharmaceutical, and Novartis Pharmaceutical
David Michelson, MD- David Michelson is a Lilly employee
Lenard Adler, MD- Lenard Adler receives grant and Research Support, is a Consultant or on Advisory Boards: Abbott Laboratories, Bristol-Myers Squibb, Eli Lilly and Co., McNeil/Johnson & Johnson, Merck & Co., Inc., Neurosearch, Novartis Pharmaceuticals Corp., Pfizer Labs, Cortex Pharmaceuticals, Cephalon and Shire Pharmaceuticals
Fred Reimherr, MD-Fred Reimherr has been part of Lilly advisory board.
Stephen J Glatt, PhD- Stephen Glatt has no conflicts of interest to declare.
Authors' contributions
Stephen V. Faraone, PhD- Steve Faraone contributed to the analysis and interpretation of the data, the drafting and revision of the manuscript.
Joseph Biederman, MD- Joseph Biederman contributed to the analysis and interpretation of the data, the drafting and revision of the manuscript.
Thomas Spencer, MD- Thomas Spencer contributed to the conception and design of the study, the acquisition of data, interpretation of data and drafting/reviewing the manuscript.
Lenard Adler, MD- Lenard Adler contributed to the conception and design of the study, the acquisition of data, interpretation of data and drafting/reviewing the manuscript.
David Michelson, MD- David Michelson contributed to the study conception design, the data acquisition as well as critical reviewing of the manuscript.
Fred Reimherr, MD- Fred Reimherr contributed to the study conception design, the data acquisition as well as critical reviewing of the manuscript.
Stephen J Glatt, PhD- Stephen Glatt contributed to the analysis and interpretation of the data, and the drafting and revision of the manuscript.
Acknowledgements
This work was supported in part by a grant from Eli Lilly (Dr. Faraone, PI)
==== Refs
Fuller R Wong D Effects of antidepressants on uptake and receptor systems in the brain Prog Neuro-Psychopharmacol & Biol Psychiat 1985 9 485 490 10.1016/0278-5846(85)90006-5
Spencer T Biederman J Wilens T Prince J Hatch M Jones J Harding M Faraone SV Seidman L Effectiveness and tolerability of tomoxetine in adults with attention deficit hyperactivity disorder American Journal of Psychiatry 1998 155 693 695 9585725
Michelson D Adler L Spencer T Reimherr FW West SA Allen AJ Kelsey D Wernicke J Dietrich A Milton D Atomoxetine in adults with ADHD: two randomized, placebo-controlled studies Biol Psychiatry 2003 53 112 120 12547466 10.1016/S0006-3223(02)01671-2
Wernicke JF Kratochivil C Safety profile of atomoxetine in the treatment of children and adolescents with ADHD Journal of Clinical Psychiatry 2002 63 50 55 12562062
Faraone SV Biederman J Spencer TJ Wilens TE The drug-placebo response curve: a new method for assessing drug effects in clinical trials Journal of Clinical Psychopharmacology 2000 20 673 679 11106140 10.1097/00004714-200012000-00014
McNeil BJ Hanley JA Statistical approaches to the analysis of receiver operating characteristic (ROC) curves Med Dec Making 1984 4 137 150
Swets JA Sensitivities and specificities of diagnostic tests JAMA 1982 248 548 549 7097897 10.1001/jama.248.5.548
Siegel B Vukicevic J Spitzer RL Using signal detection methodology to revise DSM-III-R: Re-analysis of the DSM-III-R national field trials for autistic disorder Journal of Psychiatric Research 1990 24 293 311 2090828 10.1016/0022-3956(90)90002-8
Siegel B Vukicevic J Elliott GR Kraemer HC The use of signal detection theory to assess DSM-III-R criteria for autistic disorder J Am Acad Child Adolesc Psychiatry 1989 28 542 548 2768150
Faraone SV Pliszka SR Olvera RL Skolnik R Biederman J Efficacy of adderall and methylphenidate in attention-deficit/hyperactivity disorder: a drug-placebo and drug-drug response curve analysis Journal of Child and Adolescent Psychopharmacology 2001 11 171 180 11436957 10.1089/104454601750284081
Faraone SV Short E Biederman J Findling RL Roe CM Manos M Efficacy of Adderall and Methylphenidate in attention deficit hyperactivity disorder: a drug-placebo and drug-drug response curve analysis of a naturalistic study. International Journal of Neuropsychopharmacology 2002 5 121 129 12135536 10.1017/S1461145702002845
Colditz GA Miller JN Mosteller F Measuring gain in the evaluation of medical technology. The probability of a better outcome Int J Technol Assess Health Care 1988 4 637 642 10291102
Hanley JA McNeil BJ The meaning and use of the area under a receiver operating characteristic (ROC) curve Radiology 1982 143 29 36 7063747
|
16202140
|
PMC1262689
|
CC BY
|
2021-01-04 16:39:21
|
no
|
Behav Brain Funct. 2005 Oct 3; 1:16
|
utf-8
|
Behav Brain Funct
| 2,005 |
10.1186/1744-9081-1-16
|
oa_comm
|
==== Front
Behav Brain FunctBehavioral and brain functions : BBF1744-9081BioMed Central London 1744-9081-1-171621266310.1186/1744-9081-1-17Short PaperBranching projections of ventrolateral reticular neurons to the medial preoptic area and lumbo-sacral spinal cord Russo Antonella [email protected] Rosalia [email protected] Rosa [email protected] Stefania [email protected] André [email protected] Department of Physiological Sciences, University of Catania, Catania, Italy2 Institute of Neurological Science, Research National Council, Catania, Italy3 Department of Anatomy, Diagnostic Pathology, Phorens Medicine, Hygiene and Public Health, University of Catania, Catania, Italy4 Laboratoire de Physiologie Neurovégétative, UMR 6153-CNRS 1147-INRA, Université Aix-Marseille III, Faculté des Sciences St. Jerôme, Marseille, France2005 7 10 2005 1 17 17 24 6 2005 7 10 2005 Copyright © 2005 Russo et al; licensee BioMed Central Ltd.2005Russo et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Different findings indicate that rostral ventrolateral reticular nucleus (RVL) is neuronal substrate of integration and regulation of the cardiovascular functions. Some efferent RVL neurons project to the thoraco-lumbar spinal cord and excite preganglionic sympathetic neurons, to the spinal phrenic motor neurons involved in inspiratory function and increase the activity of vasoconstrictor fibres innervating blood vessels in the skin and skeletal muscle. Our study was aimed at revealing presence of neurons within RVL supplying branching collateral input to the medial preoptic area (MPA) and to the lumbo-sacral spinal cord (SC-L) in the rat. All animal experiments were carried out in accordance with current institutional guidelines for the care and use of experimental animals. We have employed double fluorescent-labelling procedure: the projections were defined by injections of two retrograde tracers: Rhodamine Labelled Bead (RBL) and Fluoro Gold (FG) in the MPA and SC-L, respectively. Our results showed the presence of few single FG neurons and single RBL neurons in the RVL. The size of FG-neurons and RBL-neurons was medium (25–30 μm) and large (50 μm).
Few double-projecting neurons were distributed in the middle third of RVL nucleus, their size was 30–40 μm. The results demonstrate that pools of neurons in the RVL have collateral projections to the MPA and SC-L and they are involved in ascending and descending pathway. These data suggest that these neurons could play a role in maintaining activity of central and peripheral blood flow.
ventrolateral reticular nucleus (RVL)retrograde fluorescent tracermedial preoptic areaspinal cordrat.
==== Body
Introduction
Several findings indicate that rostral ventrolateral reticular medulla (RVL) is the fundamental neuronal substrate of the regulation of circulation and cardiovascular functions, that mediates vasomotor reflexes [1,2]. Different methods permitted to identify afferent projections to RVL, from the hypothalamic paraventricular nucleus [3], the lateral hypothalamic area [4], the dorsal raphe nucleus [5] and from other regions [4]. Some efferent RVL neurons project to the intermediolateral cell column of the thoraco-lumbar spinal cord and excite preganglionic sympathetic neurons [6-8], spinal phrenic motor neurons involved in inspiratory function [9] and increase the activity of vasoconstrictor fibres innervating blood vessels in the skin and skeletal muscle [10]. Ascending RVL efferent projections convey informations from RVL to diencephalic nuclei [4,11] and collateralized fibres were found in RVL, after injections within the cerebellar fastigial nucleus and the superior colliculus [12]. In fact, electrical stimulation of RVL in rats elicits an excitatory response on the renal sympathetic nerve activity [7,13]. The MPA is also involved in controlling of numerous functions: neuroendocrine, sexual, maternal behaviour and other activities [14-16]. The aim of the present study is to demonstrate the existence of direct projections from RVL to the MPA and, via collaterals, the existence of direct projections to MPA and SC-L. In addition, previous work has shown that RVL adrenergic (C1) and non-adrenergic neurons were found to participate in these projections [17-19]. The brainstem neurons may be involved in simultaneous transmission of autonomic-related signals, in fact catecholaminergic and non-catecholaminergic neurons were found to provide branching collaterals to the central nucleus of the amygdala and to the hypothalamic paraventricular nucleus [11].
In the present study we have used two retrograde tracers to determine the distribution of RVL neurons that project via collaterals to the MPA and SC-L.
Results
Only those cases (8 rats) where microscopic analysis of the injection sites revealed that the tracer deposits were correctly positioned were included in the study. Fig. 1 shows the injection sites of the retrograde tracers RLB in the MPA (A) and FG in the SC-L (B). The results showed a large number of retrogradely-labelled neurons in the whole ipsi- and contralateral RVL (stereotaxic planes: -11.60/-11.96). FG-labelled neurons from the ipsilateral and contralateral RVL were numerous (39.42 ± 3.3); moreover, these neurons ranged in size from medium (20–30 μm) to large (50 μm) and were rather sparse within RVL. RLB-labelled neurons were mostly packed within the boundaries of the field examined, consistently of medium size (20–35 μm) and uniformly distributed in ipsi- and contralateral RVL (38.2 ± 2.85). The fluorescence microscopy revealed a substantial number of RVL double-labelled neurons (18.6 ± 3.13), evidencing the presence of collateralization to the MPA and SC-L. These neurons generally were of medium to large size (30–40 μm) and were localized in the middle third of RVL.
Figure 1 Microphotographs show injection zones: in (A) RLB injection site and drawing in MPA (black area); Scale bar: 400 μm. In (B) FG injection site and drawing in SC-L (black area). Scale bar: 90 μm.
Fig. 2 shows schematic drawing of frontal brain section including RVL region with different symbol (circle, triangle and star) indicating the labelled neurons presence.
Figure 2 Schematic drawing atlas of frontal brain section including RVL region: circle indicates FG labelled neurons; triangle indicates RLB labelled neurons; star indicates FG-RLB labelled neurons.
A relatively low number of FG/RLB (4.6 ± 1.49 neurons) double-labelled neurons (30–40 μm) were scattered mainly at the ipsilateral RVL level; an example is showed in Fig. 3.
Figure 3 Microphotographs of double-labelled neurons FG-RLB in reticular ventrolateral nucleus (RVL). (A) cells stained positively to FG (excitation wavelength 330 nm); (B) the same cells stained positively to RLB (excitation wavelength 560 nm) indicating the existence of a collateral axon. Scale bar: 50 μm.
Discussion
The present study provides direct evidence, based on retrograde tracing technique, that: 1) RVL single neurons directly project to the medial preoptic area; 2) RVL single neurons directly project to spinal cord, confirming previous results [4]; 3) RVL neurons supply, via collaterals, branching inputs to the MPA and SC-L. Sympatho-excitatory neurons of the RVL, like in rostral ventromedial medulla [20] and in caudal ventrolateral medulla [21], are important for the maintenance of tonic levels of arterial pressure they are intrinsic pacemaker activity and discharge continuously [2]. In conclusion, after the injections of FG and RLB, we show a cluster of branching RVL neurons in the rat brain. The fact that RVL contains neurons with a collateral fibers to both the spinal cord and to hypothalamic MPA, suggests that these neurons might play a role in maintaining activity of central and peripheral blood flow simultaneously.
Materials and methods
All experiments were carried out in accordance with current institutional guidelines for the care and use of experimental animals. Experiments were performed on 10 adult male Wistar rats weighing 250–300 g (Morini, Italy), maintained under controlled conditions of room temperature (23 ± 1°C) and lighting (lights on 07:00 – 19:00 h); laboratory chow diet and water were available ad libitum; the in vivo experimental procedure was performed during daytime (10:00 – 13:00 a.m.).
Animals were anaesthetized with chloral hydrate (400 mg/Kg, i.p.). Two fluorescent tracers were injected to the same rat: Fluoro Gold (FG) was injected into the lumbo-sacral spinal cord on one side, Rhodamine Labeled Bead (RLB) into the medial preoptic area of the same side. Rats were placed in a Kopf stereotaxic frame and injected with 0.08 μl of undiluted RLB in the MPA, at the following coordinates (0.30; L = 0.5; V = -8.5) [22]; freshly dissolved FG (0.15 μl at 6% in saline) was injected into the SC-l. Both tracers were pressure-injected at a rate of 50 nl/min. using 1 μl Hamilton microsyringes.
Seven days after the injections, the animals were reanaesthetized and perfused through the ascending aorta with saline (60 ml), followed by ice-cold 4% paraformaldehyde phosphate buffer (300 ml; pH 7.4). The brains were removed, immersed in the same fixative for 3–4 h and cryoprotected overnight in phosphate-buffered with 20% sucrose solution.
Coronal sections (40 μm) were cut on a cryostat (Reichert), mounted on slides and observed under fluorescent microscope (Polyvar Reichert) for identification of injection zones. The sections were air-dried, mounted and observed with a Reichert fluorescence microscope equipped with filter combinations revealing red (RLB), yellow (FG). For each animal, three non-adjacent sections were evaluated and the labelled cells plotted onto schematic drawings of the RVL region level (stereotaxic planes: -11.60 / -11.96) [5]. Thus, cell numbers were expressed as the average number/section calculated from these three sections.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AR, RP, RR, SS and AJ jointly conceived and executed this study and helped draft the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We thank Mr. S. Bentivegna for helping with the photographic work. This study was supported by MIUR.
==== Refs
Aicher SA Milner TA Pickel VM Reis DJ Anatomical substrates for baroreflex sympathoinhibition in the rat Brain Res Bull 2000 51 107 110 10709955 10.1016/S0361-9230(99)00233-6
Reis DJ Golanov EV Ruggiero DA Sun MK Sympatho-excitatory neurons of the rostral ventrolateral medulla are oxygen sensors and essential elements in the tonic and reflex control of the systemic and cerebral circulations J Hypertens 1994 S159 180
Yang Z Coote JH Influence of the hypothalamic paraventricular nucleus on cardiovascular in the rostral ventrolateral medulla of the rat J Physiol (Lond) 1998 531 521 530 9807000 10.1111/j.1469-7793.1998.521bb.x
Dampney RAL Functional organization of central pathways regulating the cardiovascular system Physiol Rev 1994 74 323 364 8171117
Underwood MD Arango V Bakalian MJ Ruggiero DA Mann JJ Dorsal raphe nucleus serotonergic neurons innervate the rostral ventrolateral medulla in rat Brain Res 1999 824 45 55 10095041 10.1016/S0006-8993(99)01181-6
Minson JB Arnolda LF Llewellyn-Smith IJ Neurochemistry of nerve fibers apposing sympathetic preganglionic neurons activated by sustained hypotension J Comp Neurol 2002 449 307 318 12115667 10.1002/cne.10282
Comer AM Gibbons HM Qi J Kawai Y Win J Lipski J Detenction of mRNA species in bulbospinal neurons isolated from the rostral ventrolateral medulla using single-cell RT-PCR Brain Res Brain Res Prot 1999 4 367 377 10.1016/S1385-299X(99)00042-2
Rajakumar N Hrycyshyn AW Flumerfelt BA Afferent organization of the lateral reticular nucleus in the rat: an anterograde tracing study Anat and Embryol (Berl) 1992 185 25 37 1736682 10.1007/BF00213598
Sun MK Reis DJ Excitatory amino acid-mediated chemoreflex excitation of respiratory neurones in rostral ventrolateral medulla in rats J Physiol 1996 492 559 571 9019550
Dampney RAL McAllen RM Differential control of sympathetic fibres supplying hindlimb skin and muscle by subretrofacial neurones in the cat J Physiol (Lond) 1998 395 41 56 2900889
Petrov T Krukoff TL Jhamandas JH Branching projections of catecholaminergic brainstem neurons to the paraventricular hypotalamic nucleus and the central nucleus of the amygdala in the rat Brain Res 1993 609 81 92 8099526 10.1016/0006-8993(93)90858-K
Stanzani S Russo A Pellitteri R Storaci G Cataudella T Branching projections of catecholaminergic ventrolateral reticular neurons to the fastigial nucleus and superior colliculus in the rat: triple labelling procedure Neurosci Lett 2001 307 135 138 11438382 10.1016/S0304-3940(01)01899-7
Granata AR Denavit-Saubie M Bulbospinal sympathoexcitatory pathways in the rat Brain Res Bull 1994 34 601 605 7922604 10.1016/0361-9230(94)90146-5
Ding YQ Wang D Jun-Qing X Gong J Direct projections from the medial preoptic area to spinally- projecting neurons in Barrington's nucleus: an electron microscope study in the rat Neurosci Lett 1999 271 175 178 10507697 10.1016/S0304-3940(99)00562-5
Dominguez JM Hul EM Stimulation of the medial amygdala enhances medial preoptic dopamine release: implications for male rat sexual behavior Brain Res 2001 917 225 229 11640908 10.1016/S0006-8993(01)03031-1
Grattan DR Rocca MS Sagrillo CA McCarthy MM Selmanoff M Antiandrogen microimplants into rostral medial preoptic area decrease gamma-aminobutyric acidergic neuronal activity and increase luteinizing hormone secretion in the intact male rat Endocrinology 1996 137 4167 4173 8828473 10.1210/en.137.10.4167
Aicher SA Hahn B Milner TA N-methyl-D-aspartate-type glutamate receptors are found in post-synaptic targets of adrenergic terminals in the thoracic spinal cord Brain Res 2000 856 1 11 10677605 10.1016/S0006-8993(99)02145-9
Morrison SF Callaway J Milner TA Reis DJ Rostral ventrolateral medulla: a source of the glutamatergic innervation of the sympathetic intermediolateral nucleus Brain Res 1991 562 126 135 1724740 10.1016/0006-8993(91)91196-8
Sun MK Young BS Hackett JT Guyenet PG Rostral ventrolateral medullary neurons with intrinsic pacemaker properties are not catecholaminergic Brain Res 1988 451 345 349 2908028 10.1016/0006-8993(88)90781-0
Babic T Ciriello J Medullary and spinal cord projections from cardiovascular responsive sites in the rostral ventromedial medulla J Comp Neurol 2004 469 391 412 14730590 10.1002/cne.11024
Sun W Panneton WM Defining projections from the caudal pressor area of the caudal ventrolateral medulla J Comp Neurol 2005 482 273 93 15690490 10.1002/cne.20434
Paxinos G Watson C The rat brain in sterotaxis coordinates 1986 Academic Press: Sydney
|
16212663
|
PMC1262690
|
CC BY
|
2021-01-04 16:39:21
|
no
|
Behav Brain Funct. 2005 Oct 7; 1:17
|
utf-8
|
Behav Brain Funct
| 2,005 |
10.1186/1744-9081-1-17
|
oa_comm
|
==== Front
BMC BiochemBMC Biochemistry1471-2091BioMed Central London 1471-2091-6-171617658910.1186/1471-2091-6-17Research ArticleProbing stereoselective inhibition of the acyl binding site of cholesterol esterase with four diastereomers of 2'-N-α-methylbenzylcarbamyl-1, 1'-bi-2-naphthol Chiou Shyh-Ying [email protected] Cheng-Yue [email protected] Long-Yau [email protected] Gialih [email protected] Institute of Medicine and Department of Neurosurgery, Chung Shan Medical University, Taichung 402, Taiwan2 Department of Chemistry, National Chung-Hsing University, Taichung 402, Taiwan2005 22 9 2005 6 17 17 28 6 2005 22 9 2005 Copyright © 2005 Chiou et al; licensee BioMed Central Ltd.2005Chiou et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Recently there has been increased interest in pancreatic cholesterol esterase due to correlation between enzymatic activity in vivo and absorption of dietary cholesterol. Cholesterol esterase plays a role in digestive lipid absorption in the upper intestinal tract, though its role in cholesterol absorption in particular is controversial. Serine lipases, acetylcholinesterase, butyrylcholinesterase, and cholesterol esterase belong to a large family of proteins called the α/β-hydrolase fold, and they share the same catalytic machinery as serine proteases in that they have an active site serine residue which, with a histidine and an aspartic or glutamic acid, forms a catalytic triad. The aim of this work is to study the stereoselectivity of the acyl chain binding site of the enzyme for four diastereomers of an inhibitor.
Results
Four diastereomers of 2'-N-α-methylbenzylcarbamyl-1, 1'-bi-2-naphthol (1) are synthesized from the condensation of R-(+)- or S-(-)-1, 1'-bi-2-naphthanol with R-(+)- or S-(-)-α-methylbenzyl isocyanate in the presence of a catalytic amount of pyridine in CH2Cl2. The [α]25D values for (1R, αR)-1, (1R, αS)-1, (1S, αR)-1, and (1S, αS)-1 are +40, +21, -21, and -41°, respectively. All four diastereomers of inhibitors are characterized as pseudo substrate inhibitors of pancreatic cholesterol esterase. Values of the inhibition constant (Ki), the carbamylation constant (k2), and the bimolecular rate constant (ki) for these four diastereomeric inhibitors are investigated. The inhibitory potencies for these four diastereomers are in the descending order of (1R, αR)-1, (1R, αS)-1, (1S, αR)-1, and (1S, αS)-1. The k2 values for these four diastereomers are about the same. The enzyme stereoselectivity for the 1, 1'-bi-2-naphthyl moiety of the inhibitors (R > S, ca. 10 times) is the same as that for 2'-N-butylcarbamyl-1, 1'-bi-2-naphthol (2). The enzyme stereoselectivity for the α-methylbenzylcarbamyl moiety of the inhibitors is also R > S (2–3 times) due to the constraints in the acyl binding site.
Conclusion
We are the first to report that the acyl chain binding site of cholesterol esterase shows stereoselectivity for the four diastereomers of 1.
==== Body
Background
Recently there has been increased interest in pancreatic cholesterol esterase (CEase, EC 3.1.1.13) due to correlation between enzymatic activity in vivo and absorption of dietary cholesterol [1,2]. Physiological substrates include cholesteryl esters, retinyl esters, triacylglycerols, vitamin esters, and phospholipids [3-5]. CEase plays a role in digestive lipid absorption in the upper intestinal tract, though its role in cholesterol absorption in particular is controversial [1,6]. A recent report indicates that CEase is directly involved in lipoprotein metabolism, in that the enzyme catalyzes the conversion of large LDL to smaller, denser, more cholesteryl ester-rich lipoproteins, and that the enzyme may regulate serum cholesterol levels [7,8]. Serine lipases, acetylcholinesterase, butyrylcholinesterase, and CEase belong to a large family of proteins called the α/β-hydrolase fold [9,10], and they share the same catalytic machinery as serine proteases in that they have an active site serine residue which, with a histidine and an aspartic or glutamic acid, forms a catalytic triad [11,12]. The conservation of this catalytic triad suggests that as well as sharing a common mechanism for substrate hydrolysis, that is, formation of a discrete acyl enzyme species via the active site serine hydroxy group, serine proteases, CEase, and lipases may well be expected to be inhibited by the same classes of mechanism-based inhibitors such as phosphorothiolates [13], pyrones [14], fluoroketones [15], boronic acids [16], and carbamates [16-29].
The crystal structure of the active site region of pancreatic CEase [30,31] is similar to Torpedo californica acetylcholinesterase (AChE) [32], Candida rugosa lipase (CRL) [33,34], Geotrichum candidum lipase (GCL) [35], and Pseudomonas species lipase (PSL) [36,37]. Moreover, the active site of CEase like CRL, GCL, PSL, and acetylcholinesterase may consist of at least five major binding sites (Figure 1) [23,24,30,31]: (a) an acyl chain binding site (ABS) that binds to the acyl chain of the substrate and is opened by the removal of C-terminal 574–579 in which is bent in shape and contains a deep, wide hole from the evacuation of Phe579, (b) an oxyanion hole (OAH), the H-bonding peptide NH functions of Gly107, Ala108, and Ala195, that stabilizes the tetrahedral species, (c) an esteratic site or the catalytic triad (ES), comprised of Ser194-His435-Asp320, that is involved in nucleophilic attack to the substrate carbonyl group and in general acid-base catalysis, and (d) a leaving group binding site (LBS) or/and the second alkyl chain or group binding site (SACS) that binds to the cholesterol part of cholesterol ester or the second fatty acid chain of triacylglycerol and is located at the opposite direction of ABS.
Figure 1 Possible interactions for cholesteryl linoleate in the active site of CEase [30,31].
Previous work has shown that CEase is stereoselectively inhibited by the two atropisomers (or enantiomers) of 1, 1-bi-2-naphtyl carbamates due to the stereoselective binding at LBS of the enzyme [20,22]. Doorn et al. have also reported that CEase is stereoselectively inhibited by the four diastereomers of isomalathion due to stereoselectivity for both ES and LBS of the enzyme [13]. The aim of this study is to extend the stereoselectivity to the four diastereomers of inhibitors by adding two extra bonds between a chiral center and (or a chiral axis of the inhibitors. In other words, we may probe the double selectivity for both ABS and LBS of the enzyme. Thus, four diastereomers of 2'-N-α-methylbenzylcarbamyl-1, 1'-bi-2-naphthol (1), 2'-N-(R)-α-methylbenzylcarbamyl-(R)-1, 1'-bi-2-naphthol ((1R, αR)-1), 2'-N-(S)-α-methylbenzylcarbamyl-(R)-1, 1'-bi-2-naphthol ((1R, αS)-1), 2'-N-(R)-α-methylbenzylcarbamyl-(S)-1, 1'-bi-2-naphthol ((1S, αR)-1), and 2'-N-(S)-α-methylbenzylcarbamyl-(S)-1, 1'-bi-2-naphthol ((1S, αS)-1) (Figure 2), are synthesized from condensation of (R)- or (S)-1, 1'-bi-2-naphthol with (R)-or (S)-α-methylbenzyl isocyanate in the presence of pyridine in dichloromethane. The stereoselectivity of CEase inhibition by the four diastereomers of 1 is evaluated kinetically.
Figure 2 Structures of the four diastereomers of carbamates 1 and the two atropisomers of 2.
Most carbamate inhibitors are characterized as the pseudo substrate inhibitors of CEase (Figure 3) [16-29] and meet some of the criteria proposed by Abeles and Maycock [38]. First, the inhibition is time-dependent and follows pseudo-first-order kinetics; second, with increasing concentration of inhibitor the enzyme displays saturation kinetics; third, the enzyme is protected from inhibitions by carbamate by binding of a competitive inhibitor such as trifluoroacetophenone (TFA). The Ki step leads to the tetrahedral intermediate and the k2 step leads to the carbamyl enzyme intermediate. Moreover, values of Ki and k2 can be calculated from Equation 1 [16-29]:
Figure 3 Kinetic scheme for the pseudo substrate inhibition of CEase.
kapp = k2 [I]/(Ki(1+ [S]/Km)+ [I]) (1)
In Equation 1, kapp values are first-order rate constants which can be obtained as described in Hosie et al. [17]. Bimolecular rate constant, ki = k2/Ki, is related to overall inhibitory potency.
Results
For the first time, we synthesize four optical pure diastereomers of 1. (1R, αR)-1, (1R, αS)-1, (1S,αR)-1, and (1S, αS)-1 (Figure 2) are synthesized from the condensation of R-(+)- or S-(-)-1, 1'-bi-2-naphthanol with R-(+)- or S-(-)-α-methylbenzyl isocyanate in the presence of a catalytic amount of pyridine in CH2Cl2. The [α]25 Dvalues for (1R, αR)-1, (1R, αS)-1, (1S,αR)-1, and (1S, αS)-1 are +40, +21, -21, and -41°, respectively.
Like most carbamates, the four diastereomers of 1 are characterized as the pseudo substrate inhibitors of CEase (Figures 3 and 4) and meet some of the criteria proposed by Abeles and Maycock [38]. When CEase is incubated with a carbamate in the presence of TFA (2 μM), a known competitive inhibitor of CEase [22] before the inhibition reaction, the enzyme is protected from inhibition by carbamate by binding of TFA as described in Hosie et al. [17] (Figure 4B).
Figure 4 A: The kapp vs. [I] plot for inhibition of the CEase-catalyzed hydrolysis of PNPB by (1R, αR)-1. [PNPB] = 50 μM. The solid line is a least-squares fit to Eq. (1) [17]; the parameters of the fit are Ki = 0.27 ± 0.01 μM and k2 = (2.0 ± 0.2) × 10-3 s-1. B: % activity of CEase vs. the time period for inhibition of the enzyme with (1R, αR)-1 (50 nM) in the absence and presence of TFA (2 μM). [PNPB] = 50 μM. All the procedures followed those of Hosie et al. [17]. For the control experiments (squares), CEase was incubated alone at 25.0°C for a period of time before the inhibition reaction (CEase + PNPB + (1R, αR)-1). For the carbamate inhibition experiments (triangles), CEase was incubated with (1R, αR)-1 (50 nM) at 25.0°C for a period of time before the enzyme reaction (CEase + PNPB). For the protection experiments (circles), CEase was incubated with (1R, αR)-1 (50 nM) and TFA (2 μM) at 25.0°C for a period of time before the enzyme reaction (CEase + PNPB).
The inhibition data for CEase by the four diastereomers of 1 and the two enantiomers of 2 are summarized (Table 1). The stereochemical preference of CEase for the binaphthyl moiety of 1 (R > S, ca. 10 times) is the same as that for 2 [20,22]. The stereoselectivity of CEase for the α-methylbenzyl moiety of 1 is also the R-form (2–3 times over S-form).
Table 1 Inhibition constants for CEase-catalyzed hydrolysis of PNPB in the presence of the four diastereomers of 1 and the two enantiomers of 2
Inhibitor Ki(μM) k2(10-3s-1) ki(103 M-1s-1)
(1R, αR)-1 0.20 ± 0.01 2.0 ± 0.2 10 ± 1
(1R, gαS)-1 0.50 ± 0.03 2.0 ± 0.2 4.0 ± 0.4
(1S, gαR)-1 2.0 ± 0.1 2.0 ± 0.2 1.0 ± 0.1
(1S, gαS)-1 6.0 ± 0.4 1.8 ± 0.2 0.30 ± 0.03
(R)-2a 0.8 ± 0.1 10 ± 1 12 ± 2
(S)-2a 1.3 ± 0.1 6.0 ± 0.5 5.0 ± 0.6
aTaken from references [20,22].
Among the four diastereomers of 1, (1R, αR)-1 is the most potent inhibitor and its overall inhibitory potency (ki) is about the same as that of R-2 (Table 1). On the other hand, (1S, αS)-1 is the least potent inhibitor of CEase and its overall inhibitory potency is about 17-fold lower than that of S-2. All k2 values for the CEase inhibition by1 are about the same (Table 1).
Discussion
According to the X-ray crystal structure, CEase-catalyzed hydrolysis of cholesteryl linoleate has been proposed (Figure 1) [30,31]. Like most carbamates, the four diastereomers of 1 are characterized as the pseudo substrate inhibitors of CEase (Figures 3 and 4) [16-29] and meet some of the criteria proposed by Abeles and Maycock [38]. Therefore, the CEase inhibition by the four diastereomers of 1is proposed (Figure 5) [4]. In this mechanism, the α-methylbenzylcarbamyl moiety of 1 is proposed to bind to ABS of the enzyme, and the binaphthyl moiety of 1 is proposed to bind to LBS of the enzyme. The stereochemical preference of CEase for the binaphthyl moiety of 1 (1R > 1S in Table 1) at LBS of the enzyme is therefore identical to that of 2 (R > S) due to the fact that the nucleophilic attack of the Ser194 of the enzyme to the carbonyl group of the inhibitor sterically hinder from one of the naphthyl group of the inhibitors (Figure 5A) [20,22]. Since 4-nitrophenyl-N-benzyl-carbamate is a very potent pseudo substrate inhibitor of CEase [21,25,26], the benzylcarbamyl moiety of the inhibitor is believed to bind tightly to ABS of the enzyme. Similarly, the α-methylbenzylcarbamyl moiety of 1 is also believed to bind to ABS of the enzyme. The stereochemical preference of CEase for the α-methylbenzylcarbamyl moiety of 1 at ABS of the enzyme is also R > S (αR > αS in Table 1). The possible reason for this is the fact that one of the naphthyl group and the α-methyl group of (1S, αS)-1 are located at the same side of the nucleophilic attack of Ser194 when the inhibitor binds to CEase and therefore these two groups of the inhibitor sterically hinder the nucleophilic attack of Ser194 to the inhibitor (Figure 5A). On the other hand, (1R, αR)-1 does not have any hindrance for the nucleophilic attack of Ser194 (Figure 5B) and therefore (1R, αR)-1 is the most potent inhibitor among the four diastereomers of 1 (Table 1).
Figure 5 Possible interactions between the stereoisomers of 1 and CEase [31,31]. (A) CEase and (1S, αS)-1. The methyl benzyl moiety of the inhibitor binds to ABS of the enzyme. Three unfavorable repulsions (in red) from the methyl moiety and Ser194, the naphthyl moiety and Ser194, and the naphthyl moiety and His435 hinder the nucleophilic attack of Ser194 to the carbonyl group of the inhibitor. (B) CEase and (1R, αR)-1. There is no unfavorable repulsion for the nucleophilic attack of Ser194 to the carbonyl group of the inhibitor.
The stereoselectivity of CEase at ABS of the enzyme for the α-methylbenzyl group of 1 (R > S) (Table 1) is the same as that of CRL at its ABS for 2-methyl-6-(2-thienyl) hexanate [39]. For the Ki step (Figure 3), (1R, αR)-1 and (1S, αR)-1 bind to CEase 2.5 and 3 times more tightly than (1R, αS)-1 and (1S, αS)-1, respectively. The Ki value with regard to the chiral center at the α-position of 1 is quite low compared to that with regard to the binaphthol chiral axis of 1 (Table 1) [20,22] and to that with regard to the phosphorus chiral center of isomalathion [13]. Therefore, we propose that ABS of CEase does not show high selectivity for the chiral acyl group due to a narrow and hydrophobic binding pocket for ABS [30,31], which selectively and tightly binds to the benzyl phenyl moiety of the inhibitor and results in the discrimination of stereoselectivity by either the hydrogen atom or the methyl group at the α-position of the four diastereomers of 1 (Figure 5).
(1R, αR)-1 and (1R, αS)-1 are bound to CEase 10 and 12 times more tightly than (1S, αR)-1 and (1S, αS)-1, respectively (Table 1); however, R-2 is bound to CEase only 1.6 times more tightly than S-2 [20,22]. The possible reason is that the binding of the phenyl moiety of the α-methylbenzylcarbamyl group of 1 to ABS (Figure 5) constrains the binaphthol moiety of 1 to a more favorable conformation to bind with LBS, on the other hand, the n-butyl carbamyl of 2 has lots of room to "breathe" in ABS and therefore the binaphthol moiety of2 has many conformations and results in loosely binding to LBS.
The k2 values for the four diastereomers of 1 are about the same. This means that the k2 step is insensitive to the stereochemistry of 1. In other words, the stereoselectivity of CEase for (1R, αR)-1 primarily results from the Ki step. The k2 values for all diastereomers of1 are lower than those for the two atropisomers of 2 (Table 1). The possible reason is that the n-butylcarbamyl enzyme from both atropisomers of 2 is relatively more stable than the α-methylbenzylcarbamyl enzymes from the four diastereomers of 1.
Overall, we report that CEase has two stereoselective binding sites at LBS and ABS for the four diastereomers of 1. CEase [13], Chromobacterium viscosum lipase, and Rhizopus oryzal lipase [40] also show two stereoselective binding sites at LBS and ES for organic phosphorus compounds. Therefore, it is possible that CEase and lipase may contain totally three stereoselective binding sites at ABS, ES, and LBS for the six diastereomers of substrates or inhibitors.
Conclusion
Four diastereomers of 1 are synthesized and characterized as the pseudo substrate inhibitors of pancreatic cholesterol esterase. The inhibitory potencies for these four diastereomeric inhibitors are in the descending order of (1R, αR)-1, (1R, αS)-1, (1S, αR)-1, and (1S, αS)-1. The enzyme stereospecificity toward the 1, 1'-bi-2-naphthyl moiety of the inhibitors is the R-form and is the same as that for 2. The enzyme stereospecificity toward the α-methylbenzylcarbamyl moiety of the inhibitors is also R-form. For the first time, we observe that the acyl binding site of cholesterol esterase shows stereospecificity for diastereomeric inhibitors.
Methods
Materials
Porcine pancreatic CEase (ca. 70% pure since the observed Km value for this enzyme catalyzed hydrolysis of PNPB is 1.4 times higher than that for the pure enzyme [17]) and PNPB were obtained from Sigma; TFA and other chemicals were obtained from Aldrich. Silica gel used in liquid chromatography (Licorpre Silica 60, 200–400 mesh), medium pressure liquid chromatography column (LiChroprep Si 60) and thin layer chromatography plates (Kieselgel 60 F254) were obtained from Merck. An UV lamp as well as an UV detector (Linear UV-106 or ISCO UA-6) was used in detection. Hexane-ethyl acetate solvent gradient was used in liquid chromatography and medium pressure liquid chromatography. Other chemicals were of the highest quality available commercially. Carbamates 2 were synthesized as described before [20,22].
Instrumental methods
1H and 13C NMR spectra were recorded at 300 and 75.4 MHz (Varian-VXR 300 spectrometer), respectively. The 1H and 13C NMR chemical shifts were referred to internal Me4Si. UV spectra were recorded on an UV-visible spectrophotometer (Hewlett Packard 8452A or Beckman DU-650) with a cell holder circulated with a water bath. High resolution mass spectra were recorded at 70 eV on a Joel JMS-SX/SX-102A mass spectrophotometer. Elemental analyses were preformed on a Heraeus instrument.
Synthesis of four diastereomers of 1
(1R, αR)-1, (1R, αS)-1, (1S, αR)-1, and (1S, αS)-1 (Figure 2) were prepared from the condensation of R-(+)- or S-(-)-α-methylbenzyl isocyanate ([α]20 D = +10° or -10°) with 1 equivalent of R-(+)- or S-(-)-1, 1'-bi-2-naphthol ([α]20 D = +34° or -34°) in the presence of a catalytic amount of pyridine in CH2Cl2 at 25°C for 24 h (80–95 % yield). All products were purified by liquid chromatography or medium pressure liquid chromatography (silica gel, hexane-ethyl acetate) and characterized by 1H and 13C NMR spectra and high resolution mass spectra.
(1R, αR)-1, (1R, αS)-1, (1S, αR)-1, and (1S, αS)-1: 1H NMR (CDCl3, 300 MHz) δ/ppm 1.02 (d, J = 6.6 Hz, 3H, CH(Ph)CH3), 4.48 (quintet, J = 7 Hz, 1H, CH(Ph)CH3), 5.27 (d, J = 8.1 Hz, 1H, NH), 7.07–8.06 (m, 17H, aromatic H); 13C NMR (CDCl3, 75.4 MHz) δ/ppm 21.88 (CH3), 50.36 (CH(Ph)CH3), 122.45, 123.51, 125.43, 125.69, 126.08, 126.48, 126.60, 127.10, 127.24, 127.91, 128.18, 128.37, 128.53, 129.40, 131.40, 133.30, 133.41, 142.98, and 147.20 (aromatic Cs), 153.91 (C = O); High resolution mass spectra: Found: 433.1674; C29H23NO3 requires 433.1678. [α]25 D = +40, +21, -21, and -41° for (1R, αR)-1, (1R, αS)-1, (1S, αR)-1, and (1S, αS)-1, respectively. The stability of these compounds is very high at -20°C (no significant change for the optical rotation in 1 month).
Enzyme kinetics and data reduction
All kinetic data were obtained by using an UV-visible spectrophotometer that was interfaced to a computer. Microcal Origin (version 6.0) was used for all least squares curve fittings. The CEase inhibition was assayed as described in Hosie et al. [17]. The temperature was maintained at 25.0°C by a refrigerated circulating water bath. All reactions were performed in sodium phosphate buffer (pH 7.0) containing NaCl (0.1 M), acetonitrile (2% by volume), substrate PNPB (50 μM), triton X-100 (0.5 % by weight) and varying concentration of inhibitors (from 0.1 to 10 μM). The Km value for CEase-catalyzed hydrolysis of PNPB was calculated to be 140 ± 10 μM from the Michaelis-Menten equation. Requisite volumes of stock solution of substrate and inhibitors in acetonitrile were injected into reaction buffers via a pipet. CEase was dissolved in sodium phosphate buffer (0.1 M, pH 7.0). Reactions were initiated by injecting enzyme and monitored at 410 nm on the UV-visible spectrometer. First-order rate constants (the kapp values) for inhibition of CEase were determined as described by Hosie et al. [17] Values of Ki and k2 can be obtained by the parameters of non-linear least squares curve fittings of kapp vs. [I] plot to Equation (1) (Figure 4A). Duplicate sets of data were collected for each inhibitor concentration.
List of abbreviations used
ABS, acyl chain binding site; AChE, acetylcholinesterase, BChE, butyrylcholinesterase; CEase, cholesterol esterase; CRL, Candida rugosa lipase; ES, catalytic or esteratic site; GCL, Geotrichum candidum lipase; kapp, first-order rate constants; k2, carbamylation constants; ki, bimolecular rate constant; LHIS, leaving group hydrophilic binding site; LBS, leaving group binding site; 2'-N-(R)-α-methylbenzylcarbamyl-(R)-1, 1'-bi-2-naphthol ((1R, αR)-1); 2'-N-(S)-α-methylbenzylcarbamyl-(R)-1, 1'-bi-2-naphthol ((1R, αS)-1); 2'-N-(R)-α-methylbenzylcarbamyl-(S)-1, 1'-bi-2-naphthol ((1S, αR)-1); 2'-N-(S)-α-methylbenzylcarbamyl-(S)-1, 1'-bi-2-naphthol ((1S, αS)-1); OAH, the oxyanion hole; PSL, Pseudomonas species lipase; PNPB, p-nitrophenyl butyrate; PSL, Pseudomonas species lipase; SACS, the second acyl chain binding site; TFA, trifluoroacetophenone.
Authors' contributions
SYC carried out the enzyme kinetic studies. CYL participate in the synthesis of 4 diastereomers of carbamate inhibitors. LYL participated in the design of some parts of the study. GL drafted the manuscript and designed most parts of the study. All authors read and approved the final manuscript.
Acknowledgements
The authors thank the National Science Council of Taiwan for financial support.
==== Refs
Hui DY Molecular biology of enzymes involved with cholesterol esterase hydrolysis in mammalian tissues Biochim Biophys Acta 1996 1303 1 14 8816847
Lopez-Candales A Bosner MS Spilburg CA Lange LG Cholesterol transport function of pancreatic cholesterol esterase: directed sterol uptake and esterification in Enterocytes Biochemistry 1993 32 12085 12089 8218286 10.1021/bi00096a019
Brockerhoff H Jensen RG Cholesterol esterase Lipolytic Enzymes 1974 New York: Academic Press
Wang C-S Hartsuck JA Bile salt-activated lipase. A multiple function lipolytic enzymes Biochim Biophys Acta 1993 1166 1 19 8431483
Quinn DM Sutton LD Stout JS Calogeropoulou T Wiemer DF Phospholipase A2 activity and catalytic mechanism of pancreatic cholesterol esterase Phosphorus Sulfur Silicon 1990 51/52 43 46
Howles PN Carter CP Hui DY Dietary free and esterified cholesterol absorption in cholesterol esterase (bile salt-stimulated lipase) gene-targeted mice Biol Chem 1996 271 7196 7202 10.1074/jbc.271.12.7196
Miura S Chiba T Mochizuki N Nagura H Nemoto K Tomita I Ikeda M Tomita T Cholesterol-mediated changes of neutral cholesterol esterase activity in macrophages. Mechanism for mobilization of cholesterol esteryl esters in lipid droplets by HDL Arterioscler Thromb Vasc Biol 1997 17 3033 3040 9409290
Brodt-Eppley J White P Jenkins S Hui DY Plasma cholesterol esterase level is a determinant for an atherogenic lipoprotein profile in normolipidemic human subjects Biochim Biophys Acta 1995 1272 69 72 7548236
Cygler M Schrag JD Sussman JL Harel M Silman I Gentry MK Doctor BP Relationship between sequence conservation and three-dimensional structure in a large family of esterases, lipases, and related proteins Protein Sci 1993 2 366 382 8453375
Ollis DL Cheah E Cygler M Dijkstra B Frolow F Franken SM Harel M Remington SJ Silman I Schrag JD Sussman JL Vershueren KHG Goldman A The α/β hydrolase fold Protein Eng 1992 5 197 211 1409539
Svendsen A Woolley P, Petersen SB Sequence comparison with the lipase family Lipases, Their Structure Biochemistry and Application 1994 Cambridge :Cambridge University Press 1 21
Svendsen A Lipase protein engineering Biochim Biophys Acta 2000 1543 223 238 11150608
Doorn JA Talley TT Thompson CM Richardson RJ Probing the active sites of butyrylcholinesterase and cholesterol esterase with isomalathion: conserved Stereoselective inactivation of serine hydrolases structurally related to acetylcholinesterase Chem Res Toxicol 2001 14 807 813 11453726 10.1021/tx015501s
Deck LM Baca ML Salas SL Hunsaker LA Vander Jagt DL 3-Alkyl-6-chloro-2-pyrones: selective inhibitors of pancreatic cholesterol esterase J Med Chem 1999 42 4250 4256 10514295 10.1021/jm990309x
Sohl J Sutton LD Burton DJ Quinn DM Haloketone transition state analog inhibitors of cholesterol esterase Biochim Biophys Res Comm 1988 151 554 560
Feaster SR Quinn DM Mechanism-based inhibitors of mammalian cholesterol esterase Methods Enzymol 1997 286 231 252 9309653
Hosie L Sutton LD Quinn DM p-Nitrophenyl and cholesteryl-N-alkyl carbamates as inhibitors of cholesterol esterase J Biol Chem 1987 262 260 264 3793726
Feaster SR Lee K Baker N Hui DY Quinn DM Molecular recognition by cholesterol esterase of active site ligands: structure-reactivity effects for inhibition by aryl carbamates and subsequent carbamylenzyme turnover Biochemistry 1996 35 16723 16734 8988009 10.1021/bi961677v
Lin G Lai C-Y Hammett analysis of the inhibition of pancreatic cholesterol esterase by substituted phenyl-N-butylcarbamate Tetrahedron Lett 1995 36 6117 6120 10.1016/0040-4039(95)01233-8
Lin G Liu H-C Tsai Y-C Atropisomeric carbamoyl type inhibitors of pancreatic cholesterol esterase Bioorg Med Chem Lett 1996 6 43 46 10.1016/0960-894X(95)00549-9
Lin G Lai C-Y Linear free energy relationships of the inhibition of pancreatic cholesterol esterase by 4-nitrophenyl-N-alkylcarbamate Tetrahedron Lett 1996 37 193 196 10.1016/0040-4039(95)02126-4
Lin G Tsai Y-C Liu H-C Liao W-C Chang C-H Enantiomeric inhibitors of cholesterol esterase and acetylcholinesterase Biochim Biophys Acta 1998 1388 161 174 9774723
Lin G Shieh C-T Tsai Y-C Hwang C-I Lu C-P Structure-reactivity probes for active site shapes of cholesterol esterase by carbamate inhibitors Biochim Biophys Acta 1999 1431 500 511 10350625
Lin G Shieh C-T Ho H-C Chouhwang J-Y Lin W-Y Lu C-P Structure-reactivity relationships for the inhibition mechanism at the second alkyl chain binding site of cholesterol esterase and lipase Biochemistry 1999 38 9971 9981 10433704 10.1021/bi982775e
Lin G Lai C-Y Liao W-C Kao B-H Lu C-P Structure-reactivity relationships as probes for the inhibition mechanism of cholesterol esterase by aryl carbamates. I. Steady-state kinetics J Chin Chem Soc 2000 47 489 500
Lin G The Hammett-Taft cross-interactions for the inhibition mechanism of cholesterol esterase by substituted phenyl-N-substituted carbamates J Phys Org Chem 2000 13 313 321 10.1002/1099-1395(200006)13:6<313::AID-POC246>3.0.CO;2-F
Lin G Liao W-C Chiou S-Y Quantitative structure-activity relationships for the pre-steady-state inhibition of cholesterol esterase by 4-nitrophenyl-N-substituted carbamates Bioorg Med Chem 2000 8 2601 2607 11092545 10.1016/S0968-0896(00)00196-6
Lin G Liu Y-C Wu Y-G Ortho effects and cross interaction correlations for the mechanism of cholesterol esterase inhibition by aryl carbamates J Phys Org Chem 2004 17 707 714 10.1002/poc.740
Lin G Lee Y-R Liu Y-C Wu Y-G Ortho effects for inhibition mechanisms of butyrylcholinesterase by o-substituted phenyl N-butylcarbamates and comparison with acetylcholinesterase, cholesterol esterase, and lipase Chem Res Toxicol 2005
Wang X Wang C-S Tang J Dyda F Zhang XC The crystal structure of bovine bile salt activated lipase: insights into the bile salt activation mechanism Structure 1997 5 1209 1218 9331420 10.1016/S0969-2126(97)00271-2
Chen JC-H Miercke LJW Krucinski J Starr JR Saenz G Wang X Spilburg CA Lange LG Ellsworth JL Stroud RM Structure of bovine pancreatic cholesterol esterase at 1.6Å: novel structural features involved in lipase activation Biochemistry 1998 37 5107 5117 9548741 10.1021/bi972989g
Sussman JL Harel M Frolow F Oefner C Goldman A Toker L Silman I Atomic structure of acetylcholinesterase from Torpedo californica: a prototypic acetylcholine-binding protein Science 1991 253 872 879 1678899
Grochulski P Li Y Schrag JD Bouthillier F Bouthillier F Smith P Harrison D Rubin B Cygler M Insights into interfacial activation from an 'open' structure of Candida rugosa lipase J Biol Chem 1993 268 12843 12847 8509417
Grochulski P Bouthillier F Kazlauskas RJ Serreqi AN Schrag JD Ziomek E Cygler M Analogs of reaction intermediates identify a unique substrate binding site in Candida rugosa lipase Biochemistry 1994 33 3494 3500 8142346 10.1021/bi00178a005
Schrag JD Li Y Cygler M 1.8Å refined structure of the lipase from Geotrichum candidum J Mol Biol 1993 230 575 591 8464065 10.1006/jmbi.1993.1171
Schrag JD Li Y Cygler M Lang D Burgdorf T Hecht H-J Schmid R Schomburg D Rydel TJ Oliver JD Strickland LC Dunaway M Larson SB Day J McPherson A The Open Conformation of a Pseudomonas Lipase Structure 1997 5 187 202 9032074 10.1016/S0969-2126(97)00178-0
Lang DA Mannesse MLM De Haas GH Verheij HM Dijkstra BW Structure Basis of the Chiral Selectivity of Pseudomonas cepacia Lipase Eur J Biochem 1998 254 333 340 9660188 10.1046/j.1432-1327.1998.2540333.x
Abeles RH Maycock AL Suicide enzyme inactivators Acc Chem Res 1976 9 313 319 10.1021/ar50105a001
Berglund P Holmquist M Hult K Reversed enantiopreference of Candida rugosa lipase supports different modes of binding enantiomers of a chiral acyl donor J Mol Cat B Enzym 1998 5 283 287 10.1016/S1381-1177(98)00095-2
Stadler P Zandonella G Haalck L Spener F Hermetter A Paltauf F Inhibition of microbial lipases with stereoisomeric triglycerol analog phosphonates Biochim Biophys Acta 1996 1304 229 244 8982269
|
16176589
|
PMC1262691
|
CC BY
|
2021-01-04 16:26:24
|
no
|
BMC Biochem. 2005 Sep 22; 6:17
|
utf-8
|
BMC Biochem
| 2,005 |
10.1186/1471-2091-6-17
|
oa_comm
|
==== Front
BMC BiochemBMC Biochemistry1471-2091BioMed Central London 1471-2091-6-191619119110.1186/1471-2091-6-19Research ArticleBiochemical characterization of Cdk2-Speedy/Ringo A2 Cheng Aiyang [email protected] Shannon [email protected] Philipp [email protected] Mark J [email protected] Department of Molecular Biophysics and Biochemistry, Yale University School of Medicine, 333 Cedar Street, New Haven, CT 06520-8024, USA2 Mouse Cancer Genetics Program, National Cancer Institute, Frederick, MD 21702-1201, USA3 Department of Biological Sciences, University of Rhode Island, Kingston, RI 02881, USA2005 28 9 2005 6 19 19 31 3 2005 28 9 2005 Copyright © 2005 Cheng et al; licensee BioMed Central Ltd.2005Cheng et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Normal cell cycle progression requires the precise activation and inactivation of cyclin-dependent protein kinases (CDKs), which consist of a CDK and a cyclin subunit. A novel cell cycle regulator called Speedy/Ringo shows no sequence similarity to cyclins, yet can directly bind to and activate CDKs. Speedy/Ringo proteins, which bind to and activate Cdc2 and Cdk2 in vitro, are required for the G2 to M transition during Xenopus oocyte maturation and for normal S-phase entry in cultured human cells.
Results
We have characterized the substrate specificity and enzymatic activity of human Cdk2-Speedy/Ringo A2 in order to gain insights into the possible functions of this complex. In contrast to Cdk2-cyclin A, which has a well-defined consensus target site ((S/T)PX(K/R)) that strongly favors substrates containing a lysine at the +3 position of substrates, Cdk2-Speedy/Ringo A2 displayed a broad substrate specificity at this position. Consequently, Cdk2-Ringo/Speedy A2 phosphorylated optimal Cdk2 substrates such as histone H1 and a KSPRK peptide poorly, only ~0.08% as well as Cdk2-cyclin A, but non-canonical Cdk2 substrates such as a KSPRY peptide relatively well, with an efficiency of ~80% compared to Cdk2-cyclin A. Cdk2-Speedy/Ringo A2 also phosphorylated authentic Cdk2 substrates, such as Cdc25 proteins, which contain non-canonical CDK phosphorylation sites, nearly as well as Cdk2-cyclin A. Phosphopeptide mapping indicated that Cdk2-Speedy/Ringo A2 and Cdk2-cyclin A phosphorylate distinct subsets of sites on Cdc25 proteins. Thus, the low activity that Cdk2-Speedy/Ringo A2 displays when assayed on conventional Cdk2 substrates may significantly underestimate the potential physiological importance of Cdk2-Speedy/Ringo A2 in phosphorylating key subsets of Cdk2 substrates. Unlike Cdk2-cyclin A, whose activity depends strongly on activating phosphorylation of Cdk2 on Thr-160, neither the overall catalytic activity nor the substrate recognition by Cdk2-Speedy/Ringo A2 was significantly affected by this phosphorylation. Furthermore, Cdk2-Speedy/Ringo A2 was not a suitable substrate for metazoan CAK (which phosphorylates Cdk2 at Thr-160), supporting the notion that Speedy/Ringo A2 activates Cdk2 in a CAK-independent manner.
Conclusion
There are major differences in substrate preferences between CDK-Speedy/Ringo A2 and Cdk2-cyclin complexes. These differences may accommodate the CAK-independent activation of Cdk2 by Speedy/Ringo A2 and they raise the possibility that CDK-Speedy/Ringo A2 complexes could phosphorylate and regulate a subset of non-canonical CDK substrates, such as Cdc25 protein phosphatases, to control cell cycle progression.
==== Body
Background
Eukaryotic cell cycle progression is under the control of cyclin-dependent kinases (CDKs). In higher eukaryotic cells, Cdc2, Cdk2, Cdk4, and Cdk6 control cell cycle progression. Their activities are regulated through a variety of mechanisms, including association with regulatory subunits (cyclins, inhibitors, and assembly factors), subcellular localization, transcriptional regulation, selective proteolysis, and reversible protein phosphorylation (reviewed in [1-7]).
Binding of a cyclin to a CDK is a crucial step in its activation and leads to its phosphorylation on multiple sites [5-8]. Activating phosphorylation within the activation segment (also called the T-loop) on a conserved threonine residue (Thr-160 in Cdk2) is required for CDK activities in vitro and in vivo [9-16] and is carried out by the Cdk-activating kinase (CAK), which is composed of Cdk7, cyclin H, and Mat1 in most eukaryotes. Inhibitory phosphorylations occur on sites equivalent to Thr-14 and Tyr-15 in human Cdc2 or Cdk2 and are carried out by Wee1-like protein kinases (Wee1 and Myt1), and removed by members of the Cdc25 phosphatase family (reviewed in [6,7]). In addition to activating CDKs, cyclins also contribute to their substrate specificities [17-20] and to their subcellular localizations [21]. A general consensus sequence for efficient phosphorylation by Cdc2 and Cdk2 is (K/R)(S/T)PX(K/R) [22-24], in which a basic residue at the +3 position three amino acids C-terminal to the phosphorylated Ser or Thr is particularly important.
In addition to their activation by cyclins, CDKs can be activated by a novel cell cycle regulator called Speedy or Ringo (Rapid inducer of G2/M progression in oocytes), despite the lack of any primary sequence homology between cyclins and Speedy/Ringo proteins [25-28]. Speedy/Ringo proteins were initially described in Xenopus based on their ability to promote the G2 to M transition during oocyte maturation [25,26]. Xenopus Speedy/Ringo bound to and activated Cdc2 and Cdk2 in vitro [26,27]. A human Speedy/Ringo homologue (Spy1) is essential for S-phase entry in cultured somatic cells in a Cdk2-dependent manner: overexpression of human Spy1 accelerated S-phase entry and cell proliferation, and its inhibition by RNAi caused a cell cycle delay at G1/S [28]. These biological functions of Speedy/Ringo proteins were dependent on CDK activity since kinase-inactive forms of Cdc2 and Cdk2 abolished the effects of Spy1 on cell cycle transitions [26,28]. Interestingly, unlike CDK-cyclin complexes whose phosphorylation on the activating site is essential for activity, Speedy/Ringo proteins can activate Cdc2 and Cdk2 in vitro in the absence of activating phosphorylation of the CDK [27]. Thus, Xenopus Speedy/Ringo could activate both wild type Cdk2 and Cdk2T160A equally well and preincubation of Cdk2 with budding yeast CAK (Cak1p) had no effect on its kinase activity toward substrates such as histone H1 [27]. Cdk2-Speedy/Ringo complexes are also resistant to the inhibitory effects of p21Waf1/Cip1 and Wee1 [27]. Paradoxically, although this ability of CDK-Speedy/Ringo complexes to bypass many forms of regulation imposed on CDK-cyclin complexes might suggest that CDK-Speedy/Ringo complexes would be highly active, only very low enzymatic activity toward conventional substrates is associated with CDK-Speedy/Ringo complexes isolated from cells (unpublished data).
The importance of Speedy/Ringo proteins for cell cycle progression combined with the low protein kinase activity of CDK-Speedy/Ringo complexes toward conventional substrates raised the possibility that there might be major biochemical differences between CDK-Speedy/Ringo complexes and CDK-cyclin complexes. These differences might explain how CDK-Speedy/Ringo complexes recognize their substrates and promote cell cycle progression. Major differences are to be expected since the activating phosphorylation that is dispensable for Cdk2-Speedy/Ringo activity plays a key role in substrate recognition by Cdk2-cyclin complexes [29,30]. Activating phosphorylation also enhances substrate binding by CDK-cyclin complexes [30]. The crystal structure of Cdk2-cyclin A with an optimal peptide substrate showed that the activating phosphate forms a hydrogen bond with the side-chain of the lysine at the +3 position of the substrate [31], providing insight into how the activating phosphate participates in substrate recognition. Although Cdk5 can be activated by p35/p25 (cyclin subunit of Cdk5) without activating phosphorylation, a glutamic acid side-chain in p35/p25 mimics the activating phosphate and interacts with a basic residue at the +3 position of the substrate [32]. It will be interesting to know how CDK-Speedy/Ringo complexes recognize substrates efficiently and how an apparently low level of CDK-Speedy/Ringo activity can have profound effects on cell cycle progression. A detailed biochemical understanding of CDK-Speedy/Ringo activity should help us to understand the biological functions of Speedy/Ringo proteins during cell cycle transitions.
In this study, we explored the biochemical properties of human Cdk2-Speedy/Ringo A2. We found that neither its overall catalytic activity nor its substrate recognition required the activating phosphorylation of Cdk2. In fact, Cdk2-Speedy/Ringo A2 was a poor substrate for phosphorylation by metazoan CAK. Cdk2-Speedy/Ringo A2 tolerated almost any amino acid residue at the +3 position of substrates, which is strikingly different from the rigid requirement of Cdk2-cyclin A and Cdk2-cyclin E for a basic residue (and, in particular, for a lysine) at the +3 position. Although Cdk2-Speedy/Ringo A2 phosphorylated canonical Cdk2-cyclin A substrates such as histone H1 and a KSPRK peptide quite poorly, it phosphorylated non-canonical CDK-cyclin substrates, including Cdc25 protein phosphatases, nearly as well as Cdk2-cyclin A. These observations raise the possibility that Cdk2-Speedy/Ringo could phosphorylate and regulate a subset of non-canonical CDK substrates, such as Cdc25 protein phosphatases.
Results
Characterization of purified proteins
We initiated our work by purifying and characterizing the Cdk2, cyclin A, and Speedy/Ringo A2 proteins that would be used throughout these studies. Unphosphorylated GST-Cdk2 ([unP]Cdk2) was expressed in E. coli and purified via its GST tag. To produce Thr160-phosphorylated Cdk2 ([pT160]Cdk2), we used a previously described Cdk2-Cak1p co-expression system [31,33] that can produce essentially fully phosphorylated Cdk2 [31]. As previously reported [29,33], [pT160]Cdk2 by itself displays a low but detectable histone H1 kinase activity (Fig. 1A–B). Cyclin A stimulated the histone H1 kinase activity of [pT160]Cdk2 (solid circle) about 50-fold. The activation of [pT160]Cdk2 by cyclin A plateaued when the ratio of Cdk2 to cyclin A was about 1:1. In contrast to [pT160]Cdk2, the histone H1 kinase activity of [unP]Cdk2 (open squares) remained very low, even at high cyclin A concentrations, confirming the importance of the activating phosphorylation of Cdk2 on Thr-160 (Fig. 1A).
Figure 1 Characterization of [unP]Cdk2, [pT160]Cdk2, Speedy/Ringo A2, and K2A2coexp. (A) Activation of [unP]Cdk2 and [pT160]Cdk2 by cyclin A. 0.05 μg of [unP]Cdk2 or [pT160]Cdk2 was preincubated with the indicated amounts of cyclin A prior to determination of its histone H1 kinase activity. (B) Activation of [unP]Cdk2 and [pT160]Cdk2 by Speedy/Ringo A2. 0.5 μg of [unP]Cdk2 or [pT160]Cdk2 was preincubated with the indicated amounts of GST-Speedy/Ringo A2 prior to determination of its histone H1 kinase activity. (C) Gel filtration analysis of K2A2coexp. Cdk2 and Speedy/Ringo A2-His6 were coexpressed in E. coli and purified on a metal affinity column. The purified K2A2coexp was loaded on a Superdex-200 column; one-ml fractions were collected. Proteins from 20 μl of the input (lane L) or from 40 μl of fractions 9–21 were resolved by SDS-PAGE, transferred to a PVDF membrane, and detected with a Cdk2-specific antibody (lower panel), or by staining with Coomassie Brilliant Blue R-250 (upper panel). (D) Time course of ATPase activity of [pT160]Cdk2-cyclin A and K2A2coexp. 16.7 μM of [pT160]Cdk2-cyclin A or K2A2coexp was incubated with [γ-32P]ATP for the indicated times. Samples were chromatographed to resolve 32Pi from [γ-32P]ATP. The rate of ATP hydrolysis was quantitated by phosphorimaging analysis. (E) The relative ATPase activities of [pT160]Cdk2-cyclin A and K2A2coexp were calculated from the slopes in panel D. The ATPase activity of [pT160]Cdk2-cyclin A was set to 100%.
We next determined the histone H1 kinase activities of [unP]Cdk2 and [pT160]Cdk2 after incubation with GST-Speedy/Ringo A2, which was expressed and purified from E. coli. We have found that the previously reported mammalian Speedy/Ringo protein, Spy1, is expressed as two closely related proteins resulting from alternative splicing [34]. These two proteins, which we term Speedy/Ringo A1 and A2, differ only at their extreme C-termini and appear to be indistinguishable in all functional respects. Spy1 corresponds to Speedy/Ringo A1. As shown in Fig. 1B, GST-Speedy/Ringo A2 activated [unP]Cdk2 (open squares) and [pT160]Cdk2 (solid circles) equally well, indicating that Speedy/Ringo A2 can bind to and activate Cdk2 without regard to its activating phosphorylation. Cdk2 activity plateaued at a GST-Speedy/Ringo A2 to GST-Cdk2 ratio of at least 4:1. We suspect that an excess of GST-Speedy/Ringo A2 over GST-Cdk2 was required because much of the GST-Speedy/Ringo A2 protein aggregates, as we have observed by gel-filtration chromatography (data not shown), thereby reducing the effective concentration of active Speedy/Ringo A2.
As an alternative way to produce functional Cdk2-Speedy/Ringo A2 complexes, we also developed a coexpression system in which a C-terminally hexahistidine-tagged form of Speedy/Ringo A2 was coexpressed with untagged Cdk2 in E. coli. [unP]Cdk2-Speedy/Ringo A2-his6 (K2A2coexp) was purified on a metal affinity column and by gel filtration chromatography. Fractions were resolved by 10% SDS-PAGE and transferred to a PVDF membrane. Total proteins were detected by staining with Coomassie Brilliant Blue R-250 (Fig. 1C top panel) and Cdk2 was detected by immunoblotting with anti-Cdk2 antibodies (Fig. 1C lower panel). Speedy/Ringo A2-his6, which migrated as a 42 kDa protein on 10% SDS-PAGE, co-eluted with Cdk2 from the gel filtration column. The apparent MW for the peak fractions (fractions 14–15) was ~80 kDa, close to the combined molecular weights of monomeric Cdk2 (33 kDa) and Speedy/Ringo A2-his6 (42 kDa). The diffuse appearance of Speedy/Ringo A2 is probably due to phosphorylation of Speedy/Ringo A2 by Cdk2 (data not shown). The ratio of Cdk2 to Speedy/Ringo A2-his6 appears to be close to 1:1 in fractions 14 and 15, suggesting that one molecule of Speedy/Ringo A2 binds one molecule of Cdk2. K2A2coexp displayed very similar protein kinase activity to in vitro-assembled Cdk2-Speedy/Ringo A2 complexes (see below), suggesting that both types of complexes may be fully active. Note that because of the coexpression of Cdk2 and Speedy/Ringo A2, and the poor phosphorylation of Cdk2 bound to Speedy/Ringo A2 by both metazoan CAK (Cdk7/cyclinH/Mat1) and budding yeast Cak1p (see below), we could only produce the unphosphorylated form of K2A2coexp. This [unP]Cdk2-Speedy/Ringo A2 heterodimer was used as a control throughout these studies.
We measured the ATPase activities of Cdk2-cyclin A and Cdk2-Speedy/Ringo A2 to assess the relative fractions of functional complexes. We reasoned that ATPase activities would be less subject to substrate specificity effects conferred by cyclin A or Speedy/Ringo A2 than protein kinase assays. The ATPase activity of monomeric [unP]Cdk2 is low; both activating phosphorylation and cyclin-binding increase the turnover of ATP by 20–25 fold [29,30]. Therefore, ATPase activity should be a reasonable indicator for the adoption of the catalytically active conformation of Cdk2 in Cdk2-Speedy/Ringo A2 complexes. We determined the ATPase activities of [pT160]Cdk2-cyclin A and K2A2coexp. A time course experiment (Fig. 1D) showed that the rate of ATP hydrolysis remained linear during the assay. The relative ATPase activities of [pT160]Cdk2-cyclin A and K2A2coexp were calculated from the slopes in Fig. 1D and are shown in Fig. 1E. The ATPase activity of K2A2coexp was slightly higher than that of [pT160]Cdk2-cyclin A, indicating that Cdk2 in K2A2coexp is as active as that in [pT160]Cdk2-cyclin A. We should emphasize that in vitro-assembled [unP]Cdk2-Speedy/Ringo A2 and K2A2coexp displayed very similar activities and substrate specificities (see below, and data not shown). Thus, it appears that in vitro-assembled Cdk2-Speedy/Ringo A2 complexes are fully active and it is reasonable to assume that they will display physiological biochemical characteristics.
Substrate specificity of Cdk2-Speedy/Ringo A2
We used a systematic panel of CDK substrates to investigate the substrate specificity of Cdk2-Speedy/Ringo A2 complexes. These substrates contained a pentapeptide of the form XSPXX (where X indicates any amino acid) fused to the C-terminus of GST [24]. Substrates were purified as GST-fusion proteins from E. coli. The optimal substrates for [pT161]Cdc2-cyclin B and [pT160]Cdk2-cyclin A have been identified as (K/R)(S/T)PX(K/R). Substitutions of the basic residue at the +3 position (with respect to the phosphorylation site) had the greatest effects on phosphorylation efficiencies, both by Cdc2 and by Cdk2 [24]. Thus, KSPRK is considered the "wild-type" substrate in these studies. We determined which position was most important for phosphorylation by Cdk2-Speedy/Ringo A2 using alanine substitution substrates in which the charged residues at -1, +2, and +3 were individually replaced with alanine. We compared the abilities of these substrates to be phosphorylated by Cdk2-Speedy/Ringo A2 (assembled in vitro), K2A2coexp, [pT160]Cdk2-Speedy/Ringo A2, and [pT160]Cdk2-cyclin A (Fig. 2). (Note that we are comparing relative substrate preferences and not absolute phosphorylation efficiencies, which vary greatly between Cdk2 bound to cyclin A and to Speedy/Ringo A2 (see below and Table 1).) For all of these enzymes, replacement of the lysine at the -1 position (ASPRK) had no effect on phosphorylation efficiency, substitution at the +2 position (KSPAK) had a modest effect, and substitution at the +3 position (KSPRA) produced the severest effects (Fig. 2). Cdk2-cyclin A was much more sensitive to the +2 and +3 substitutions than any of the forms of Cdk2-Speedy/Ringo A2. Similar effects were observed using K2A2coexp and [unP]Cdk2-Speedy/Ringo A2, suggesting that the different ways in which these complexes were produced had little effect on their resulting substrate specificities. Phosphorylation of Cdk2 in Cdk2-Speedy/Ringo A2 complexes had virtually no effect on substrate specificity, which was surprising given the role of this phosphate in recognition of the +3 position of substrate by Cdk2-cyclin A [30,31]. Nevertheless, the +3 position was the most important position for both [unP]Cdk2-Speedy/Ringo A2 and [pT160]Cdk2-Speedy/Ringo A2, just as it is for [pT160]Cdk2-cyclin A.
Figure 2 Effects of alanine substitutions at each of the three charged positions in a KSPRK substrate on relative phosphorylation by Cdk2-cyclin A and Cdk2-Speedy/Ringo A2. The phosphorylation efficiencies of the indicated GST peptides by K2A2coexp, [unP]Cdk2-Speedy/Ringo A2, [pT160]Cdk2-Speedy/Ringo A2, and [pT160]Cdk2-cyclin A were compared. Assays were performed at substrate concentrations of 50 μM. All values are relative to the phosphorylation of the wild type (KSPRK) substrate by the same enzyme. Values represent the means ± S.E. from three separate experiments.
Table 1 Comparison of the relative phosphorylation efficiencies of the indicated substrates by [pT160]Cdk2-cyclin A, [unP]Cdk2-Speedy/Ringo A2, [pT160]Cdk2-Speedy/Ringo A2, and coexpressed [unP]Cdk2-Speedy/Ringo A2 (K2A2coexp). Values represent the means from three independent experiments.
Enzyme complex Histone H1 (5 μM) KSPRK (50 μM) KSPRR (50 μM) KSPRY (50 μM)
(pmol phosphate·min-1·μg-1 Cdk2)
[pT160]Cdk2-cyclin A 1983 1062 90 3.3
[unP]Cdk2- Speedy/Ringo A2 1.68 0.83 2.55 2.43
[pT160]Cdk2- Speedy/Ringo A2 2.05 1.02 3.13 0.68
K2A2coexp 1.55 1.40 5.5 3.15
We next evaluated the substrate specificity of Cdk2-Speedy/Ringo A2 at the +3 position in more detail using a panel of GST-KSPRX substrates [24]. We first compared the substrate specificity of [unP]Cdk2-Speedy/Ringo A2 with that of [unP]Cdk2-cyclin A, allowing us to compare the effects of the binding partner, separate from effects due to the Cdk2 phosphorylation state. The relative specificities of [unP]Cdk2-Speedy/Ringo A2 and [unP]Cdk2-cyclin A were determined at a substrate concentration of 50 μM, which is well below the KM value of Cdk2-Speedy/Ringo A2 and [pT160]Cdk2-cyclin A (see below and [24]) for these substrates, and thus within the linear range of the assay. We also compared the abilities of these kinases to phosphorylate histone H1 (5 μM). The phosphorylation of the "wild-type" substrate (GST-KSPRK) was defined as 100%. Although [unP]Cdk2-cyclin A was less active than fully activated [pT160]Cdk2-cyclin A, [unP]Cdk2-cyclin A still preferred lysine and arginine residues at the +3 position (Fig. 3A), consistent with a previous report indicating that [unP]Cdk2-cyclin A was only moderately defective in substrate binding [29]. In striking contrast to [unP]Cdk2-cyclin A, [unP]Cdk2-Speedy/Ringo A2 tolerated nearly all amino acid substitutions at the +3 position (Fig. 3A). The best substrates for [unP]Cdk2-Speedy/Ringo A2 contained tyrosine (Y) at 473 ± 82%, arginine (R) at 325 ± 74%, and tryptophan (W) at 293 ± 56%. More than half of the amino acid substitutions at the +3 position produced better substrates than lysine. In fact, 17 out of the 20 substrates were phosphorylated at least 50% as efficiently as KSPRK; only alanine, asparagine, and glutamine yielded poor substrates. In contrast, [pT160]Cdk2-cyclin A was unable to phosphorylate any substitution substrate more than 5% as efficiently as KSPRK, and most substitutions produced substrates whose phosphorylation was undetectable (< 0.01% of KSPRK).
Figure 3 Effects of amino acid substitutions at the +3 position of KSPRK on substrate utilization by [unP]Cdk2-Speedy/Ringo A2, [unP]Cdk2-cyclin A, K2A2coexp, and [pT160]Cdk2-cyclin A. (A) Comparison of the substrate specificity of [unP]Cdk2-Speedy/Ringo A2 (open bars) and [unP]Cdk2-cyclin A (solid bars). Assays were performed at substrate concentrations of 50 μM. All phosphorylation efficiencies are relative to the phosphorylation of the KSPRK substrate by the same enzyme. Values represent the means ± S.E. from three separate experiments. (B) Comparison of the substrate specificities of [unP]Cdk2-Speedy/Ringo A2, K2A2coexp, and [pT160]Cdk2-Speedy/Ringo A2. Assays were performed at substrate concentrations of 50 μM. All phosphorylation efficiencies are relative to the phosphorylation of the KSPRK substrate by the same enzyme. Values represent the means ± S.E. from three separate experiments. Single letters indicate the amino acid at the +3 position of KSPRK. H1, histone H1.
We next compared the sensitivities of in vitro-assembled [unP]Cdk2-Speedy/Ringo A2 and K2A2coexp to substitutions at the +3 position to test whether the route of formation of these complexes affected their substrate utilization. As shown in Fig. 3B, in vitro-assembled [unP]Cdk2-Speedy/Ringo A2 and K2A2coexp exhibited almost identical phosphorylation efficiency profiles for the tested substrates, including histone H1. Except for the KSPRK and KSPRT substrates, K2A2coexp phosphorylated all substrates slightly less efficiently than in vitro-assembled [unP]Cdk2-Speedy/Ringo A2, possibly due to its slightly higher enzymatic activity toward KSPRK (Table I), which was defined as 100%. Overall, the differences in substrate specificity and enzymatic activity between the in vitro-assembled complexes and K2A2coexp were marginal. Both complexes could accept any amino acid side chain except aspartate and glutamate at the +3 position.
We then determined the effect of Thr-160 phosphorylation on the substrate specificity of [pT160]Cdk2-Speedy/Ringo A2. The optimal substrates at the +3 position were arginine (R) at 250 ± 19%, histidine (H) at 107 ± 6%, and lysine (K) at 100%. Phosphorylation of KSPRW and KSPRY were reduced from the very high levels exhibited by [unP]Cdk2-Speedy/Ringo A2 to levels more typical of other amino acid substitutions. Overall, Thr-160 phosphorylation modulated the substrate specificity of Cdk2-Speedy/Ringo A2 but did not transform it into a dramatically different pattern (Fig. 3B).
To determine the absolute (rather than relative) activities of Cdk2-cyclin A and Cdk2-Speedy/Ringo A2, we compared the abilities of [pT160]Cdk2-cyclin A and Cdk2-Speedy/Ringo A2 to phosphorylate several substrates, including histone H1, KSPRK, KSPRR, and KSPRY (Table I). For Cdk2-cyclin A, histone H1 and KSPRK are the best substrates, although KSPRR also fits the consensus substrate sequence, and KSPRY is an unfavorable substrate [24]. Cdk2-Speedy/Ringo A2 phosphorylated all four substrates with similar efficiencies (Table 1). [unP]Speedy/Ringo A2, [pT160]Cdk2-Speedy/Ringo A2, and K2A2coexp phosphorylated the optimal Cdk2-cyclin A substrates histone H1 and KSPRK only about 0.1% as efficiently as [pT160]Cdk2-cyclin A (Table 1). In fact, the Cdk2-Speedy/Ringo A2 complexes were only about 6-fold more active toward histone H1 than monomeric [pT160]Cdk2 (Fig. 1B). The differences between Cdk2-cyclin A and Cdk2-Speedy/Ringo A2 decreased dramatically when activity was measured toward poorer Cdk2-cyclin A substrates. For instance, K2A2coexp phosphorylated KSPRR 6% as efficiently as [pT160]Cdk2-cyclin A and it phosphorylated KSPRY 95% as efficiently as [pT160]Cdk2-cyclin A (Table I). Thus, the relative activities of Cdk2 bound to cyclin A or to Speedy/Ringo A2 depend critically on the substrate used.
To understand why Cdk2-Speedy/Ringo A2 was much less active than [pT160]Cdk2-cyclin A toward certain substrates, we tried to assess the KM and catalytic activity (Vmax) of Cdk2-Speedy/Ringo A2. We chose several substrates including KSPRK, KSPRR, KSPRY, and histone H1 for this analysis. Phosphorylation of these substrates by [unP]Cdk2-Speedy/Ringo A2 and [pT160]Cdk2-Speedy/Ringo A2 was carried out over a wide range of substrate concentrations. The maximal concentrations were 250 μM for histone H1 and 1000 μM for the GST-peptide substrates. The phosphorylation of histone H1, KSPRY, and KSPRK by [unP]Cdk2-Speedy/Ringo A2 increased linearly up to the maximum concentration of each substrate (Fig. 4A). Interestingly, the phosphorylation of the KSPRR substrate plateaued at high substrate concentrations (open squares in Fig. 4A). We observed a similar phosphorylation efficiency-concentration relationship using K2A2coexp (data not shown), suggesting that KSPRR is a good substrate for [unP]Cdk2-Speedy/Ringo A2. The phosphorylation of KSPRY (open circles) continued to increase linearly even at substrate concentrations at which the phosphorylation of KSPRR had plateaued, indicating that [unP]Cdk2-Speedy/Ringo A2 might have a stronger affinity for the KSPRR peptide but that it can perform the chemical steps of catalysis more efficiently on KSPRY. We also determined the utilization of KSPRK, KSPRR, KSPRY, and histone H1 by [pT160]Cdk2-Speedy/Ringo A2 over the same range of substrate concentrations. The phosphorylation of these four substrates increased linearly throughout the concentration range (Fig. 4B). Surprisingly, the phosphorylation of KSPRR by [pT160]Cdk2-Speedy/Ringo A2 increased linearly throughout the concentration range, suggesting that activating phosphorylation actually decreased the affinity of Cdk2-Speedy/Ringo A2 for KSPRR. These results differ significantly from those obtained using Cdk2-cyclin A, which has a KM for phosphorylation of histone H1 of 0.8 μM [29], and for phosphorylation of KSPRK of 150 μM [24]. Thus, it appears that weak substrate binding contributed, at least partially, to the low enzymatic activity of Cdk2-Speedy/Ringo A2 toward these substrates.
Figure 4 Velocity versus concentration plots for phosphorylation of substrates by [unP]Cdk2-Speedy/Ringo A2 and [pT160]Cdk2-Speedy/Ringo A2. Velocity versus concentration plots for [unP]Cdk2-Speedy/Ringo A2 (A) and [pT160]Cdk2-Speedy/Ringo A2 (B). The substrates are histone H1 (solid circles), KSPRK (solid squares), KSPRR (open squares), and KSPRY (open circles).
Phosphorylation of Cdc25 proteins
The above results indicate that Cdk2-Speedy/Ringo A2 complexes phosphorylate CDK substrates with canonical motifs (S/T)PX(K/R) poorly compared to Cdk2-cyclin A, but that they can phosphorylate non-canonical CDK substrates relatively well. These findings suggested that Cdk2-Speedy/Ringo A2 might efficiently phosphorylate some physiological Cdk2 substrates containing non-canonical motifs. We reasoned that Cdc25 proteins might be such substrates. Autoamplification of Cdc2 activity at the G2/M transition was proposed more than a decade ago [8]. In this model, Cdc25 protein phosphatases activate Cdc2-cyclin B complexes by removing inhibitory phosphates from Cdc2. In turn, active Cdc2-cyclin B phosphorylates and activates Cdc25, which activates additional Cdc2-cyclin B complexes. Thus, a low level of CDK activity can be amplified through this positive feedback loop, leading to the concerted activation of Cdc2 and entry into mitosis. To achieve an abrupt all-or-none activation of Cdc2 at the right time, it is important that Cdc25 proteins not be activated at too low a level of Cdc2 activity. From this point of view, one would predict that Cdc25 would be a poor CDK substrate. A similar analysis applies to the activation of Cdk2 complexes by Cdc25 proteins.
All three mammalian Cdc25 isoforms (A, B, and C) can be phosphorylated on multiple sites by CDK-cyclin complexes (reviewed in [35-37]). Among 32 potential CDK phosphorylation sites ((S/T)PXX) in human Cdc25A, B, and C (Fig. 5A), only 2 sites fit the consensus CDK phosphorylation motif, and even these are weak fits for phosphorylation by Cdk2 (SPXR) since they lack a lysine at the +3 position [24]. We conclude that most CDK phosphorylation of Cdc25 proteins occurs on non-canonical motifs. We, therefore, examined whether Cdc25 proteins could be phosphorylated by Cdk2-Speedy/Ringo A2. As shown in Fig. 5B, both Cdk2-cyclin A and K2A2coexp phosphorylated GST-Cdc25A, B, and C. To compare the activities of Cdk2-cyclin A and Cdk2-Speedy/Ringo A2 toward Cdc25, we varied the amounts of [pT160]Cdk2-cyclin A and K2A2coexp used to phosphorylate constant amounts of GST-Cdc25A, B, and C. We estimated that the enzymatic activity of [pT160]Cdk2-cyclin A toward the Cdc25 proteins was 7–14 fold higher than that of K2A2coexp. In contrast, [pT160]Cdk2-cyclin A phosphorylated histone H1 >1000-fold more efficiently than Cdk2-Speedy/Ringo A2 (Fig. 5C).
Figure 5 Phosphorylation of Cdc25A, B, and C by Cdk2-Speedy/Ringo A2. (A) Potential CDK phosphorylation sites in human Cdc25A, B, and C. There are 12 (S/T)PXX sequences in Cdc25A, 14 in Cdc25B, and 6 in Cdc25C. The two sequences that come closest to fitting the consensus CDK phosphorylation sequence (S/T)PX(K/R) are underlined. (B) Phosphorylation of Cdc25A, B, and C by K2A2coexp and [pT160]Cdk2-cyclin A. Phosphorylation of GST-Cdc25A, B, and C (5 μg) by the indicated amounts of K2A2coexp (lanes 1–6) or Cdk2-cyclin A (lanes 7–12). Phosphorylated proteins were separated by SDS-PAGE and detected by autoradiography. (C) Phosphorylation of histone H1 by K2A2coexp and [pT160]Cdk2-cyclin A. Histone H1 (5 μM) was phosphorylated by the indicated amounts of K2A2coexp (lanes 1–6) or Cdk2-cyclin A (lanes 7–12). Phosphorylated proteins were separated by SDS-PAGE and detected by autoradiography.
To determine whether Cdk2-Speedy/Ringo A2 and Cdk2-cyclin A phosphorylate Cdc25 proteins on the same or different sites, we compared tryptic phosphopeptide maps of Cdc25 proteins phosphorylated by [pT160]Cdk2-cyclin A (Fig. 6A) with those phosphorylated by K2A2coexp (Fig. 6B). Each Cdc25 protein was phosphorylated to about the same extent by each Cdk2 complex. Since there are multiple potential CDK phosphorylation sites in Cdc25 proteins (Fig. 5A), it was not surprising that multiple sites were actually phosphorylated. As shown in Fig. 6, phosphorylation of each Cdc25 protein by Cdk2-Speedy/Ringo A2 and by Cdk2-cyclin A produced distinct phosphopeptide patterns. These phosphopeptides could be grouped into two categories. First, many phosphopeptides were unique for either Cdk2-cyclin A or Cdk2-Speedy/Ringo A2, indicating a very strong influence of Cdk2 partner on phosphorylation specificity. For example, spots A1, A2, B1, B2, and B3 were unique for Cdk2-cyclin A, whereas spots a1, a2, b1, b2, and c4 were exclusively associated with Cdk2-Speedy/Ringo A2. Phosphopeptides in the second group were phosphorylated by both kinases, but to different extents. For example, in the phosphopeptide maps of Cdc25A, spots A3, A4, A5, and A6 generated by Cdk2-cyclin A (Fig. 6A) appear to be identical to spots a3, a4, a5, and a6 generated by Cdk2-Speedy/Ringo A2 (Fig. 6B), respectively. While Cdk2-cyclin A phosphorylated site A3 more strongly than sites A4, A5, and A6, Cdk2-Speedy/Ringo A2 had the opposite site preference. In the phosphopeptide maps of Cdc25C, Cdk2-cyclin A phosphorylated peptide C2 more efficiently than peptide C1 whereas Cdk2-Speedy/Ringo A2 phosphorylated C1 more efficiently than C2. Therefore, Cdk2-Speedy/Ringo A2 and Cdk2-cyclin A have distinct but partially overlapping substrate preferences on natural substrates.
Figure 6 Trypic phosphopeptide mapping of Cdc25 proteins. Tryptic phosphopeptide mapping of Cdc25 phosphorylated by [pT160]Cdk2-cyclin A (A) and by K2A2coexp (B). GST-Cdc25A, B, and C (5 μg) were phosphorylated by K2A2coexp and [pT160]Cdk2-cyclin A, separated by 10% SDS-PAGE, extracted, and digested with trypsin. Phosphopeptides were separated on thin layer chromatography plates by electrophoresis followed by chromatography and were detected by autoradiography.
Cdk2 activation loop conformation in Cdk2-Speedy/Ringo A2 complexes
We used accessibility to kinases and phosphatases to probe the conformation of the Cdk2 activation loop in Cdk2-Speedy/Ringo A2 complexes. The activation loop plays an important role in substrate recognition by Cdk2-cyclin A as the Thr-160 phosphate makes direct contact with the +3 position of substrates. The strikingly different substrate preferences of Cdk2-cyclin A and Cdk2-Speedy/Ringo A2 raised the possibility that the activation loop in Cdk2-Speedy/Ringo A2 might adopt a different conformation from that exhibited by Cdk2-cyclin A. In addition, like Cdk2-Speedy/Ringo A2, monomeric [pT160]Cdk2 displayed low enzymatic activity toward histone H1 ([29,33] and Fig. 1B) and defective substrate binding [29]. Structural studies indicate that the activation loop in monomeric [pT160]Cdk2 is highly disorganized and that only a very small population of [pT160]Cdk2 is in the active conformation at any given time [33]. This comparison raised the possibility that the activation loop in Cdk2-Speedy/Ringo A2 might also be disordered.
We first probed the T-loop conformation in Cdk2-Speedy/Ringo A2 using budding yeast Cak1p. Cak1p phosphorylates monomeric Cdk2 efficiently, but this phosphorylation is inhibited over 95% in the presence of cyclin A [38], presumably because the fixed conformation of the activation loop in Cdk2-cyclin A forms a poor substrate for Cak1p. A kinase-inactive form of Cdk2 (GST-Cdk2D145N) was used in this assay to eliminate the high background phosphorylation of Speedy/Ringo A2 by wild-type Cdk2. GST-Cdk2D145N bound in vitro translated [35S]-Speedy/Ringo A2 as well as wild-type Cdk2 (data not shown). GST-Cdk2 was preincubated with increasing amounts of GST-Speedy/Ringo A2 or GST alone before phosphorylation by Cak1p. As shown in Fig. 7A, the phosphorylation of Cdk2 by Cak1p was reduced in the presence of high concentrations of Speedy/Ringo A2, indicating that the binding of Speedy/Ringo A2 to Cdk2 changed the conformation of the activation loop.
Figure 7 Effects of Speedy/Ringo A2 binding on the phosphorylation and dephosphorylation of Cdk2 on Thr-160. (A) Speedy/Ringo A2 hinders the phosphorylation of Cdk2 on Thr-160 by budding yeast Cak1p. GST-Cdk2D145N (1 μg) was incubated with increasing amounts (0, 1, 2.5, 5, and 10 μg) of GST-Speedy/Ringo A2 or GST prior to phosphorylation by Cak1p (100 ng). The reactions were terminated and phosphorylated proteins were resolved by SDS-PAGE and detected by autoradiography. (B) Speedy/Ringo A2 binding did not stimulate Thr-160 phosphorylation by mammalian CAK. GST-Cdk2D145N (1 μg; lanes 1, 4, 5) was preincubated with cyclin A (1 μg; lanes 2 and 4) or GST-Speedy/Ringo A2 (4 μg; lanes 3 and 5) before phosphorylation by CAK. Samples were processed as described in Methods. (C) Speedy/Ringo A2 hinders the dephosphorylation of Cdk2 on Thr-160 by PP2Cα. [32P-T160]GST-Cdk2D145N was preincubated with cyclin A (lanes 2 and 4), GST-Speedy/Ringo A2 (lanes 6 and 8), or buffer alone (lanes 1, 3, 5, 7) prior to addition of recombinant human PP2Cα (lanes 3, 4, 7, 8) or buffer (lanes 1, 2, 5, 6). [32P-T160]GST-Cdk2D145N was detected by autoradiography.
We next investigated the T-loop conformation of Cdk2-Speedy/Ringo A2 using mammalian CAK (Cdk7/cyclin H/Mat1). Monomeric CDKs are poor substrates for mammalian CAK whereas the binding of cyclin stimulates activating phosphorylation by CAK more than seven fold [38]. We again used a kinase-inactive form of GST-Cdk2 (GST-Cdk2D145N) to reduce background phosphorylation of GST-Speedy/Ringo A2 by wild-type Cdk2. GST-Cdk2D145N alone was slightly phosphorylated by CAK (Fig. 7B, lane 1). The phosphorylation of GST-Cdk2D145N was greatly enhanced in the presence of cyclin A (Fig. 7B, lane 4). In contrast, GST-Speedy/Ringo A2 had no effect on the phosphorylation GST-Cdk2D145N by CAK (Fig. 7B, lane 5). The defective phosphorylation of Cdk2-Speedy/Ringo A2 by mammalian CAK is unlikely to be due to a failure of Speedy/Ringo A2 and GST-Cdk2D145N to form a complex. Taking these results together, we conclude that binding of Speedy/Ringo A2 alters the activation loop conformation of Cdk2 (Fig. 7A), but in a manner distinct from how cyclin A does so (Fig. 7B). Furthermore, it appears that Cdk2-Speedy/Ringo A2 is a poor substrate for mammalian CAK.
Finally, we examined the T-loop conformation in the Cdk2-Speedy/Ringo A2 complex using serine/threonine protein phosphatase type 2C (PP2C) [39]. Mammalian PP2Cα and β can remove the activating phosphate from monomeric CDKs such as Cdk2 and Cdk6 [40]. The binding of cyclin A to Cdk2 prevents dephosphorylation by PP2C [40], presumably because the activation loop becomes locked into a conformation inaccessible to PP2C [41]. We tested whether Speedy/Ringo A2, like cyclin A, can prevent dephosphorylation by PP2Cα. GST-Cdk2D145N was labeled with [γ-32P]-ATP using Cak1p and purified by gel filtration. 32P-labeled Cdk2 was preincubated with Speedy/Ringo A2, cyclin A, or buffer before addition of PP2Cα. As described previously [40], cyclin A fully blocked the dephosphorylation of Cdk2 (Fig. 7C, compare lanes 3 and 4). The binding of Speedy/Ringo A2 partially blocked the dephosphorylation of Cdk2 by PP2Cα (Fig. 7C, lane 8). Based on this intermediate result we conclude, as with the Cak1p and CAK results above, that binding to Speedy/Ringo A2 alters the conformation of the activation loop, but that this conformation differs from that seen after binding of cyclin A.
Discussion
The recently discovered Speedy/Ringo proteins represent a novel class of non-cyclin CDK activators that play important roles in cell cycle progression. Xenopus Speedy/Ringo is necessary for G2/M progression during oocyte maturation and a human Speedy/Ringo protein (Spy1) regulates S-phase entry in cultured cells [25-28]. Although apparently not present in yeast, plants, and insects, Speedy/Ringo homologues can be found in the most primitive branching clade of chordates (Ciona intestinalis) [34], from which all vertebrates evolved. It is conceivable that Speedy/Ringo proteins regulate cell cycle progression in all vertebrates. Although there is no obvious primary sequence similarity between cyclins and Speedy/Ringo proteins, Speedy/Ringo proteins can bind to and activate CDKs directly [27,34]. In this study, we carried out a biochemical characterization of Cdk2-Speedy/Ringo A2 complexes. We verified that human Speedy/Ringo A2 can form a 1:1 complex with Cdk2 and that it can activate Cdk2 in vitro, even in the absence of phosphorylation of Cdk2 on Thr-160.
However, Speedy/Ringo A2 is not a simple replacement for certain cyclins; there are many important biochemical differences between cyclins and Speedy/Ringo proteins. First, Cdk2-Speedy/Ringo A2 displays a broad substrate specificity, which is very different from the narrow consensus CDK phosphorylation motif. Previous studies showed that Cdk2-cyclin A phosphorylated (K/R)(S/T)PX(K/R) sequences, with a strong preference for a lysine at the terminal position. We found, using a systematic peptide substrate panel, that the +3 position is also the most important residue for Cdk2-Speedy/Ringo A2 recognition. At the +3 position, the best substrates for Cdk2-cyclin A are KSPRK and KSPRR (~5% of KSPRK), which contain basic residues at the +3 position [24]. In contrast, the best substrates for Cdk2-Speedy/Ringo A2 contained tyrosine (Y), arginine (R), and tryptophan (W) in the +3 position, all of which carry bulky side chains. Furthermore, Cdk2-Speedy/Ringo A2 complexes could tolerate almost any amino acid residue at the +3 position and phosphorylated 17 of the 20 +3 substrates at least 50% as well as KSPRK. More than half of the +3 substitutions yielded substrates that were phosphorylated more efficiently than KSPRK. Only alanine, aspartate, and glutamate formed poor substrates, whose relative phosphorylation by Cdk2-Speedy/Ringo A2 was still higher than the relative phosphorylation of all but a few of the +3 substrates by Cdk2-cyclin A.
The second difference between cyclins and Speedy/Ringo A2 is that Cdk2-Speedy/Ringo A2 possesses low enzymatic activity toward conventional CDK substrates, indicating that Speedy/Ringo A2 is unlikely to be able to replace cyclins and promote the full range of Cdk2 substrate phosphorylation. For example, the activity of Cdk2-Speedy/Ringo A2 toward histone H1 was only ~0.1% the activity of [pT160]Cdk2-cyclin A in the standard histone H1 kinase assay. Therefore, it is unlikely that CDK-Speedy/Ringo can promote cell cycle progression by itself. The broad tolerance of Cdk2-Speedy/Ringo A2 for substitutions at the +3 position of substrates and its low activity toward conventional Cdk2 substrates may go hand in hand. The insensitivity of Cdk2-Speedy/Ringo A2 to phosphorylation of Thr-160 on Cdk2 may also contribute to its low activity toward conventional substrates as this phosphate plays a direct role in substrate recognition via interaction with +3 basic residues. Indeed, the activation loop conformation adopted by Cdk2 upon binding Speedy/Ringo A2 appears to differ significantly from that adopted upon cyclin A binding. Cdk2-Speedy/Ringo A2 also appeared to bind substrates, except for KSPRR, poorly, which may contribute to its low enzymatic activity.
Although it displays a low enzymatic activity toward conventional CDK substrates such as histone H1, Cdk2-Speedy/Ringo A2 actually phosphorylated non-canonical CDK substrates nearly as well as Cdk2-cyclin A. We extended this observation made using model substrates to physiological CDK substrates, the Cdc25 dual-specificity phosphatases. We found that Cdk2-Speedy/Ringo A2 could phosphorylate the three human Cdc25 proteins quite efficiently, about 1000 times more efficiently than would be expected based on the histone H1 kinase activity of Cdk2-Speedy/Ringo A2 and only about an order of magnitude less well than Cdk2-cyclin A. Furthermore, phosphopeptide mapping of Cdc25 proteins confirmed that the substrate specificity of Cdk2-Speedy/Ringo A2 overlaps with but is distinct from that of Cdk2-cyclin A. It should also be pointed out that, in addition to inherent substrate specificity, some CDKs are targeted to some of their substrates. For example, the S-phase cyclins, such as cyclin A and Clb5, possess a 'hydrophobic patch' that can interact with the 'RXL' or 'Cy' motif present in some substrates to carry out specific functions during DNA replication [18,31,42-46]. In contrast, phosphorylations of histone H1 and GST-peptide substrates are independent of any docking site. Thus, for Cdk2-Speedy/Ringo A2, the existence of a docking site on Speedy/Ringo A2 would greatly increase the phosphorylation of select substrates by Cdk2-Speedy/Ringo A2.
A third difference between cyclins and Speedy/Ringo A2 is that Speedy/Ringo A2 can activate Cdk2 independent of the activating phosphorylation on Thr-160 of Cdk2. Previous studies have shown that Xenopus Speedy/Ringo can activate Cdc2 and Cdk2 in vitro in the absence of activating phosphorylation and that it can render Cdk2 less sensitive to inhibition via inhibitory phosphorylation and binding of CKIs such as p21 [27]. We examined whether activating phosphorylation, though not necessary for phosphorylation of histone H1, might nonetheless affect the substrate specificity of Cdk2-Speedy/Ringo A2. Neither the overall catalytic activity of Cdk2-Speedy/Ringo A2 nor its substrate recognition required the activating phosphorylation of Cdk2. In fact, Thr-160 phosphorylation reduced the activity of Cdk2-Speedy/Ringo A2 toward most of the tested substrates and did not promote activity toward any particular substrate. We also found that Cdk2-Speedy/Ringo A2 is a poor substrate for metazoan CAK (Cdk7/Cyclin H/Mat1). These findings indicate that the activation of Cdk2 by Speedy/Ringo A2 does not require the activating phosphorylation by CAK and that it may proceed in the absence of this phosphorylation.
Conclusion
We have identified crucial biochemical differences between Cdk2-cyclin A and Cdk2-Speedy/Ringo A2 complexes. These findings raise the possibility that Speedy/Ringo A2 could play a significant role in the phosphorylation of CDK substrates containing non-canonical phosphorylation sites and suggest that Cdc25 proteins might be physiological targets for Cdk2-Speedy/Ringo A2 complexes. This is an intriguing possibility given the positive feedback that exists, for instance, between activation of Cdc2 by Cdc25 and of Cdc25 by Cdc2. CDK-Speedy/Ringo complexes – which do not require activating phosphorylation by CAK and are less sensitive to inhibitory phosphorylation and CDK inhibitors such as p21 – would be in a strong position to jump-start a positive feedback loop or to reverse the inhibition of Cdc25 caused by stresses such as that caused by DNA damage.
Methods
Reagents
[γ-32P]-ATP (3000 Ci/mmol) was from Dupont-NEN (Boston, MA). E. coli BL21-Codon plusRIL and BL21(DE3)-Codon plusRIL cells were from Stratagene (La Jolla, CA). Calf thymus histone H1 (Cat. #1004875) was from Roche Diagnostics Inc (Indianapolis, IN). All other chemicals were from Sigma (St. Louis, MO) unless indicated otherwise. 1 × protease inhibitor mix (PI) contained 1 mM PMSF and 10 μg/ml each of leupeptin, chymostatin, and pepstatin. EB buffer is 80 mM β-glycerophosphate, pH 7.3, 20 mM EGTA, 15 mM MgCl2, 10 mM DTT, 1 mg/ml ovalbumin, and 1 × PI. Buffer A is 20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 10 mM MgCl2, 1 mM DTT, 1 mg/ml ovalbumin, 0.1% Tween 20, 1 × protease inhibitors. 1 × TBS buffer is 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.5 mM EDTA, 0.5 mM DTT. Rabbit anti-Cdk2 antibodies were from Santa Cruz Biotech (Santa Cruz, CA). HRP-conjugated secondary antibodies and SuperSignal™ ECL reagents were from Pierce (Rockford, IL).
Protein expression and purification
GST-Speedy/Ringo A2 [39], GST-Cdk2 [31], human PP2Cα [40], GST-Cdk2 mutants [38,47], GST-Cak1p [14], Cdk7/cyclin H [38], and GST fusion substrates [24] have been described previously. The GST-Cdk2/GST-Cak1p bicistron expression plasmid was constructed by Neil Hanlon (University of Oxford, UK) and provided by Louise Johnson (University of Oxford, UK). GST-[pT160]Cdk2 was expressed as described [31]. A plasmid expressing a C-terminally his6-tagged version of the portion of bovine cyclin A3 corresponding to residues 171–432 of human cyclin A was provided by John Lew (University of California, Santa Barbara, CA) and purified as described [48]. For simplicity, we refer to this protein as cyclin A. Expression vectors for GST-Cdc25A, B, and C were provided by Anindya Dutta (University of Virginia) [49].
To coexpress Cdk2-Speedy/Ringo A2-His6 in bacteria (K2A2coexp), mouse Speedy/Ringo A2 was first amplified by polymerase chain reaction using an N-terminal primer to incorporate a HindIII site and a ribosome binding site before the start codon, and a C-terminal primer to remove the stop codon and add a XhoI site. The DNA sequences of the PCR primers are as follows (cloning sites are underlined): 5'-CCCCAAGCTTAAGGAGGGATAGCCATGGGACGGCATAATCAGATGTATTG-3' and 5'-CCCCCTCGAGTTCTTCACTCTCTGCAAACC-3'. The resulting DNA was digested with the indicated restriction enzymes and cloned into pET21d (Novagen) at the HindIII and XhoI sites, placing it in-frame with a C-terminal polyhistidine (his6) tag in the vector. Finally, HA-tagged Cdk2 was excised from an expression vector [12] using NcoI and BamHI and cloned into pET21d at the corresponding sites. The resulting K2A2coexp coexpression vector was transformed into BL21(DE3)-Codon plusRIL cells. K2A2coexp was induced using IPTG, partially purified on a metal affinity column as described [39], and loaded onto a Superdex-200 column pre-equilibrated with 1 × TBS buffer at a flow rate of 0.5 ml/min. Fractions containing Cdk2 were pooled and concentrated to about 2 mg/ml. The concentration of Cdk2 was determined by immunoblotting using GST-Cdk2 as a standard. Cdk2-Speedy/Ringo A2 complexes produced by coexpression in bacteria are designated K2A2coexp and always contained Cdk2 unphosphorylated on Thr-160.
Cdk2-Speedy/Ringo A2 complexes were also formed in vitro by mixing purified GST-Cdk2 with a three-fold molar excess of GST-Speedy/Ringo A2 in 1 × EB at room temperature for 20 min. For some experiments, the Cdk2 was first phosphorylated on Thr-160 by coexpression with Cak1p in bacteria (see above). Cdk2-cyclin A complexes were formed in vitro by mixing purified GST-[unP]Cdk2 or GST-[pT160]Cdk2 with an equal molar amount of cyclin A in 1 × EB at room temperature for 20 min.
ATPase assays
To measure the rate of ATP hydrolysis by [pT160]Cdk2-cyclin A and K2A2coexp, 16.7 μM of [pT160]Cdk2-cyclin A or K2A2coexp in 10 μl of ATPase buffer (50 mM Tris-HCl, pH 7.4, 15 mM MgCl2, 150 mM NaCl, 1 mg/ml ovalbumin, 10 mM DTT, 0.5% Tween-20, and 1 × protease inhibitors) was mixed with an equal volume of ATP mix containing 1 mM ATP and 0.5 μCi/μl [γ-32P]ATP in ATPase buffer. At each time point, 1 μl of the assay was mixed with 4 μl of Stop buffer (ATPase buffer containing 20 mM EDTA instead of 15 mM MgCl2). 1 μl of the terminated reaction was spotted onto a polyethyleneimine cellulose plate (Selecto Scientific, Norcross, GA), and chromatographed for 2 h in 50 mM HCl. Plates were dried and rechromatographed prior to PhosphorImager analysis.
In vitro kinase assay and data analysis
Cdk2-Speedy/Ringo A2 mixtures were prepared as described above. The Cdk2 concentration in the enzyme mixture was adjusted to 0.1 μg/μl for GST-Cdk2 in GST-Cdk2-Speedy/Ringo A2 complexes and to 0.055 μg/μl for Cdk2 in K2A2coexp. To determine the substrate specificity and enzymatic activity of Cdk2-Speedy/Ringo A2 (Figs. 2, 3), kinase assays were carried out by incubating 5 μl of enzymes (8.6 pmol of Cdk2) with 5 μl of substrates (13 μg GST substrates or 1.3 μg histone H1) in the presence of 0.25 μCi/μl [γ-32P]-ATP, and 0.4 mM ATP in 1 × EB at 25°C. Reactions proceeded for 10 min and were terminated by addition of 5 μl of 3 × SDS-PAGE sample buffer. Samples were resolved by 10% SDS-PAGE and analyzed by autoradiography and phosphorimaging. Substrate bands were also excised and quantified by liquid scintillation counting. To determine the phosphorylation efficiency-concentration plot (Fig. 4), 5 μl of enzymes (8.6 pmol) were incubated with 5 μl of substrates in the presence of 0.25 μCi/μl [γ-32P]-ATP, and 1 mM ATP in 1 × EB for 10 min at 25°C. The reaction was terminated by the addition of 5 μl of 3 × SDS-PAGE sample buffer. Samples containing more than 10 μg of GST substrates or 1 μg of histone H1 were diluted in 1 × sample buffer before SDS-PAGE. Samples were resolved by 10% SDS-PAGE and analyzed by autoradiography and phosphorimaging. Individual substrate bands were excised and quantified by liquid scintillation counting. Data were analyzed using the MS-excel program.
Phosphorylation of GST-Cdc25A, B, C, and histone H1
For the comparative analysis of Cdc25 and histone H1 phosphorylation by Cdk2-Speedy/Ringo A2 and [pT160]Cdk2-cyclin A (Fig. 5), 5 μl of substrate (5 μg of GST-Cdc25 or 5 μM of histone H1 in EB) was mixed with 5 μl of enzyme mix containing the indicated amounts of coexpressed Cdk2-Speedy/Ringo A2 or of [pT160]Cdk2-cyclin A in the presence of 0.25 μCi/μl [γ-32P]-ATP, 0.4 mM ATP in 1 × EB. The reactions proceeded for 10 min at room temperature and were terminated by addition of 5 μl of 3 × SDS-PAGE sample buffer. Samples were resolved by 10% SDS-PAGE and analyzed by autoradiography and phosphorimaging.
Tryptic phosphopeptide mapping of Cdc25 proteins
GST-Cdc25A, B, and C (5 μg) were phosphorylated in the presence of [γ-32P]-ATP by coexpressed Cdk2-Speedy/Ringo A2 or [pT160]Cdk2-cyclin A as described above. [32P]-Cdc25 was subjected to tryptic peptide mapping as described [50]. Briefly, samples were resolved by 10% SDS-PAGE, excised from gels, and extracted in 50 mM ammonium bicarbonate, 10% β-mercaptoethanol, and 0.2% SDS. Cdc25 was precipitated with cold trichloroacetic acid (20%) in the presence of 20 μg of bovine serum albumin as a carrier. The precipitated pellets were washed twice with ice cold acetone. Cdc25 proteins were then digested with 20 μg of sequencing grade trypsin (Promega, Cat#V5111) at 37°C overnight. The tryptic peptides were separated on thin-layer cellulose plates (EM chemicals) by horizontal electrophoresis at 1,000 V for 25 min in pH 1.9 buffer (2.5% [vol/vol] formic acid and 7.8% [vol/vol] acetic acid) followed by ascending thin-layer chromatography in 32.5% [vol/vol] n-butanol, 25% [vol/vol] pyridine, 7.5% [vol/vol] acetic acid. Following chromatography, the plate was dried and autoradiographed.
Phosphorylation/dephosphorylation of Cdk2-Speedy/Ringo A2
1 μg of GST-Cdk2D145N was preincubated with 4 μg of GST-Speedy/Ringo A2 in 10 μl of 1 × EB at room temperature for 20 min. Phosphorylation of Cdk2 on the activating phosphorylation site by yeast Cak1p and mammalian CAK was carried out as described [38]. The reaction was terminated by the addition of 3 × SDS-PAGE sample buffer. Samples were resolved in 10% SDS-PAGE and analyzed by autoradiography.
32P-labeled GST-Cdk2D145N was prepared and purified as described previously [39]. The dephosphorylation of 32P-Cdk2 by PP2C were carried out as described [39,40]. Briefly, ~80 ng of 32P-Cdk2 was incubated with an excess of cyclin A (0.5 μg), GST-Speedy/Ringo A2 (1 μg), or buffer A at room temperature for 20 min. Samples were incubated at room temperature for 10 min following the addition of 100 ng of recombinant human PP2Cα [40]. Samples were analyzed by 10% SDS-PAGE as described above.
List of abbreviations
CAK, Cdk-Activating Kinase; CDK, cyclin-dependent protein kinase; Spy1, Speedy; GST, glutathione S-transferase; PAGE, polyacrylamide gel electrophoresis.
Authors' contributions
SG carried out the experiments described in Figure 3A and performed preliminary versions of most of the experiments shown in Figures 2 and 3. PK carried out the experiment described in Figure 7B. AC carried out all other experiments. AC drafted the manuscript and prepared the figures. MS conceived the project, participated in its design and execution, and revised the manuscript. All authors participated in revising the manuscript and have read and approved the final manuscript.
Acknowledgements
We thank Louise N. Johnson, John Lew, and Anindya Dutta for expression plasmids, Stephanie Thomas for technical assistance, and Ayman S. El-Guindy and George Miller for advice on tryptic phosphopeptide mapping. For helpful discussions and critical reading of the manuscript, we thank Janet Burton, Denis Ostapenko, and Vasiliki Tsakraklides. This work was supported by grant GM47830 to M.J.S. from the National Institutes of Health, a grant to M.J.S. from the Robert Leet and Clara Guthrie Patterson Trust and by an internal grant to P.K. from the National Cancer Institute. Aiyang Cheng is a special fellow of the Leukemia & Lymphoma Society.
==== Refs
Pines J Cyclins and cyclin-dependent kinases: a biochemical view Biochem J 1995 308 697 711 8948422
Sherr CJ Roberts JM Inhibitors of mammalian G1 cyclin-dependent kinases Genes Dev 1995 9 1149 1163 7758941
Sherr CJ Roberts JM CDK inhibitors: positive and negative regulators of G1-phase progression Genes Dev 1999 13 1501 1512 10385618
King RW Deshaies RJ Peters JM Kirschner MW How proteolysis drives the cell cycle Science 1996 274 1652 1659 8939846 10.1126/science.274.5293.1652
Morgan DO The dynamics of cyclin dependent kinase structure Curr Opin Cell Biol 1996 8 767 772 8939669 10.1016/S0955-0674(96)80076-7
Morgan DO Cyclin-dependent kinases: engines, clocks, and microprocessors Annu Rev Cell Dev Biol 1997 13 261 291 9442875 10.1146/annurev.cellbio.13.1.261
Solomon MJ Kaldis P Pagano M Regulation of CDKs by phosphorylation Results and Problems in Cell Differentiation 1998 Heidelberg: Springer 79 109 9670320
Solomon MJ Glotzer M Lee TH Philippe M Kirschner MW Cyclin activation of p34cdc2 Cell 1990 63 1013 1024 2147872 10.1016/0092-8674(90)90504-8
Solomon MJ Lee T Kirschner MW Role of phosphorylation in p34cdc2 activation: identification of an activating kinase Mol Biol Cell 1992 3 13 27 1532335
Fesquet D Labbé JC Derancourt J Capony JP Galas S Girard F Lorca T Shuttleworth J Dorée M Cavadore JC The MO15 gene encodes the catalytic subunit of a protein kinase that activates cdc2 and other cyclin-dependent kinases (CDKs) through phosphorylation of Thr161 and its homologues EMBO J 1993 12 3111 3121 8344251
Poon RY Yamashita K Adamczewski JP Hunt T Shuttleworth J The cdc2-related protein p40MO15 is the catalytic subunit of a protein kinase that can activate p33cdk2 and p34cdc2 EMBO J 1993 12 3123 3132 8393783
Solomon MJ Harper JW Shuttleworth J CAK, the p34cdc2 activating kinase, contains a protein identical or closely related to p40MO15 EMBO J 1993 12 3133 3142 8344252
Solomon MJ Activation of the various cyclin/cdc2 protein kinases Curr Opin Cell Biol 1993 5 180 186 8507489 10.1016/0955-0674(93)90100-5
Kaldis P Sutton A Solomon MJ The Cdk-activating kinase (CAK) from budding yeast Cell 1996 86 553 564 8752210 10.1016/S0092-8674(00)80129-4
Espinoza FH Farrell A Erdjument-Bromage H Tempst P Morgan DO A cyclin-dependent kinase-activating kinase (CAK) in budding yeast unrelated to vertebrate CAK Science 1996 273 1714 1717 8781234
Thuret JY Valay JG Faye G Mann C Civ1 (CAK in vivo), a novel Cdk-activating kinase Cell 1996 86 565 576 8752211 10.1016/S0092-8674(00)80130-0
Peeper DS Parker LL Ewen ME Toebes M Hall FL Xu M Zantema A van der Eb AJ Piwnica-Worms H A- and B-type cyclins differentially modulate substrate specificity of cyclin-cdk complexes EMBO J 1993 12 1947 1954 8491188
Schulman BA Lindstrom DL Harlow E Substrate recruitment to cyclin-dependent kinase 2 by a multipurpose docking site on cyclin A Proc Natl Acad Sci USA 1998 95 10453 10458 9724724 10.1073/pnas.95.18.10453
Tyers M Tokiwa G Futcher B Comparison of the Saccharomyces cerevisiae G1 cyclins: Cln3 may be an upstream activator of Cln1, Cln2 and other cyclins EMBO J 1993 12 1955 1968 8387915
Levine K Huang K Cross FR Saccharomyces cerevisiae G1 cyclins differ in their intrinsic functional specificities Mol Cell Biol 1996 16 6794 6803 8943334
Pines J Hunter T Human cyclins A and B1 are differentially located in the cell and undergo cell cycle-dependent nuclear transport J Cell Biol 1991 115 1 17 1717476 10.1083/jcb.115.1.1
Nigg EA The substrates of the cdc2 kinase Semin Cell Biol 1991 2 261 270 1842344
Songyang Z Blechner S Hoagland N Hoekstra MF Piwnica-Worms H Cantley LC Use of an oriented peptide library to determine the optimal substrates of protein kinases Curr Biol 1994 4 973 982 7874496 10.1016/S0960-9822(00)00221-9
Holmes JK Solomon MJ A predictive scale for evaluating cyclin-dependent kinase substrates. A comparison of p34cdc2 and p33cdk2 J Biol Chem 1996 271 25240 25246 8810285 10.1074/jbc.271.41.25240
Lenormand JL Dellinger RW Knudsen KE Subramani S Donoghue DJ Speedy: a novel cell cycle regulator of the G2/M transition EMBO J 1999 18 1869 1877 10202150 10.1093/emboj/18.7.1869
Ferby I Blazquez M Palmer A Eritja R Nebreda AR A novel p34(cdc2)-binding and activating protein that is necessary and sufficient to trigger G(2)/M progression in Xenopus oocytes Genes Dev 1999 13 2177 2189 10465793
Karaiskou A Perez LH Ferby I Ozon R Jessus C Nebreda AR Differential regulation of Cdc2 and Cdk2 by RINGO and cyclins J Biol Chem 2001 276 36028 36034 11461916 10.1074/jbc.M104722200
Porter LA Dellinger RW Tynan JA Barnes EA Kong M Lenormand JL Donoghue DJ Human Speedy: a novel cell cycle regulator that enhances proliferation through activation of Cdk2 J Cell Biol 2002 157 357 366 11980914 10.1083/jcb.200109045
Stevenson LM Deal MS Hagopian JC Lew J Activation mechanism of CDK2: role of cyclin binding versus phosphorylation Biochemistry 2002 41 8528 8534 12081504 10.1021/bi025812h
Holmes JK Solomon MJ The role of Thr160 phosphorylation of Cdk2 in substrate recognition Eur J Biochem 2001 268 4647 4652 11532001 10.1046/j.1432-1327.2001.02392.x
Brown NR Noble ME Endicott JA Johnson LN The structural basis for specificity of substrate and recruitment peptides for cyclin-dependent kinases Nat Cell Biol 1999 1 438 443 10559988 10.1038/15674
Tarricone C Dhavan R Peng J Areces LB Tsai LH Musacchio A Structure and regulation of the CDK5-p25(nck5a) complex Mol Cell 2001 8 657 669 11583627 10.1016/S1097-2765(01)00343-4
Brown NR Noble ME Lawrie AM Morris MC Tunnah P Divita G Johnson LN Endicott JA Effects of phosphorylation of threonine 160 on cyclin-dependent kinase 2 structure and activity J Biol Chem 1999 274 8746 8756 10085115 10.1074/jbc.274.13.8746
Cheng A Xiong W Ferrell JE JrSolomon MJ Identification and comparative analysis of multiple mammalian Speedy/Ringo proteins Cell Cycle 2005 4 155 165 15611625
Busino L Chiesa M Draetta GF Donzelli M Cdc25A phosphatase: combinatorial phosphorylation, ubiquitylation and proteolysis Oncogene 2004 23 2050 2056 15021892 10.1038/sj.onc.1207394
Cans C Ducommun B Baldin V Proteasome-dependent degradation of human CDC25B phosphatase Mol Biol Rep 1999 26 53 57 10363647 10.1023/A:1006912105352
Nilsson I Hoffmann I Cell cycle regulation by the Cdc25 phosphatase family Prog Cell Cycle Res 2000 4 107 114 10740819
Kaldis P Russo AA Chou HS Pavletich NP Solomon MJ Human and yeast cdk-activating kinases (CAKs) display distinct substrate specificities Mol Biol Cell 1998 9 2545 2560 9725911
Cheng A Ross KE Kaldis P Solomon MJ Dephosphorylation of cyclin-dependent kinases by type 2C protein phosphatases Genes Dev 1999 13 2946 2957 10580002 10.1101/gad.13.22.2946
Cheng A Kaldis P Solomon MJ Dephosphorylation of human cyclin-dependent kinases by protein phosphatase type 2C alpha and beta 2 isoforms J Biol Chem 2000 275 34744 34749 10934208 10.1074/jbc.M006210200
Song H Hanlon N Brown NR Noble ME Johnson LN Barford D Phosphoprotein-protein interactions revealed by the crystal structure of kinase-associated phosphatase in complex with phosphoCDK2 Mol Cell 2001 7 615 626 11463386 10.1016/S1097-2765(01)00208-8
Chen J Saha P Kornbluth S Dynlacht BD Dutta A Cyclin-binding motifs are essential for the function of p21CIP1 Mol Cell Biol 1996 16 4673 4682 8756624
Adams PD Li X Sellers WR Baker KB Leng X Harper JW Taya Y Kaelin WG Jr Retinoblastoma protein contains a C-terminal motif that targets it for phosphorylation by cyclin-cdk complexes Mol Cell Biol 1999 19 1068 1080 9891042
Zhu L Enders G Lees JA Beijersbergen RL Bernards R Harlow E The pRB-related protein p107 contains two growth suppression domains: independent interactions with E2F and cyclin/cdk complexes EMBO J 1995 14 1904 1913 7743997
Loog M Morgan DO Cyclin specificity in the phosphorylation of cyclin-dependent kinase substrates Nature 2005 434 104 118 15744308 10.1038/nature03329
Archambault V Buchler NE Wilmes GM Jacobson MD Cross FR Two-faced cyclins with eyes on the targets Cell Cycle 2005 4 125 130 15611618
Kaldis P Cheng A Solomon MJ The effects of changing the site of activating phosphorylation in CDK2 from threonine to serine J Biol Chem 2000 275 32578 32584 10931829 10.1074/jbc.M003212200
Hagopian JC Kirtley MP Stevenson LM Gergis RM Russo AA Pavletich NP Parsons SM Lew J Kinetic basis for activation of CDK2/cyclin A by phosphorylation J Biol Chem 2001 276 275 280 11029468 10.1074/jbc.M007337200
Saha P Eichbaum Q Silberman ED Mayer BJ Dutta A p21CIP1 and Cdc25A: competition between an inhibitor and an activator of cyclin-dependent kinases Mol Cell Biol 1997 17 4338 4345 9234691
El-Guindy AS Miller G Phosphorylation of Epstein-Barr virus ZEBRA protein at its casein kinase 2 sites mediates its ability to repress activation of a viral lytic cycle late gene by Rta J Virol 2004 78 7634 7644 15220438 10.1128/JVI.78.14.7634-7644.2004
|
16191191
|
PMC1262692
|
CC BY
|
2021-01-04 16:26:24
|
no
|
BMC Biochem. 2005 Sep 28; 6:19
|
utf-8
|
BMC Biochem
| 2,005 |
10.1186/1471-2091-6-19
|
oa_comm
|
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-2341618536010.1186/1471-2105-6-234Methodology ArticleQuality determination and the repair of poor quality spots in array experiments Tom Brian DM [email protected] Walter R [email protected] Elizabeth T [email protected] James W [email protected] Medical Research Council – Biostatistics Unit, Institute of Public Health, Robinson Way, Cambridge CB2 2SR, UK2 Department of Pathology, University of Cambridge, Cambridge CB2 1QP, UK3 Washington University, School of Medicine, Department of Molecular Microbiology, 660 S. Euclid Avenue, CB 8230, St. Louis, MO 63110, USA2005 26 9 2005 6 234 234 8 3 2005 26 9 2005 Copyright © 2005 Tom et al; licensee BioMed Central Ltd.2005Tom et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
A common feature of microarray experiments is the occurence of missing gene expression data. These missing values occur for a variety of reasons, in particular, because of the filtering of poor quality spots and the removal of undefined values when a logarithmic transformation is applied to negative background-corrected intensities. The efficiency and power of an analysis performed can be substantially reduced by having an incomplete matrix of gene intensities. Additionally, most statistical methods require a complete intensity matrix. Furthermore, biases may be introduced into analyses through missing information on some genes. Thus methods for appropriately replacing (imputing) missing data and/or weighting poor quality spots are required.
Results
We present a likelihood-based method for imputing missing data or weighting poor quality spots that requires a number of biological or technical replicates. This likelihood-based approach assumes that the data for a given spot arising from each channel of a two-dye (two-channel) cDNA microarray comparison experiment independently come from a three-component mixture distribution – the parameters of which are estimated through use of a constrained E-M algorithm. Posterior probabilities of belonging to each component of the mixture distributions are calculated and used to decide whether imputation is required. These posterior probabilities may also be used to construct quality weights that can down-weight poor quality spots in any analysis performed afterwards. The approach is illustrated using data obtained from an experiment to observe gene expression changes with 24 hr paclitaxel (Taxol ®) treatment on a human cervical cancer derived cell line (HeLa).
Conclusion
As the quality of microarray experiments affect downstream processes, it is important to have a reliable and automatic method of identifying poor quality spots and arrays. We propose a method of identifying poor quality spots, and suggest a method of repairing the arrays by either imputation or assigning quality weights to the spots. This repaired data set would be less biased and can be analysed using any of the appropriate statistical methods found in the microarray literature.
==== Body
Background
Until fairly recently, little research has concentrated on design and quality issues in microarray studies. Instead, most research has been devoted to developing methods for analysis. Indeed this is understandable as the final endpoint of any study is to answer the biological questions of interest. However, the importance of experimental design and quality control cannot be over-emphasised, as experiments that have not been designed based on sound principles are more likely to produce poor quality data, which in turn affects all downstream processes (image analysis, transformation, normalization, statistical analysis) and thus lead to unreliable or misleading results. Work by researchers such as [1-3], etc. have attempted to address the deficit in the area of experimental design. However, far fewer researchers have tackled the issue of quality control, some exceptions being the work done by [4-7].
In microarray studies, researchers are often interested in comparing two or more mRNA samples either to determine which genes are differentially expressed or to detect different subtypes [8,9]. The approach adopted in analysing the data depends not only on the question(s) of interest, but also on the quality of the microarray data. Statistical methods such as cluster analysis may be quite sensitive to the process of filtering out poor quality expression data (i.e. missing data), whilst other methods such as principal component analysis and singular value decomposition cannot be used when missing data are present in the matrix of gene expressions [10].
Missing data in microarray experiments occur for a number of different reasons, including the quality of the clone preparation and of the mRNA, image corruption, the printing process, the presence of dust or scratches on the array, saturation, incomplete hybridization etc. In the pre-processing of the raw intensity data (i.e. image extraction and analysis, normalization and transformation), quality assessment of individual spots plays an integral part. Poor quality spots, either flagged up manually or through quality measures provided by the image extraction software or developed in the literature, are usually filtered out so as to prevent bias in results. Additionally, negative background-corrected intensities become undefined if a logarithmic transformation of the data is used, and must be removed. Thus this filtering process may potentially lead to a substantial amount of missing data, in particular for cDNA experiments when a relative measure of gene expression is of interest and therefore valid intensity measurements in both the red (Cy5) and green (Cy3) channels are required. Consequently, compared to the situation when there is complete data, the efficiency and power of any analysis performed on the filtered data may be substantially reduced.
Furthermore, if data "missingness" is due to the removal of less reactive spots or intensities below some pre-defined (often arbitrary) limit of detection, then any analysis performed on the incomplete data set may be subject to unwanted bias. The incomplete data set may not be representative of the complete data set, as the filtering process may be highly selective in the types of genes affected. Thus the missingness is informative, and ways of appropriately addressing the missingness are required to obtain more relevant and meaningful results. In particular, replacing undefined logarithmic data with zeros or shifting the data by a positive constant, although common approaches, may not be valid.
It is generally agreed that experimental replication (biological replicates) and repeated measurements (technical or analytic replicates) are fundamental requirements in the experimental design of microarray experiments, as they are critical for reliably distinguishing noise from other sources of variation (important or otherwise), thereby increasing the reliability and consistency of results obtained [2,11]. Furthermore, as we will show, replication may serve the additional purpose of allowing a more thorough determination of spot quality and open the way to handling poor quality spots through imputation or weighting. In this paper we describe a modelling approach to "repair" microarray data sets that, by the filtering of poor quality or negative intensities, have missing data.
Results
Example
We apply our method (see Methods section) to background uncorrected intensity data obtained from an experiment performed to observe gene expression changes with 24 hr paclitaxel (Taxol ®) treatment on a human cervical cancer derived cell line (Hela). In this experiment the cells were treated with l0 nM paclitaxel at 50% confluency and left for 24 hours prior to RNA extraction. Six independent RNA biological replicates (six flasks each extracted individually) were created for treated samples. These were compared using identically configured 4.5 K human cDNA microarrays (HGMP, Hinxton) with a common reference DMSO-treated sample used as control. Each array had 8448 probes spotted.
To calibrate and stabilize the variance of the background uncorrected intensity data, the inverse hyperbolic sine (arsinh) transformation of [12] was applied using the "vsn" package in R (Bioconductor Project). The RI (or MA) plots (figures not shown) of the transformed data for the six arrays showed that the transformation stabilized the variance reasonably well for most of them. However, Array 2 had rather different scale and offset parameters than the other five arrays. This may be indicative of a "problem" with Array 2, although not necessarily, as the arsinh transformation may have corrected for any systematic differences between the six arrays. However, as will be seen later, Array 2 was observed to have the largest proportion of spots identified as being of poor quality amongst the six arrays.
The results of fitting the mixture model described in the Methods section to our data are presented in the tables below. The parameters α and ε were set to 3 and 0.01 respectively. Estimates of some of the location parameters, the variances and the component probabilities are shown in Table 1. The automated approach estimated that approximately 81% of the intensities in channel F635 (Cy5-channel) were of good quality, whilst 98% of the intensities in channel F532 (Cy3-channel) were estimated to be of good quality. Table 2 displays the number and proportion of poor quality (failures) spots identified by our proposed method under stochastic flagging. We observe that there are significantly more poor quality data predicted for the F635 channel than the F532 channel. Further, we observe that Array 2 has the highest proportion of failures in both channels amongst the six arrays (54% and 3% in the F635 and F532 channels respectively), thus confirming our earlier observation. This array is partly responsible for only 81% of the intensities overall in F635 being of good quality. When Array 2 is removed and the quality assessment is repeated the failure proportions for each channel are more evenly spread across the five remaining arrays (data not shown) and the percentage of good quality intensities in channel F635 improves to 90%, whilst in channel F532 it becomes 97%. Interestingly, of the double failures occurring in Array 2 through to Array 6, the percentage of unreliably low poor quality spots in both channels were approximately 5%, 83%, 60%, 33% and 87% respectively.
Table 1 Parameter estimates from the model. Estimated parameters from the mixture model used to assess quality. The μ1ck estimates are not shown, since there are a large number of genes.
F635 Channel (c = 1) F532 Channel (c = 2)
parameter estimate estimate
μ0c -0.37 -0.72
μ2c 0.99 1.45
0.11 0.23
0.11 0.23
0.22 0.47
πc(0) 0.013 0.010
πc(1) 0.808 0.976
πc(2) 0.178 0.013
Table 2 Failures in the six arrays. Numbers and proportions predicted as failing in each channel for the six arrays.
F635 Channel Failures F532 Channel Failures Double Channel Failures
Array Failures % Total Failures % Total Failures % Total
1 30 0.4 8 0.1 0 0
2 4539 54 225 3 204 2
3 1345 16 43 0.5 29 0.3
4 1180 14 13 0.2 5 0.1
5 191 2 13 0.2 9 0.1
6 444 5 58 0.7 23 0.3
Total 7729 360 270
It was reassuring that the majority of spots identified by eye as of poor quality (data not shown) prior to undertaking the automatic quality assessment analysis were also identified through the proposed method.
An example of the type of spots clearly identified by eye and also identified through use of the model as being of poor quality is shown in Figure 1. Here dust on the slide at this spot position has caused an obvious saturated flash in both channels of Array 2. By imputation under stochastic flagging, we have replaced these poor intensities with more reasonable intensity measurements that are comparable to the same spot on the other five arrays. In addition to identifying these clear "outlier" spots, note that more subtle differences between replicates that could not be picked up manually by "eyeing" the data are easily detected with the automated approach.
Figure 1 Spot quality identification. Spot quality identification. The spot on Array 2 has been identified as being of poor quality in both channels due to dust on the slide at that position.
The resulting bivariate scatter plot of the transformed data using our model (under stochastic flagging) to predict the quality labels is shown in Figure 2. Most of the data fall in the middle box, whilst the majority of data are in the upper right four panels. This indicates that most of the poor quality data were of the unreliably high variety. Note the strong positive correlation between yik1 and yik2 in each cell of Figure 2. This clearly reflects underlying correlations in the μ1k1 and μ1k2, for example. Note that this does not conflict with the conditional independence assumption for our model, which as stated in the Methods section, concerns the residuals rik1 and rik2 Note also that overlap in intensity measurements across panels is allowable, as each spot intensity measurement for a particular probe is made relative to the replicate spot intensities for that probe across arrays.
Figure 2 Bivariate scatter plot of transformed data. The bivariate scatter distribution of the transformed intensity data, yikc. z = 0, 1, 2 correspond to the poor component with unreliably low intensities, the good component and the poor component with unreliably high intensities.
Figure 3 shows the resulting bivariate residual scatter plot using the predicted quality labels. Note that because of the way spots were classified into quality categories, we observe an apparent vertical or horizontal "boundary" in some of the panels where the residuals cannot go beyond (e.g. in Panels 5 and 6). Quantile-quantile plots for the residuals predicted as good quality are shown in Figure 4. Note that there appears to be reasonable fit between the observed residuals and the expected residuals from a standard normal distribution, except in the tails. The deviations in the tails are partly due to the way spots are classified into quality categories (boundary effect described above), partly due to the constraints placed on the parameters and also due to the influence of Array 2 (especially in Channel F635).
Here the strong positive correlations that were observed in Figure 2 is now only apparent in the middle panel and the lower left-hand panel. We have experimented with more elaborate models that take account of this dependence, but the results were not substantially changed. We therefore prefer to stay with the model described in the Methods section, as it is simpler.
Figure 3 Bivariate scatter plot of residuals. The bivariate scatter distribution of the residuals, rikc. z = 0, 1, 2 correspond to the poor component with unreliably low intensities, the good component and the poor component with unreliably high intensities.
Figure 4 Quantile-Quantile plots. Quantile-Quantile plot for spots predicted as good quality in each channel.
In the microarray literature, a number of approaches have been developed for missing value estimation or imputation. These approaches range from the simple "replace missing entries with zeroes" and row-average approaches to K-nearest neighbourhood (KNN) approach and its variants such as the sequential KNN (SKNN) approach [10,13], to singular value decomposition (SVD) and Bayesian principal component analysis (BPCA) methods [10,14], to least square, regression and maximum likelihood approaches [15-18]. The most popular of these is the KNN approach, which was shown by [10] to outperform the row-average and SVD approaches. However, in terms of root mean square error (RMSE), it was shown not to perform as well as the more complex and time consuming approaches [14-18] that have been recently proposed or its variant SKNN approach, which was shown to have improved accuracy in estimation of missing data with high computational speed. Additionally, an imputation approach using Gaussian mixture clustering [19] has been developed and found to be more accurate than the SVD and KNN approaches. This imputation method is similar in spirit to our mixture modelling approach.
We have compared the imputation part of our approach with the KNN and SKNN approaches for missing data estimation. We assume that the imputed data set constructed from our example above is a "true" data set. From this true data set, we randomly select 1000 spots, and perturb their F635 (Cy5) intensities by any independent Gaussian random variable with mean ± 1 and variance 2. These probes may then indicate possible "high" or "low" poor quality spots depending on how extreme the applied perturbations. From this newly pertubated data set, we apply our method to flag poor quality spots. Of the 1000 pertubed spots, those flagged were then filtered and the KNN and SKNN methods were applied to repair the data for these spots. The root mean square errors (RMSEs) of the KNN and SKNN methods were then calculated for these flagged spots and compared to the RMSE obtained from our mixture model assuming that the flagged intensity data were replaced by the appropriate "good quality" spots' mean parameters. We repeated the above ten times and the average root mean square errors were calculated for the three imputation approaches.
It was found that on average 539 of the 1000 randomly chosen spots were flagged by the quality assessment step of our method. The RMSE for our mixture model approach was 0.475, which was substantially smaller than the RMSEs for the KNN and SKNN approaches, which were 0.733 and 0.727 respectively.
Discussion
We have demonstrated an alternative approach to assessing the spot quality in cDNA microarray experimentation. The method requires replicate arrays in order to assess whether a spot signal is a true signal or not. Its strength lies in the use of information found within and, also importantly, between arrays. Thus we are able to separate different components of variability found in microarray experiments. This then allows us to be able to identify subtle problems that cannot be detected by considering each array separately, as well as the more obvious problems such as dust and comet tails. Data that appear to be good when assessed within an array, need not be reliable when assessed against corresponding data from replicate arrays. Thus replication increases the power of detecting poor and good quality spots and therefore reduce the false positive and false negative rates.
The data from replicate arrays, in its raw or background corrected form, may in general not be comparable because of the need for separate calibration and normalization of the arrays. We chose to use the inverse hyperbolic sine transformation of [12] to do the necessary calibration and normalization. This transformation was shown to be very effective and robust when compared to alternative transformations discussed in the literature.
Our approach has the additional advantage of not filtering out data but instead imputing new data to replace the spots (in either or both channels) identified as being unreliable. Of course, it is necessary to acknowledge that these imputed data are not real data and therefore suitable measures must be taken to account for the resulting uncertainty in further analyses using these repaired data. We advocate the use of multiple imputation as a way to avoid spuriously precise results. Furthermore our method performed favourably to the KNN and SKNN missing data/imputation approaches, with the added generality/advantage that it not only repairs poor quality spots but identifies them.
Alternatively instead of imputing new data, our approach can be used to assign weights to each spot. These quality weights can then be used under various strategies to down-weight spots thought to be of uncertain quality in downstream microarray processes, such as normalization and statistical analyses. Some researchers have advocated filtering of unreliable spots to avoid biasing results. We believe that the filtering of spots does not necessarily remove biases, but may actually introduce bias if the data filtering process is informative. That is, for example, if the filtering process is highly selective in removing certain types of genes and therefore the resulting filtered data set will not be representative of the true data. As our approach is dependent on having replicates, it is natural to ask how many replicates are required. However, the number of replicates required depends on a number of factors, such as the type of microarray experiment to be performed (i.e. design and analysis issues), the reliability of the experimental system used (i.e. taking into account quality issues), the cost, etc. In the example above, five or six replicates appeared to be a reasonable number. However in other types of experiments a larger number of replicates may be required. [20] provide a useful discussion regarding this question.
Conclusion
As the quality of microarray experiments affect downstream processes, it is essential to have a reliable and automatic method of flagging and then repairing poor quality spots. We have proposed a mixture model method to accomplish this two-step process of identification and imputation, and thereby producing a repaired/complete data set which is less biased than before.
Methods
Quality assessment and the mixture model
At present, microarray spot quality is assessed via two approaches. The first is based on the physical characteristics of the spot. That is, the noise, the size, shape and position of each spot, and the development of a composite quality score that reflects these features [5,20]. The second is based on assessing the quality of spots through consistency (in terms of whether or not a spot is expressed) of replicated results from a number of similar arrays [4,7]. Below, we describe an alternative approach to quality control that addresses the identification of poor quality data and how to replace them. We discuss this approach in the case of two-channel cDNA microarrays, but believe that the method can be extended to other types of array experiments (oligonucleotide, Affymetrix chips etc.), with minor modifications. In a two-channel cDNA microarray experiment, two samples of mRNA are labelled with different fluorescent dyes, commonly Cy3 (green dye) and Cy5 (red dye), and co-hybridized onto a microarray of thousands of known cDNA clones (probes) immobilised on glass supports. The image data obtained from the experiment for each spot on the array are in the form of (Cy3, Cy5) spot or target-intensity pairs, representing the expression levels of the corresponding genes in the two mRNA samples. For each channel (Cy3 or Cy5) in each spot, the observed intensity may be thought of as one of three quality types:
Type 0 poor quality, where the observed intensity is unreliably low in relation to other replicates;
Type 1 good quality, where the observed intensity is valid; and
Type 2 poor quality, where the observed intensity is unreliably high in relation to other replicates.
Fundamentally, we maintain that omitting any of the above categories of poor quality data can lead to serious biases. Therefore we do not filter out poor quality data, but instead we model the distribution of the data in each channel conditionally independently (given the spot) as a three-component mixture distribution. For this approach to be useful, replication is required. A fuller description of the model we propose is outlined below.
Furthermore, pre-processing (normalization and transformation) of the replicate arrays is required to make them comparable to each other. We have chosen to use the inverse hyberbolic sine (arsinh) transformation of [12] (also see [21]), instead of the logarithmic transformation, to transform and normalize the data. This transformation can be written mathematically as
where aic and bic are array-dependent (i.e. ith array) and channel-specific (i.e. cth channel) parameters, is the original intensity reading for the kth spot in the cth channel of the ith array and yikc is the corresponding transformed value. This transformation is used for three reasons. Firstly, it is defined over the entire real line and thus the problem of obtaining undefined values with logarithmic transformations is avoided. Secondly, this transformation has been shown to be more effective at stabilizing the variance over the entire range of intensities for various types of microarray experiments, thus removing any relationship between the variance of the spot intensities with their means. Also this transformation behaves similarly to the logarithmic transformation for large intensities. Finally, this transformation, due to its specific array-dependent parameters, can robustly and independently calibrate (normalize) the data from each microarray. [22] have investigated, via simulation, the usefulness of different transformations and found that, in a variety of situations, the arsinh transformation performs well in terms of straightening the curvature seen in RI (MA) plots and in stabilizing the variance of the microarray data. Additionally, they found it to be one of the transformations providing the greatest increase in power (compared to the logarithmic transformation) to distinguish differential genes from non-differential genes, with the detectable fold change being only slightly reduced. Of course, variance stabilization following the use of the arsinh transformation nevertheless needs to be checked in each case by, for example, looking at the MA plot.
We assume that the observed data on this transformed scale can be modelled independently for each channel, conditional on the spot's true mean μ1kc, as a mixture of three normal distributions corresponding to the three quality types defined above. Essentially, therefore, our assumption of conditional independence concerns the independence of the residual noise in each observation (see equation (3)), not of the observations themselves. The assumption of normality is in keeping with how this transformation was developed, and fits in with the standard methodological assumptions made when fitting microarray data. Note that, at the outset, we do not know to which of the three components each observed intensity belongs. Our task is to infer this from the replicate data. Denoting an observed (Cy3, Cy5) transformed target-intensity pair for the kth spot in the ith array by the bivariate response yik = (yik1, yik2), then the mixture probability density, , is the product of the mixture probability densities of the Cy3 intensity, and of the Cy5 intensity, . For conciseness, we denote these three densities as f(yik1, yik2), f1(yik1) and f2(yik2) respectively, where we suppress the dependence on the parameters. Mathematically, we write
where
and where fc(yikc|zikc) is the conditional distribution of yikc, given that yikc, is of type Zikc and πc(z) is the prior probability that an observation from channel c is of type z, where z = 0,1, or 2.
Now the conditional distributions fc(yikc|zikc) for the Cy3 and Cy5-spot intensities are assumed to have the same form. They are described as shown below
where the means μ1k1 and μ1k2 depend on k, but all the variances: and , and the remaining means: μ01, μ02, μ21 and μ22 do not. However, to prevent non-identifiability the following constraints on μ0c, μ2c, and πc(z), are specified: μ0c ≤ min(μ1kc: ∀k), μ2c ≥ ασ1c for c = 1, 2 where α is a user-specified positive parameter, and πc(z) ≥ ε for c = 1, 2 and z = 0, 1, 2, where ε is a user-specified parameter in the range (0, 1/3). For example, α = 3 indicates that the unrelaibly high mean should be three standard deviation (σ1c) away from the true signal mean, μ1kc. An ε of 0.01 would indicate that the proportion of poor quality spots in our arrays will not be less than 2% of the total number of probes. Additionally, the following constraints on the variance parameters are added: and for c = 1, 2. That is, the measurement error associated with Type 1 data should not be greater than the variability attached to poor quality data.
We believe that the above conditional distributions have biological plausibility. Spot intensities of Type 0 are affected by either incomplete hybridization or suboptimal incorporation of the dye or strongly affected by high background noise. Our model (2.1) asserts that an observed intensity of Type 0 does not contain any information about the target at that spot. Therefore we assume that all spot intensities of Type 0 in a particular channel will have a common mean and also a common variance.
Most of the data should be of Type 1 in well performed experiments. Our model (2.2) asserts that an observed intensity of Type 1 reliably reflects biologically meaningful information about the target at that spot. We have assumed that the spot intensities for each probe (across replicate arrays) in a channel will have a probe-specific mean signal (representing gene-specific levels of up or down regulation), but a variance which is common across probes. The assumption of common variance appears a reasonable one to make especially after the transformation and normalization step is performed on the data.
Spot intensities of Type 2 are affected by dust or scratches, etc.. Our model (2.3) asserts that such intensities reflect a biologically meaningful (true) signal, μ1kc, plus a bias, μ2c, due to unwanted signal effects caused by the dust or scratches. Note that the discrimination between these three types can only be achieved through replicate data. Replication allows us to assess the reproducibility/reliability of the observed spot intensities.
Under our proposed model, we are able to construct residuals of the form
which will allow us to assess the appropriateness of our model assumptions, through, for example, quantile-quantile plots (i.e. Q-Q plots) and other graphical methods. Note that these graphical methods may only be useful after the components where spots belong are identified.
Assuming that we have N technical or biological replicates (i.e. N arrays), then the observed joint probability density (or observed likelihood) is a product of the mixture distribution (1) over the K spots and the N replicates. That is, the observed likelihood, L, takes the form
Our aim is to identify for each spot its most likely component. Where an observation is predicted to be of Type 0 or Type 2, we aim to replace it with an imputed value. To achieve this goal, we adopt the strategy below, where Points 3a and 3b represent two alternative and independent versions of assigning intensities to type.
1. Estimate the mean, variance and component probability parameters, (μ, σ2, π), through maximum likelihood, using the likelihood (4), subject to the constraints being satisfied.
2. Calculate the channel-specific posterior probabilities, θikcl = p(zikc = l|yikc, μ, σ2, π), for belonging to each of the three components for each spot in each replicate, when given the observed intensity data and the estimates obtained in Step 1 above. These posterior probabilities are given by
3. The mixture component for each channel can be assigned to each spot in each array in either of two independent ways:
a. Deterministic Flagging: Assign intensity yikc to the Type l having maximum posterior probability θikcl; or
b. Stochastic Flagging: Assign intensity yikc to the Type l, where l is sampled with probability θikcl for l = 0, 1, 2.
4. Where the Type l assigned to yikc is not 1, yikc is replaced by an imputed value sampled independently from (2.2).
Alternatively instead of following Points 3 and 4 of the above strategy, the user can assign a weight wik = θik1lθik2l' to each spot. These weights can then be used in downstream analyses with already developed software packages (e.g. LIMMA).
These strategies are implemented through use of an Expectation-Maximization (E-M) algorithm [23,24]. In our implementation, the algorithm has been modified to take into account the constraints imposed on the parameters to avoid non-identifiability. Also, to avoid the potential problem of the unboundedness of the log-likelihood, the maximum likelihood estimate of the variance parameters have been modified as shown below.
where the σ*2s are the modified (weighted) versions of σ2s, β = M/K - 1 = N - 1, for c = 1, 2, and M is the total number of observations. This modification of the variance terms is motivated by Bayesian arguments. The constraints placed on the original variance parameters also apply to these modified variances.
Note that our approach allows us to borrow strength across probes/spots in order to estimate most of the parameters in our model, the exception being the estimation of the μ1k1 and μ1k2 parameters which rely only on the observations from the replicates of the kth spot.
A variety of different stopping strategies may be applied to determine when the parameter estimates in our model have converged. These range from looking at the relative changes in the log-likelihood from one iteration of the E-M algorithm to the next, to multivariately assessing the changes in all the parameter estimates from one iteration to the next, through the use of an appropriately defined distance metric. We prefer to use the change in the log-likelihood to determine convergence. However, since for the constrained E-M algorithm the log-likelihood need not increase at each iteration, we instead stop the algorithm if at the current iteration the log-likelihood obtained is larger than at previous iterations and when the change between this maximum current value of the log-likelihood and the previous maximum value of the log-likelihood (over the previous iterations) is smaller than a pre-specified convergence value, or when a pre-specified number of iterations have been completed. Some fine tuning of the convergence value may be required, as also may be the case for the choice of initial values for the parameters.
Repairing the microarray data-set via imputation
Following the above rules in Points 3 (either 3a or 3b) and 4 above, we obtain a repaired complete intensity data set. If a single imputation of the data set is all that is required, then either deterministic flagging (Point 3a) or stochastic flagging (Point 3b) can be used.
Note that a drawback of the single imputation approach is that the imputed values are treated as if known, and therefore in future analyses using this singly repaired data set, no acknowledgement will be made of the uncertainty that results from imputing the values. That is, these analyses will ignore the variability due to imputation [24] and estimates obtained may be spuriously over-precise. Thus many researchers working in the area of missing data, recommend the use of multiple imputation over single imputation. Thus, it may be preferable to generate a few multiple repaired data sets (Point 4) using, say "stochastic flagging" as in Point 3b above to identify spots which require imputation. Subsequent analyses may then be performed on each repaired data set. Results may then be compared between these data sets to ensure that any conclusions are consistent across these data sets. Alternatively, more formal methods can be used [25]. In the example described earlier, we generate just a single imputation although our algorithm may be used to generate multiple imputations.
Authors' contributions
BDMT and WRG developed the methodology for this paper, analysed the data, and drafted the paper. ETBP and JWA designed and carried out the microarray experiments, provided the data, and contributed to the drafting of the paper. All authors read and approved the final manuscript.
Acknowledgements
JWA and ETBP were supported by an award from the Biotechnology and Biological Sciences Research Council (BBSRC), United Kingdom.
==== Refs
Kerr KM Churchill GA Experimental design for gene expression microarrays Biostatistics 2001 2 183 201 12933549 10.1093/biostatistics/2.2.183
Yang YH Speed T Design issues for cDNA microarray experiments Nature Rev Genetics 2002 3 579 588 12154381
Lee MT Whitmore GA Power and sample size for DNA microarray studies Statistics in Medicine 2002 21 3543 3570 12436455 10.1002/sim.1335
Lee MT Kuo FK Whitmore GA Sklar J Importance of replication in microarray gene expression studies: Statistical methods and evidence from repetitive cDNA hybridizations PNAS 2000 97 9834 9839 10963655 10.1073/pnas.97.18.9834
Wang X Ghosh S Guo S Quantitative quality control in microarray image processing and data acquisition Nucleic Acids Res 2001 29 E75 E82 11470890 10.1093/nar/29.15.e75
Hautaniemi S Edgren H Vesanen P Wolf M Järvinen A Yli-Harja O Astola J Kallioniemi O Monni O A novel strategy for microarray quality control using Bayesian networks Bioinformatics 2003 19 2031 2038 14594707 10.1093/bioinformatics/btg275
Li C Tseng GC Wong WH Speed T Model-based analysis of oligonucleotide arrays and issues in cDNA microarray analysis Statistical Analysis of Gene Expression Microarray Data 2003 Chapman & Hall/CRC 1 34
Quackenbush J Computational analysis of microarray data Nat Rev Genet 2001 2 418 427 11389458 10.1038/35076576
Slonim DK From patterns to pathways: gene expression data analysis comes of age Nat Genet 2002 32 502 508 12454645 10.1038/ng1033
Troyanskaya O Cantor M Sherlock G Brown P Hastie T Tibshirani R Botstein D Altman RB Missing value estimation methods for DNA microarrays Bioinformatics 2001 17 520 525 11395428 10.1093/bioinformatics/17.6.520
Churchill GA Fundamentals of experimental design for cDNA microarrays Nat Genet 2002 32 490 495 12454643 10.1038/ng1031
Huber W von Heydebreck A Sültmann H Poustka A Vingron M Variance stabilization applied to microarray data calibration and to the quantification of differential expression Bioinformatics 2002 18 S96 S104 12169536
Kim KY Kim BJ Yi GS Reuse of imputed data in microarray analysis increases imputation efficiency BMC Bioinformatics 2004 5 160 15504240 10.1186/1471-2105-5-160
Oba S Sato M Takemasa I Monden M Matsubara K Ishii S A Bayesian missing value estimation methods for gene expression profile data Bioinformatics 2003 19 2088 2096 14594714 10.1093/bioinformatics/btg287
Bø TH Dysvik B Jonassen I LSimpute: accurate estimation of missing values in microarray data with least squares methods Nucleic Acids Res 2004 32 e34 14978222 10.1093/nar/gnh026
Kim H Golub GH Park H Missing value estimation for DNA microarray gene expression data: local least squares imputation Bioinformatics 2005 21 187 198 15333461 10.1093/bioinformatics/bth499
Sehgal MSB Gondal I Dooley LS Collateral missing value imputation: a new robust missing value estimation algorithm for microarray data Bioinformatics 2005 21 2417 2423 15731210 10.1093/bioinformatics/bti345
Zhou X Wang X Dougherty E Missing-value estimation using linear and non-linear regression with Bayesian gene selection Bioinformatics 2003 19 2302 2307 14630659 10.1093/bioinformatics/btg323
Ouyang M Welsh WJ Georgopoulos P Gaussian mixture clustering and imputation of microarray data Bioinformatics 2004 20 917 923 14751970 10.1093/bioinformatics/bth007
Wang X Hessner MJ Wu Y Pati N Ghosh S Quantitative quality control in microarray experiments and the application in data filtering, normalization and false positive prediction Bioinformatics 2003 19 1341 1347 12874045 10.1093/bioinformatics/btg154
Durbin BP Hardin JS Hawkins DM Rocke DM A variance-stabilizing transformation for gene-expression microarray data Bioinformatics 2002 18 S105 S110 12169537
Cui X Kerr MK Churchill GA Transformations for cDNA microarray data Statistical Applications in Genetics and Molecular Biology 2003 2 1 20
Dempster AP Laird N Rubin DB Maximum likelihood from incomplete data via the EM algorithm Journal of the Royal Statistical Society, Series B 1977 39 1 38
Tanner MA Tools for Statistical Inference 1996 3 New York: Springer-Verlag
Little RJ Rubin DB Statistical Analysis with Missing Data 2002 2 New Jersey: John Wiley and Sons
|
16185360
|
PMC1262693
|
CC BY
|
2021-01-04 16:27:27
|
no
|
BMC Bioinformatics. 2005 Sep 26; 6:234
|
utf-8
|
BMC Bioinformatics
| 2,005 |
10.1186/1471-2105-6-234
|
oa_comm
|
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-2351618802910.1186/1471-2105-6-235Methodology ArticleAn analysis of extensible modelling for functional genomics data Jones Andrew R [email protected] Norman W [email protected] School of Computer Science, University of Manchester, Manchester, UK2005 27 9 2005 6 235 235 30 6 2005 27 9 2005 Copyright © 2005 Jones and Paton; licensee BioMed Central Ltd.2005Jones and Paton; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Several data formats have been developed for large scale biological experiments, using a variety of methodologies. Most data formats contain a mechanism for allowing extensions to encode unanticipated data types. Extensions to data formats are important because the experimental methodologies tend to be fairly diverse and rapidly evolving, which hinders the creation of formats that will be stable over time.
Results
In this paper we review the data formats that exist in functional genomics, some of which have become de facto or de jure standards, with a particular focus on how each domain has been modelled, and how each format allows extensions. We describe the tasks that are frequently performed over data formats and analyse how well each task is supported by a particular modelling structure.
Conclusion
From our analysis, we make recommendations as to the types of modelling structure that are most suitable for particular types of experimental annotation. There are several standards currently under development that we believe could benefit from systematically following a set of guidelines.
==== Body
Background
The advent of large scale approaches investigating biological systems has generated a requirement for standard data formats that has been recognised by the bioinformatics community for several years. It is a major challenge to create standards that are stable and "future proof" for considerable lengths of time. In this document, we review the models associated with standard data formats for microarrays, proteomics and metabolomics (collectively known as functional genomics). The experimental techniques in these areas are evolving rapidly, different laboratories use different instruments and software, and a single experiment can produce a wide range of heterogeneous data types. This causes problems because data produced in one laboratory often cannot be interpreted by other groups or compared with other data sets produced in a different setting. Proposals have been made for data standards for microarrays (MAGE-ML [1]), protein-protein interactions (the Molecular Interaction format [2]), mass spectrometry (most recently mzData [3] and mzXML [4]), and protein separation proteomics (PEDRo [5]). There have also been proposed extensions to MAGE-ML to accommodate other types of experiment (FGE-OM [6] and SysBio-OM [7]). Data standards for metabolomics are at an early stage, but there are three models that could contribute to a data standard: SysBio-OM models metabolome data arising from NMR (nuclear magnetic resonance) and mass spectrometry; CCPN [8] is a comprehensive model of NMR data for macromolecules; and ArMet [9] covers data arising from metabolomic studies on plants.
Each proposal has been developed using various modelling strategies that enable unanticipated types of data to be encoded or to allow the model to be extended in the future. This extensibility of the models is an essential component because new functional genomics techniques are frequently developed, new instruments and software have parameters and data types that must be stored, and there are no limits on the types of biological samples that can be tested. Sufficient annotation must be captured to allow data sets to be interpreted, queried and analysed, the context must be unambiguous, and the format should capture sufficient detail about the provenance of data.
In this paper, we first describe the modelling structures that allow for extensions and the tasks that may be carried out over biological data. We analyse how well each task can be supported if information is captured within one of the extensible structures. The following section examines the extensible structures employed in the current models and highlights potential problems, in terms of tasks that may not be adequately supported. We then make recommendations as to the modelling structures that best support the most important tasks for common parts of a functional genomics workflow, and discuss the relevance of such structures to the development of new data standards.
Modelling constructs for extensibility
There are several structures that can be incorporated into models that allow additional types of data to be encoded without affecting the core schema. In this section, we first describe the modelling constructs that allow for extensibility, and then describe the kinds of tasks that may be performed over experimental data and its associated annotations, with a view to clarifying which extensibility features support which tasks.
External ontologies
External ontologies are widely used for making extensible models. Ontologies are structured controlled vocabularies containing defined terms. Each term may be associated with a set of rules or relationships to other terms that allow logical questions to be asked of the ontology. Terms from an ontology can be imported into a model, which is advantageous because the term has a meaning beyond the scope of the source system. Furthermore, where there is a standard ontology, data produced by different laboratories will use the same terms, promoting greater uniformity across different systems. The following example demonstrates the use of ontologies within MAGE-ML:
<BioSource identifier="BioSource:Drosophila:OregonR" name="Drosophila strain, Oregon R">
<MaterialType>
<OntologyEntry category="MGED:MaterialType" value="Organism"/>
</MaterialType>
<Characteristic_assnlist>
<OntologyEntry category="NCBI:Taxonomy" value="Drosophila melanogaster"/>
<OntologyEntry category="Flybase:Genotype" value="wild type"/>
<OntologyEntry category="Flybase:Strain" value="Oregon R"/>
</Characteristic_assnlist>
</BioSource>
This example demonstrates the specification of a source of material (flies) of a particular strain. The element <Characteristic_assnlist> contains a set of characteristics of the biological material using terms obtained from two different controlled vocabularies. The FlyBase ontology [10] has a definition of the "wild type" genotype and the "Oregon R" strain of flies. The NCBI taxonomy [11] is used to specify which species is being studied. The definitions can be retrieved if required to ensure that the species, strain and genotype are unambiguously described.
Name-Value-Type triples
Many of the data standards listed in the introduction have name, value, type (NVT) triples that allow additional parameters or data types to be added by the user which do not exist in a publicly available controlled vocabulary or ontology. "Name" stores the item that must be captured, "value" is the data value and "type" is a qualifier or unit. The following example is taken from the mzXML format:
<nameValue name ='heatedCapillaryTemperature' value='203.4' type='Celsius'>
In this example, there is an additional property (heatedCapillaryTemperature) that must be encoded in the data format but was not incorporated in the core schema. The parameter has a parent element that corresponds to the mass spectrometry device, demonstrating that NVT is usually context-sensitive.
External files
Additional information not covered in the data schema can be captured in separate files that are referenced from the source document. Many data formats are encoded with Extensible Markup Language (XML), which is a fairly verbose format. In some instances, information is captured in separate tab-delimited files, spreadsheets, word processing documents or image files. For example, both MAGE-OM and PEDRo specify that image data should be stored in a separate file and referenced by a URI (Uniform Resource Indicator).
Inheritance
Inheritance is used in software engineering to reduce the size of a model and make explicit areas of overlap by re-using certain components. Models are often represented in the Unified Modeling Language [12] (UML), which facilitates the design of software systems in a platform independent manner. The example in Figure 1 demonstrates how inheritance has been used in MAGE-OM, the object model that is part of the microarray data standard. The classes LabeledExtract, BioSource and BioSample are all subclasses of the general class BioMaterial. The associations between BioMaterial and other classes are inherited by LabeledExtract, BioSource and BioSample. These three classes have additional properties that make them more specific than BioMaterial.
Figure 1 The BioMaterial package in MAGE-OM. There are three subclasses BioSample, LabeledExtract and BioSource of the superclass BioMaterial.
Inheritance could be used to make a model extensible by designing a set of generic classes that describe components shared across all possible domains that use the format. Such a data model could grow over time by the addition of new subclasses containing attributes that are specific to a particular domain or to a newly emerging technology. This would have the effect that previous versions of the standard should still be supported by software, and that the standard can continuously evolve. An example of a standard developed in this way is the Geography Markup Language [13] (GML), which contains a modular structure allowing developers to use the subsets of the model that apply to their domain of interest. Inheritance has also been used extensively in MAGE-OM, the microarray object model, and the model has been extended by the addition of new subclasses in FGE-OM and SysBio-OM that do not affect classes defined in MAGE-OM. While there is no current proposal in functional genomics for the development of an evolving standard, the release of the FGE-OM and SysBio-OM models raises interesting questions as to whether this may be a feasible methodology for defining an extensible data standard. In the results section, we define this kind of extensibility as Extend Model Inheritance (EMI).
Tasks
We have identified a set of tasks that a data model must support for users. In some instances it is assumed that the data model has been implemented in a system, such as a database. The tasks are as follows:
• Search: performing of simple searches over the attributes coded in extensible structures to retrieve particular data sets.
• Share: sharing of data sets between different research groups.
• Read: manual reading of data files (or the extracted text) to understand the intention and execution of an experiment.
• Repeat experiment: the provision of sufficient detail on methods and protocols to allow an experiment to be repeated.
• Compare experiments manually: manually determine how similar different experiments are.
• Compare experiments automatically: automatically determine using a software system if two experiments are sufficiently similar to allow results to be directly compared. Although there are many other issues that may prevent automatic comparison of results, such as the use of incompatible accession numbers from different databases to identify the same objects.
• Query: querying the parts of an experiment that have been encoded in extensible structures to retrieve particular subsets of data or to ask more complex questions about the structure of the data.
• Analyse: performing of statistical or analytical processes over the data set.
• Browse: manually browsing the contents of a set of data files to find relevant experiments.
• Populate: creating data sets conforming to the standard.
There are also several tasks that fall into a different category, relating to the development and management of the model.
• Modelling: the ease with which the model can be created.
• Data capture interface: the cost to develop the user interface for populating the data format.
• Query or browse interface: the cost to develop the user interface for browsing or querying the data format.
• Data management: the cost of data management in terms of time for developers to implement changes to the database schema or additional software required for parsing.
• Quality assurance: the ability of the representation to prevent inclusion of an incorrect or imprecise value.
Support for tasks
This section presents an analysis of how well each task is supported by the different extensible structures. The support for each task by each extension is described in Table 1 and Table 2. For many of the tasks, we differentiate between performing within an organisation (the local case) and the task being performed by a user from a different organisation from where the data are produced (the non-local case).
Table 1 The support for different tasks offered by different modelling structures: NVT (Name-Value-Type), ontologies, external files and extend model inheritance (EMI).
Extension Support for task: Search
NVT Different sources will differ in attribute and value, therefore good for local data because NVT can be used to encode arbitrary properties as long as local users are aware of the data types that can be searched. Poor for non-local searches, as inconsistent attributes and values are likely to be used.
Ontology Okay if searched with exact matching terms; more difficult to support non-exact match because the search engine is unlikely to search within the ontology structure.
External file Not good; there may be no access to the structure of the file. Only information retrieval style requests can be made.
EMI Extensions can be searched locally but non-local searches will not be possible unless the extended models are shared.
Extension Support for task: Share
NVT Good for local sharing, poor for sharing externally because properties may be encoded in NVT in inconsistent ways.
Ontology Good if terms agree (if the same ontology has been used).
External file Okay if file is in a standard format, otherwise bad (information may be difficult to access).
EMI Good for local sharing; cannot be shared externally unless the extended models are shared.
Extension Support for task: Read
NVT Generally good because writer can be expressive (NVT is better than plain text); only problem is misinterpretation if NVT is used inconsistently.
Ontology Good because terms are well defined.
External file Good if file is in a standard format, otherwise bad. Other software may be required to access the file, such as for images, archive files, spreadsheets and so on.
EMI Good because writer can be as expressive as required.
Extension Support for task: Repeat Experiment
NVT Okay for local case (especially good if data capture is automated); in general it is a hard problem for the non-local case.
Ontology Good for the non-local case. May be less good for local case if local terms are converted to ontology terms and cannot be converted back (ontology may not be able to express all local data in a lossless manner).
External file Okay if file is in a standard format, otherwise bad.
EMI Good for local case, poor for non-local case unless extensions are widely shared.
Extension Support for task: Compare experiments manually
NVT Okay, but inconsistencies could be problematic if data types are encoded differently in different settings.
Ontology Good because terms are well defined and standard.
External file Okay if the file is in a standard format that can be easily processed.
EMI Generally good because the model developer can be expressive.
Extension Support for task: Compare experiments automatically
NVT Good for local case; not good for the non-local case because NVT is likely to have been implemented differently.
Ontology Good (consistent representation from different experiments).
External file Okay if data are stored in a spreadsheet or tab-delimited text and descriptive metadata are stored correctly within the data format, or if the external file is in a standard format that can be easily processed.
EMI Good for local case; cannot be done for the non-local case unless the extensions are widely shared.
Extension Support for task: Query
NVT Worse than problem for search because queries are generally more precise.
Ontology Generally good, but must query more than one language and the software for query evaluation may not be able to call out to a reasoning service (to make use of the ontology structure).
External file Not good (it must be assumed that there is no access to structure).
EMI Good for local case, cannot be queried non-locally unless the extensions are shared.
Extension Support for task: Analyse
NVT Not possible; generic analyses must not depend on such data.
Ontology May not be relevant; analysis is not usually over ontology terms (but much better than NVT if it is).
External file Okay if data are stored in a spreadsheet or tab-delimited text and metadata are stored correctly within the data format, or if the external file is in a standard format that can be easily processed.
EMI Okay for local analysis but additional wrappers may be required to allow generic analysis software to access the data. Poor for non-local case as the format will have to be interpreted and software must be written.
Extension Support for task: Browse
NVT Okay (probably better than plain text).
Ontology Good, less chance of misinterpretation than NVT.
External file Not good unless file is immediately readable.
EMI Good because writer can be expressive.
Extension Support for task: Populate
NVT Easy to populate but hard to enforce consistency.
Ontology Easy as long as ontology is in place and easily accessible.
External file Easy to populate but hard to enforce consistency.
EMI Easy.
Table 2 The relationship between the development of systems to support a data standard and the different modelling structures that could be used: NVT (Name-value-type), ontologies, external files and extend model inheritance (EMI).
Extension Support for task: Modelling
NVT Near zero cost.
Ontology Expensive (hard to develop ontology).
External file No cost.
EMI Fairly high cost because additional modelling in advance and the developer must understand the core model and how it can be extended.
Extension Support for task: Interface (for populating)
NVT Fairly easy as the code need not reflect the attributes, but difficult to ensure consistency as there is no explicit prompting from a controlled vocabulary.
Ontology Some additional costs (importing ontology or calling an ontology service)
External file Very easy (just upload the file).
EMI Changes required to the interface to reflect the extensions unless the interface is created automatically from the model.
Extension Support for task: Interface (for query/browse)
NVT Few additional costs as the interface code need not reflect the attributes.
Ontology Low cost as the queries can be generated from the model and ontology.
External file The default is no functionality over the file otherwise extra coding is required which may be relatively costly.
EMI Additional costs as interface code must be written to cover the extension unless the interface is generated from the model.
Extension Support for task: Data management
NVT Low cost as no changes are required to the schema.
Ontology No changes to database schema but small additional costs because the ontology has to be stored locally or linked externally.
External file No changes required to the schema but small additional cost because more than one storage mechanism must be managed (database and file system).
EMI Changes are required to the schema, which are likely to be expensive.
Extension Support for task: Quality assurance
NVT None; terms should be used with caution. NVT cannot restrict the cardinality or possible values.
Ontology Good because a domain value can be enforced.
External file No constraint or value checking.
EMI Some quality enforcement because there will be guidelines as to the types of extensions allowed to a model and the model will enforce constraints on the value stored.
Extensibility in biological models
MAGE-OM
A standard has been developed for microarray data, of which one part is an object model, called MAGE-OM (MicroArray and Gene Expression – Object Model), which is expressed in UML. The developers of MAGE-OM recognised that microarray technology was still evolving, that the types of experiments were fairly diverse, and that the biological samples on which experiments could be performed are practically infinite, yet all the information should be captured in a structure that would support many of the tasks described above. Therefore, several modelling constructs have been used in MAGE-OM to create a highly extensible object model.
Ontologies in MAGE-OM
MAGE-OM has many specified places in which parts of external ontologies can be imported. Examples include the characteristics of biological samples, types of biological material or compounds, and taxonomic classifications of organisms. A term can be obtained from any ontology as long as the source of the term is specified. This allows the object model to be stable but the external ontologies can grow over time with contributions from domain experts to increase the coverage of the data standard. Changes to the ontology are unlikely to cause software to fail whereas most software is dependent on the structure of the object model.
NVT triples in MAGE
The Extendable class in MAGE-OM has a relationship to a class (NameValueType) that has the attributes name, value and type (NVT). All other classes in MAGE-OM are subclasses of Extendable and inherit this relationship, allowing additional properties to be captured in NVT triples with no restrictions. In the various data repositories that support MAGE, there have been few, if any, reported uses of general NVT triples because there are usually specific classes that have been used to capture a particular concept. The inclusion of the NVT triple class could cause problems as experimental parameters encoded in this way could not be automatically compared with other experiments that have modelled parameters correctly and the values may not be capable of being queried.
External files
MAGE-OM represents processed data, resulting from image analysis, in external files containing tab-delimited data. The model captures metadata to describe what each column refers to, which is essential to ensure that when the data files are re-analysed there should be no misinterpretation of what is contained within external files. This design is advantageous because tab-delimited data files are more compact than XML. MAGE-OM also allows external image files to be specified (the raw data from the experiment), as image files tend to be in standard formats that can be interpreted by widely available software.
Extension to MAGE-OM through inheritance
There has been no formal attempt to evolve the MAGE-OM standard by the addition of new classes that inherit from parts of the core model, but there have been two proposals that have extended MAGE into other areas of functional genomics, called FGE-OM and SysBio-OM. Both models cover microarrays and proteomics, and SysBio-OM additionally covers metabolomics. In several places the two models have extended MAGE-OM through the use of inheritance. For example, both proposals include new subclasses of BioMaterial (shown in Figure 1) to model substances specific to proteome studies, such as spots on a two-dimensional gel and fractions from a column separation. The two models also create new subclasses of classes modelling a generic laboratory treatment, the inputs to the treatment and the output. A similar design is used in PEDRo (see below). It is interesting to note that several different designs have arrived at a similar method for specifying laboratory treatments, raising the possibility that MAGE-OM could become a standard that grows over time through the addition of new subclasses modelling inputs, treatments and outputs.
PEDRo
Overview
The PEDRo (Proteomics Experiment Data Repository) model was released in early 2003 to stimulate community involvement in the development of a data standard for proteomics. PEDRo consists of an object model expressed in UML, which covers protein separation techniques, such as gel electrophoresis and liquid chromatography, and protein identification using mass spectrometry. Around the same time the Proteomics Standards Initiative (PSI [14]) was founded by the Human Proteome Organisation (HUPO) to develop data standards for proteomics in the context of protein-protein interactions and mass spectrometry (MS). PEDRo has been accepted as the working model of PSI for protein separation based experiments. PEDRo is divided into four sections capturing (i) the design of the experiment and source of material, (ii) protein separation, (iii) the experimental setup for MS, and (iv) database identification of proteins with MS data. The design methodology of PEDRo is significantly different from MAGE-OM. PEDRo has detailed classes containing attributes that specify exactly what data type should be stored in which position. The model is very tightly specified and it is unlikely that experimental annotation encoded in PEDRo would be open to widespread misinterpretation. However, the model is relatively rigid and cannot easily be extended to cover unanticipated data types. PEDRo does not utilise extensible structures to describe biological samples, and therefore cannot store a structured description of all types of sample that may be used in proteomics.
Ontologies, NVT and inheritance in PEDRo
PEDRo uses ontologies in a small number of positions, such as additional parameters for database searches or unanticipated types of laboratory treatment. There are no instances of ontology usage in the database implementation (PEDRoDB [15]), due to the lack of controlled vocabularies in the proteomics field at present. There are no positions at which NVT triples are employed in PEDRo. PEDRo uses inheritance by including superclasses that capture (i) the concept of a substance (Analyte) used in a proteomics experiment and (ii) the type of processing or technique used (AnalyteProcessingStep). An AnalyteProcessingStep takes instances of Analyte as input and output. The specific details of each processing step or substance are captured in subclasses. This design could in theory be extended by adding new subclasses of AnalyteProcessingStep and Analyte. An evolving standard may be possible, although the overhead of vetting, discussing and finalising additions to the model may be prohibitively costly.
External files
Images of electrophoresis gels are represented in separate files in PEDRo, which is an acceptable solution because most users will have software that can view the majority of image file formats. PEDRo also specifies that a file containing the input parameters for MS instruments or database searches can be specified. This could cause problems if the file is not in a standard format because it will support very few of the important tasks for the user, such as query or compare experiments automatically. If the file is a proprietary format, the information may not be readable or accessible to some users at all.
Models for mass spectrometry
There have been several proposals in the past covering general MS data formats, including SpectroML [16] and ANDI [17]. We focus on three recent proposals for MS data standards: mzXML produced by the Institute for Systems Biology, mzData developed by PSI and AniML developed by ASTM [18] (an internationally recognised standards organisation). The mzXML format is a superset of the data formats produced by different instrument manufacturers, and software has been developed to convert many of the vendor specific data formats to mzXML. It is planned for future versions that controlled vocabularies will be used for vendor specific details, such as the name and type of instrument used. However, the current version of the mzXML schema does not use ontologies to capture additional information; instead, the options for terms are included within the schema. This design means that the schema can be used immediately with no additional resources required but that it cannot be extended to cover new types of technology without releasing a new schema. Additional information can be captured in the format using an NVT element that has no restrictions.
The mzData format has a similar goal to mzXML, namely to provide a single encoding of information from the different output formats produced by MS instruments. Controlled vocabularies will be used to populate many parts of the format including lists of instrument parameters, the detection mechanism, and the type of MS analysis. Supplementary information can also be captured for several objects in an element that captures the name of the object, the value and the simple data type (String, Boolean, float etc). This might cause problems because, as stated above, NVT triples may not be open to automated analysis. The source file from which the mzData file is created can be referenced using a URI. Source files are usually proprietary formats that cannot be processed by other groups. As such, there will be limited benefit in relating the mzData file back to its source, except for the purposes of local laboratory management.
The mzXML and mzData formats have very limited descriptions of the biological samples used in the experiment because it is intended that they will be used in conjunction with another data standard, such as the PSI-OM [19] model of proteome data. Various instrument parameters can be captured in both models using NVT triples, which could cause problems for querying or comparison of different data files. However, instrument parameters are unlikely to be used for searching or querying and rarely for analysis; therefore, it is possible that NVT triples are an adequate structure for encoding such information.
AniML is a model for analytical chemistry data, including the output from mass spectrometry, NMR and chromatography. AniML consists of a flexible core defined by an XML schema. There are extensions for different experimental techniques, which are XML instance documents, rather than XML schemas, defining the allowable values. This approach could be viewed as a combination between using inheritance and ontologies because specific terms are defined that should be used in particular places in the format. In this context, this is an extension of the core schema by providing more strict requirements in the form of controlled vocabularies. However, the controlled vocabularies are effectively hard-coded in the extensions.
Convergence of mass spectrometry formats
It is essential that the three formats converge to some extent to allow standardisation of mass spectrometry data files. One of the main differences is the method in which controlled vocabularies are referenced. The mzData format can include references to an external list of terms with accession numbers. In contrast, mzXML includes the terms hard-coded within the schema, although it is planned for mzXML version 2 that external CV terms will be used. AniML has specific terms in the technology specific extensions. The advantage of placing the terms outside of the schema, as in mzData and in the AniML technology instance documents, is that changes can be made to the list of terms without releasing a new schema. This has the disadvantage that additional software is required to verify that external terms have been used correctly. The mzXML format can be validated using only a standard XML Schema parser but if new terms are required, a new version of the schema must be released.
There has recently been an agreement that the same terms will ultimately be used by mzData, mzXML and AniML. Furthermore, future versions of mzXML will include references to external vocabularies, hence becoming closer to mzData in structure. It should be possible to write software that converts data between the different formats, although it is unlikely that all formats will have exactly the same coverage. It is hoped that the different organisations continue to collaborate to bring about the unification of the formats.
Molecular interaction format
A standard data format for protein interaction experiments, such as Yeast Two-Hybrid [20], has been developed by PSI called the Molecular Interaction Format (MIF), which is defined by an XML Schema. The first release of the format (level 1) covers the data that is available in most of the publicly accessible databases. PSI has developed a controlled vocabulary of terms which are used at specific places in MIF. An example term is the name of the experimental method but the format does not have a detailed description of the experimental protocols or the biological samples used. Descriptions of experimental protocols will be required in future versions because the results of protein interaction experiments are highly dependent on the technique used [21]. The Gene Ontology [22] (GO) will be used for describing genes and proteins and the NCBI Taxonomy will be used to standardise the names of species. Extensible structures may be less important for MIF because its primary use is the transfer of data between pre-existing databases. As such, the format's requirements are known in advance to some extent. If extensions are required, they can be accommodated in the next release of the standard.
Metabolomics
There are three data models that have relevance for the metabolomics community: SysBio-OM, ArMet and CCPN. SysBio-OM is an extension of MAGE-OM with the addition of new classes to model NMR data that may arise in a metabolome investigation. ArMet is a proposal from the plant metabolomics community to capture the large volumes of data that are being produced as a result of GC-MS (Gas Chromatography – Mass Spectrometry) experiments on plants. CCPN is a data model produced by the NMR community to capture details of the starting sample, the input parameters and the output from the instrument. CCPN could be used to capture metabolome data because NMR is a commonly used technique for analysing the metabolites present in a sample.
SysBio-OM has a close correspondence with MAGE-OM, and shares the same kinds of extensible modelling structures. Therefore, the comments about NVT, ontologies and external files for MAGE-OM are also relevant for SysBio-OM. CCPN contains a fairly detailed object model, and many classes have a large number of attributes that specify exactly the data types that can be captured. It is similar to the design of PEDRo in that it uses few extensible structures, although references to external databases for molecules or chemical compounds are allowed. Data files produced from CCPN are likely to be consistent and open to querying, although the format may need constant updates if there are changes in technology.
The ArMet proposal specifies that controlled vocabularies can be used for describing biological samples and chemical compounds but uses few of the extensible structures described above. The developers of ArMet suggest that the format may evolve and it could be extended through inheritance. If extensions are developed to the model, it is vital that they are widely publicised to prevent the development of different dialects of the format that cannot be compared.
Results
In this section we examine various parts of a generic functional genomics experiment (Figure 2 displays a summary of certain types of experiments), and determine the relative importance of each task listed above. We have identified the following areas that have highly similar annotation requirements across all types of experiment: the experimental hypothesis, the source of biological material, experimental protocols, numerical data and machine or software parameters.
Figure 2 The shared components in different types of functional genomics experiments. The immunohistochemistry images were obtained from .
Experimental hypothesis
The purpose of a functional genomics experiment (the hypothesis) is typically to discover the genes, proteins or metabolites that are present or expressed in a sample of interest, or those that are altered in one set of conditions compared with another. The critical difference between the conditions must be open to searching and querying. The hypothesis is often the first text that will be viewed by someone accessing the data set to determine its relevance, and it is therefore of primary importance that it can be read and browsed.
The relative importance of each task is summarised in Table 3. We believe that querying, searching, browsing and sharing are of greatest importance for experimental hypotheses. If these components are to be captured in extensible structures, ontologies are the only option that allow all these tasks to be well supported. NVT triples should not be used because querying or searching this information would be hindered.
Table 3 Importance of tasks for annotation about an experimental hypothesis (*hypothesis unlikely to be analysed).
Task Search Share Read Repeat Comp man Comp auto Query Analyse Browse Populate
Importance High High High Med High Med High Low* High High
Source of biological material
The source of material is a critical part of the experimental annotation because the results of a functional genomics investigation only have any validity within the context of the sample from which they were generated. It is important that biological samples can be queried or searched to enable users to retrieve relevant data sets, and samples must be described in a manner that allows automated comparison of experiments (Table 4). We stated that querying and searching are best supported over ontology terms and that NVT triples or external files will cause problems. It seems appropriate that efforts are focussed on designing ontologies that contain terms to describe samples, for instance as extensions to the MGED Ontology [23].
Table 4 Importance of tasks for descriptions of biological material (*analysis unlikely to be over biological samples).
Task Search Share Read Repeat Comp man Comp auto Query Analyse Browse Populate
Importance High High High Med Med High High Low* High High
Experimental protocols
The basic protocols employed in an experiment have fairly similar semantics across all functional genomics experiments, and similar representations of experimental protocols are present in several models. It is unlikely that fine details of protocols will often be searched or queried but the protocol text must be easily readable to allow manual comparison of results (Table 5). A well structured description of protocols may allow results from different experiments to be compared automatically. If NVT is used to express protocols, reading and manual comparison of experiments will be fairly well supported but automatic comparison will not be possible. The use of ontologies to capture protocols would improve facilities for automated comparison of experimental results, but the cost to model all types of protocol with controlled terms may be prohibitively high. There would be limited benefit storing protocols in external files, such as a word processing documents, compared with storing plain text within the core data format (apart from formatting).
Table 5 Tasks for experimental protocols.
Task Search Share Read Repeat Comp man Comp auto Query Analyse Browse Populate
Importance Low Med High High Med High Low Low Med High
Numerical data
The importance of the provision of support for tasks over numerical data is presented in Table 6. Examples in this context could be raw or processed data, such as the ratios of fluorescence from a microarray scan, or quantification data from a proteomics experiment. The important tasks over numerical data are analyse, share and query. There may be metadata that describes the semantics of the values, which should be described using ontology terms if possible to allow queries over the data. The actual values could be stored in an external file, such as tab-delimited text or a spreadsheet over which standard analyses can usually be performed.
Table 6 Tasks for numerical data.
Task Search Share Read Repeat Comp man Comp auto Query Analyse Browse Populate
Importance Med High Low Low Low High Med High Low High
Machine parameters
Many types of instrument and software have a set of input parameters. The most important uses for the parameters are to allow the experiment to be repeated, and to underpin automated comparison of results between two or more experiments. In many cases the equivalence of results can only be established if all the parameters are equal. Tasks such as query, search, read or browse are much less relevant (Table 7). Non-local repetition of experiments requires encodings using ontologies, and NVT should only be used to allow local repetition. However, it is unlikely that controlled vocabularies, containing parameters from all types of instrument, will exist. In this case, the use of NVT is preferable to storage in external files or in extensions to the model, because NVT should allow experiments to be compared manually, and the parameters can be accessed more easily if encoded in NVT rather than in a proprietary format.
Table 7 Tasks for machine parameters.
Task Search Share Read Repeat Comp man Comp auto Query Analyse Browse Populate
Importance Low Med Med High Low High Low Low Low Med
Discussion
It is important that data standards are created that allow flexibility in the data types that can be captured. This issue is particularly important for experiments such as proteomics, in which large volumes of data are created but the experimental methodology is frequently changing. Data models must allow for extensions that cover new technologies, otherwise a "data standard" will only cover a subset of experiment types that exist. New proposals for standards would continue to arise, or intrusive changes would be required on a regular basis. Developers should also be wary of creating models that are overly general or give users too many options for how information can be encoded. In these cases, dialects of models could arise where information can be encoded sufficiently but interpretation by different groups is difficult. We have examined data structures that allow for extensibility, identifying how well a set of tasks can be supported by data encoded in each type of structure (Table 1). The general findings are as follows. For most parts of experimental annotation, ontologies give significant advantages to users because the standardisation of terms allows for improved searching and querying, and reduces the chance of terms being misinterpreted. Ontologies also allow software to perform automated analysis to determine the similarity between different experiments. The disadvantage is that ontologies are expensive for developers to create, and present some additional costs to the user in data management (Table 2). Furthermore, ontologies will never be able to cover all the terms required by all users, because ontology development will always lag behind the creation of new experimental techniques, software or instruments. There are also issues of consistency and maintenance of ontologies which are unlikely to be resolved by official standards organisations due to the costs involved.
NVT triples give considerable flexibility to the user and they are preferable to the storage of parameters in proprietary formats because NVT encodings can be read and browsed. NVT triples are difficult to search or query though and they should not be used for data types that will be used frequently to retrieve data sets. It is important that data formats are well documented to ensure that there are guidelines for "reasonable" usage of NVT triples. In several of the data formats supplementary information about objects can be provided using ontology terms where they exist or NVT for user-defined terms. If the same user-defined terms are used frequently by different groups, this can be a mechanism for discovering new terms that should be updated in the ontology.
External files are an acceptable solution for images and for tab-delimited data if there are facilities within the core schema for capturing metadata describing the data type in each column in the file. External files should not contain information that is required for querying and searching, and they should be in a standard format that all users can process easily.
We have suggested that a data model could be developed incrementally using inheritance to add new classes that capture technology specific details. A model developed using this strategy must include generic classes that capture the concept of a laboratory technique, biological substances, raw and processed data. The core of the model should not contain details that are specific to a particular technology. Any extensions that are developed must be carefully managed to ensure that parallel development of different models covering a single domain is avoided. There must be strict guidelines for the kinds of extensions that are allowed and good documentation describing the intended usage of the core model.
We briefly touched on the issue of data quality. Data quality is a very broad concept that can be measured in a variety of ways relating to the consistency and credibility of a record [24]. Consistency can be classified into format and value consistency. Format consistency comprises rules for how the data should be parsed in terms of simple data type usage (string, integer or float), cardinality and so on. Format consistency could be verified over ontologies (if rules about the allowed syntax are encoded in the ontology) and extensions to the model, but only to a limited extent on NVT and not at all over external files. The use of an ontology is the only extensible solution which allows verification of whether values meet semantic rules (value consistency), for instance whether a genuine taxonomic name has been given for a "species" data type. Other aspects of data quality, such as credibility, often depend upon domain specific knowledge and they are difficult to control at the level of model development. The result is that sufficient annotation must be stored in structures that can be browsed, searched or queried to allow users to make judgements about data quality, and that data are open to statistical analyses.
Conclusion
We have presented a classification of structures that allow for extensibility within models that are used to create standard file formats for functional genomics. We hope that the classification will help to guide the development of new data models and standards. The first version of a protein separation standard will be released by PSI within the next year, the second version of the microarray standard MAGE-OM is also under development, and metabolomics standards are being discussed. The guidelines we have presented should maximise the potential use of data sets, while allowing good expression of the data semantics as required by the users of each format.
Methods
The modelling constructs were identified by reading the published literature and the technical documentation for each of the current proposals. We examined data repositories and software that have implemented a format to elucidate the tasks that are frequently performed. The results have been generated by cross-referencing the most important tasks for specified parts of an experiment with the modelling structures that can support those tasks.
Authors' contributions
The analysis was performed jointly by AJ and NP. AJ drafted the manuscript and revisions were made by NP.
Acknowledgements
This work was supported by BBSRC, whose support we are pleased to acknowledge. The authors would like to thank Paolo Missier for discussions on data quality issues.
==== Refs
Spellman PT Miller M Stewart J Troup C Sarkans U Chervitz S Bernhart D Sherlock G Ball C Lepage M Swiatek M Marks WL Goncalves J Markel S Iordan D Shojatalab M Pizarro A White J Hubley R Deutsch E Senger M Aronow BJ Robinson A Bassett D Stoeckert CJ JrBrazma A Design and implementation of microarray gene expression markup language (MAGE-ML) Genome Biol 2002 RESEARCH0046 12225585
Hermjakob H Montecchi-Palazzi L Bader G Wojcik J Salwinski L Ceol A Moore S Orchard S Sarkans U von Mering C Roechert B Poux S Jung E Mersch H Kersey P Lappe M Li Y Zeng R Rana D Nikolski M Husi H Brun C Shanker K Grant SG Sander C Bork P Zhu W Pandey A Brazma A Jacq B Vidal M Sherman D Legrain P Cesareni G Xenarios I Eisenberg D Steipe B Hogue C Apweiler R The HUPO PSI's molecular interaction format-a community standard for the representation of protein interaction data Nat Biotechnol 2004 22 177 183 14755292 10.1038/nbt926
PSI-MS XML Data Format
Pedrioli PG Eng JK Hubley R Vogelzang M Deutsch EW Raught B Pratt B Nilsson E Angeletti RH Apweiler R Cheung K Costello CE Hermjakob H Huang S Julian RK Kapp E McComb ME Oliver SG Omenn G Paton NW Simpson R Smith R Taylor CF Zhu W Aebersold R A common open representation of mass spectrometry data and its application to proteomics research Nat Biotechnol 2004 22 1459 1466 15529173 10.1038/nbt1031
Taylor CF Paton NW Garwood KL Kirby PD Stead DA Yin Z Deutsch EW Selway L Walker J Riba-Garcia I Mohammed S Deery MJ Howard JA Dunkley T Aebersold R Kell DB Lilley KS Roepstorff P Yates JR 3rdBrass A Brown AJ Cash P Gaskell SJ Hubbard SJ Oliver SG A systematic approach to modeling, capturing, and disseminating proteomics experimental data Nat Biotechnol 2003 21 247 254 12610571 10.1038/nbt0303-247
Jones A Hunt E Wastling JM Pizarro A Stoeckert CJ Jr An object model and database for functional genomics Bioinformatics 2004 20 1583 1590 15145818 10.1093/bioinformatics/bth130
Xirasagar S Gustafson S Merrick BA Tomer KB Stasiewicz S Chan DD Yost KJ 3rdYates JR 3rdSumner S Xiao N Waters MD CEBS Object Model for Systems Biology Data, CEBS MAGE SysBio-OM Bioinformatics 2004 20 2004 2015 15044233 10.1093/bioinformatics/bth189
Fogh R Ionides J Ulrich E Boucher W Vranken W Linge JP Habeck M Rieping W Bhat TN Westbrook J Henrick K Gilliland G Berman H Thornton J Nilges M Markley J Laue E The CCPN project: an interim report on a data model for the NMR community Nat Struct Biol 2002 9 416 418 12032555 10.1038/nsb0602-416
Jenkins H Hardy N Beckmann M Draper J Smith AR Taylor J Fiehn O Goodacre R Bino RJ Hall R Kopka J Lane GA Lange BM Liu JR Mendes P Nikolau BJ Oliver SG Paton NW Rhee S Roessner-Tunali U Saito K Smedsgaard J Sumner LW Wang T Walsh S Wurtele ES Kell DB A proposed framework for the description of plant metabolomics experiments and their results Nat Biotechnol 2004 22 1601 1606 15583675 10.1038/nbt1041
FlyBase, A Database of the Drosophila Genome
The NCBI Taxonomy Homepage
Unified Modeling Language
GML – the Geography Markup Language
Orchard S Hermjakob H Apweiler R The proteomics standards initiative Proteomics 2003 3 1374 6 12872238 10.1002/pmic.200300496
Garwood K McLaughlin T Garwood C Joens S Morrison N Taylor CF Carroll K Evans C Whetton AD Hart S Stead D Yin Z Brown AJ Hesketh A Chater K Hansson L Mewissen M Ghazal P Howard J Lilley KS Gaskell SJ Brass A Hubbard SJ Oliver SG Paton NW PEDRo: A database for storing, searching and disseminating experimental proteomics data BMC Genomics 2004 5 68 15377392 10.1186/1471-2164-5-68
Standards for Exchange of Instrument Data and NIST Chemical Reference Data – SpectroML
Standards for chromatography data systems, ASTM adopts protocols for analytical data interchange (Andi)
Analytical Information Markup Language (AnIML)
The Proteomics Standards Initiative
Fields S Song O A novel genetic system to detect protein-protein interactions Nature 1989 340 245 246 2547163 10.1038/340245a0
Cho S Park SG Lee do H Park BC Protein-protein Interaction Networks: from Interactions to Networks J Biochem Mol Biol 2004 37 45 52 14761302
The Gene Ontology Consortium Creating the gene ontology resource: design and implementation Genome Res 2001 11 1425 1433 11483584 10.1101/gr.180801
Stoeckert CJ Parkinson H The MGED ontology: a framework for describing functional genomics experiments Comp Funct Genom 2003 4 127 132 10.1002/cfg.234
Wang RY Strong DM Beyond Accuracy What Data Quality Means to Data Consumers Journal of Management Information Systems 1996 12 5 33
|
16188029
|
PMC1262694
|
CC BY
|
2021-01-04 16:27:26
|
no
|
BMC Bioinformatics. 2005 Sep 27; 6:235
|
utf-8
|
BMC Bioinformatics
| 2,005 |
10.1186/1471-2105-6-235
|
oa_comm
|
==== Front
BMC BioinformaticsBMC Bioinformatics1471-2105BioMed Central London 1471-2105-6-2381619119310.1186/1471-2105-6-238Methodology ArticleMathematical design of prokaryotic clone-based microarrays Pieterse Bart [email protected] Elisabeth J [email protected] Frank HJ [email protected] der Werf Mariët J [email protected] Wageningen Centre for Food Sciences. Diedenweg 20, 6700 AN Wageningen, The Netherlands2 TNO Quality of Life. Utrechtseweg 48, 3700 AJ Zeist, The Netherlands3 BioDetection Systems, Kruislaan 406, 1098 SM, Amsterdam, The Netherlands4 Wageningen University and Research Centre, Systems and Control Group, Department of Agrotechnology and Food Sciences, Bornsesteeg 59, 6708 PD Wageningen, The Netherlands5 HAS Den Bosch, Onderwijsboulevard 221, 5200 MA, Den Bosch, The Netherlands2005 28 9 2005 6 238 238 19 5 2005 28 9 2005 Copyright © 2005 Pieterse et al; licensee BioMed Central Ltd.2005Pieterse et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Clone-based microarrays, on which each spot represents a random genomic fragment, are a good alternative to open reading frame-based microarrays, especially for microorganisms for which the complete genome sequence is not available. Since the generation of a genomic DNA library is a random process, it is beforehand uncertain which genes are represented. Nevertheless, the genome coverage of such an array, which depends on different variables like the insert size and the number of clones in the library, can be predicted by mathematical approaches. When applying the classical formulas that determine the probability that a certain sequence is represented in a DNA library at the nucleotide level, massive amounts of clones would be necessary to obtain a proper coverage of the genome.
Results
This paper describes the development of two complementary equations for determining the genome coverage at the gene level. The first equation predicts the fraction of genes that is represented on the array in a detectable way and cover at least a set part (the minimal insert coverage) of the genomic fragment by which these genes are represented. The higher this minimal insert coverage, the larger the chance that changes in expression of a specific gene can be detected and attributed to that gene. The second equation predicts the fraction of genes that is represented in spots on the array that only represent genes from a single transcription unit, which information can be interpreted in a quantitative way.
Conclusion
Validation of these equations shows that they form reliable tools supporting optimal design of prokaryotic clone-based microarrays.
==== Body
Background
In the past decade, whole transcriptome comparison by microarray hybridizations has proven to be an effective tool for studying genome wide gene responses. The general approaches for the development of microarrays are based on the completely annotated genome sequence of an organism. Usually each spot on the array represents one open reading frame (ORF). Whereas this approach has clear advantages for strains for which the complete annotated genome sequence is available, it is not applicable to strains for which this is not the case.
A method that allows for the rapid construction of microarrays for which the completely annotated genome sequence is not required is by the construction of a clone-based array. In this approach, a chromosomal DNA library is constructed from the strain of interest. From this library the genomic fragments, the inserts, are amplified from the clones by PCR with generic primers and spotted on the array-slide [1,2].
The major differences between ORF-based and clone-based arrays with respect to the data interpretation are that in case of clone-based arrays the differential signals can only be linked to a specific gene after the DNA fragment from the spot of interest on the array has been sequenced, and that it is beforehand uncertain whether a gene is represented on the array. Moreover, whereas ORF-based microarrays exclusively generate gene specific data, a differential signal within a spot on a clone-based array can originate from multiple genes on the insert that are not necessarily linked at the transcriptional level.
The extent of these limitations can be quantified by estimating the genome coverage by the spots present on the array. The standard formulas for estimating the genome coverage of a DNA library, the Clark-Carbon equation [3] and the Lander-Waterman equation [4], determine this coverage at the nucleotide level. In other words, they consider the genome as a set of nucleotides, which is useful when the library is to be used for genome sequencing. However, these formulas will overestimate the required number of clones for hybridization purposes. The reason for this is that for hybridization purposes small overlapping fragments that allow for specific binding of the labeled cDNA suffice. Akopyants et al. [5] developed an equation for the estimation of the fraction of genes that are at least partially represented. This formula is directly derived from classical probability calculations and contains the organism specific variables genome size and average gene size. Due to the fact that Akopyants et al. determine the genome coverage at the gene level, and consider a gene represented if a fragment is present that is large enough to hybridize to and large enough to identify the gene, the required number of clones to obtain a certain coverage is reduced.
A general drawback of these three formulas is that they give no insight into the fraction of genes for which specific data can be generated in a transcriptomics experiment. The data from a spot are considered specific if the expression ratios from the quantified signal from that spot can directly be related to the gene(s) represented by the spot. This is not the case if DNA from multiple (neighboring) transcription units is present in one spot, since it would be uncertain which gene is responsible for which part of the total signal from that spot.
In this paper, two formulas were developed that enable for mathematical predictions of genome coverage by a prokaryotic clone based-array at the gene level as a function of genome size, number of clones, insert size, and either the minimal part of the insert that is covered by the gene or the minimal overlap of the gene and the insert: the minimal insert coverage (MIC) equation, and the gene specific information (GSI) equation.
In order to develop equations that are applicable to a broad range of microorganisms, model datasets were generated for 15 prokaryotes originating from several genera (Table 1) that functioned as templates on which the MIC- and GSI-equations were fitted. The resulting formulas were validated on 10 other prokaryotic species.
Table 1 Overview of prokaryotes from several genera with their genes/transcription unit-ratio. Microorganisms that were used for model development (M) or validation (V) of the MIC- and the GSI-equation are depicted in the list.
Genus
Organism
genes/TU (R) Model (M) or validation (V) strain
Proteobacteria Gammaproteobacteria Enterobacteriales Escherichia coli K-12 MG1655 1.6
Escherichia coli O157:H7 EDL933 1.6
Escherichia coli CFT073 1.6 M
Salmonella typhi CT19 1.4
Salmonella typhimurium LT2 1.6
Yersinia pestis CO92 1.4
Shigella flexneri 2a str. 2457T 1.5
Buchnera aphidicola Sg 1.5 V
Wigglesworthia glossinidia 1.5
ProteobacteriaGammaproteobacteria Pasteurellales Haemophilus influenzae Rd 1.7
Pasteurella multocida PM70 1.7 V
Proteobacteria Gammaproteobacteria Xanthomonadales Xylella fastidiosa 9a5c 1.5
Xanthomonas campestris ATCC33913 1.5 V
Proteobacteria Gammaproteobacteria Vibrionales Vibrio cholerae El Tor N16961 1.8 M
Vibrio parahaemolyticus RIMD2210633 1.5
Vibrio vulnificus CMCP6 1.5
Proteobacteria Gammaproteobacteria Pseudomonadales Pseudomonas aeruginosa PA01 1.6 M
Pseudomonas putida KT2440 1.6
Proteobacteria Gammaproteobacteria Legionellales Coxiella burnetii RSA 493 1.6
Proteobacteria Betaproteobacteria Neisseria meningitidis Z2491 1.6 M
Ralstonia solanacearum GMI1000 1.6
Proteobacteria Epsilonproteobacteria Helicobacter pylori 26695 2.3 M
Campylobacter jejuni NCTC11168 2.7 M
Proteobacteria Alphaproteobacteria Rickettsia prowazekii Nadrid E 1.4 V
Sinorhizobium meliloti 1021 1.5
Agrobacterium tumefaciens C58 1.5
Brucella suis 1330 1.5
Caulobacter crescentus 1.5
Firmicutes Bacillales Bacillus subtilis 168 1.6 M
Oceanobacillus iheyensis HTE831 1.6
Stapylococcus aureus MW2 1.6
Listeria monocytogenes EGD-e 1.8 M
Listeria innocua Clip11262 1.8
Firmicutes Clostridia Clostridium acetobutylicum ATCC824 1.6
Clostridium tetani E88 1.6
Thermoanaerobacter tengcongensis MB4T 2.0
Firmicutes Lactobacillales Lactococcus lactis IL1403 1.5 M
Streptococcus agalactiae 2603 1.8
Streptococcus pneumoniae R6 1.8
Lactobacillus plantarum WCFS1 1.6 M
Enterococcus faecalis V583 1.8
Firmicutes Mollicutes Mycoplasma pneumoniae M129 2.1 M
Mycoplasma genitalium G37 3.1 V
Mycoplasma penetrans HF-2 1.6
Ureaplasma urealyticum (serovar 3) 2.1
Actinobacteria Mycobacterium tuberculosis H37Rv 1.7 M
Corynebacterium glutamicum ATCC 13032 1.5
Streptomyces coelicolor A3(2) 1.4
Tropheryma whipplei Twist 1.9
Bifidobacterium longum NCC2705 1.3 V
Fusobacteria Fusobacterium nucleatum ATCC25586 2.0 V
Chlamydia Chlamydia trachomatis (serovar D) 1.6
Chlamydophila pneumoniae AR39 1.6
Spirochete Borrelia burgdorferi B31 1.8
Treponema pallidum Nichols 1.9
Leptospira interrogans 56601 1.5
Bacteroid Bacteroides thetaiotaomicron VPI-5482 1.8 M
Cyanobacteria Thermosynechococcus elongatus BP-1 1.6
Nostoc sp. PCC 7120 1.2
Green sulfur bacteria Chlorobium tepidum TLS 1.6
Deinococcus Deinococcus radiodurans R1 1.5 V
Hyperthermophilic bacteria Aquifex aeolicus VF5 2.1
Thermotoga maritima MSB8 3.0 V
Archae Euryarchaeota Methanococcus jannaschii DSM2661 1.8 M
Pyrococcus furiosus DSM3638 2.0 M
Archaeoglobus fulgidus DSM4304 2.1
Thermoplasma acidophilum DSM1728 1.5
Methanosarcina acetivorans C2A 1.3 V
Methanosarcina mazei Goe1 1.3
Pyrococcus abyssi 2.1
Archae Crenarchaeota Aeropyrum pernix K1 2.0
Sulfolobus solfotaricus P2 1.6
Pyrobaculum aerophilum IM2 1.7
Description of the developed equations
Minimal Insert Coverage (MIC)-equation
Since the generation of inserts for a genomic DNA library is a random process, a large part of the represented genes may be co-represented with other genes by one spot on the microarray. This complicates data interpretation since it introduces an uncertainty on which gene or genes are responsible for differential signals from these spots. The impact of differential expression of a specific gene on the observed difference of the signal from a spot will be larger when a larger part of the genome fragment in that spot is covered by that gene. Moreover, the larger the part of the insert that is covered by a specific gene, the larger the chance that differential signals for the spot can be attributed to that gene, and the higher the chance that differential expression levels from that gene result in a statistically significant differential signal on the array.
The MIC-equation anticipates to this effect by predicting the number of genes that are (at least partially) present on an insert and cover at least a predefined part of the insert (DIC). This predefined part is defined as a percentage of the total insert. E.g. if the insert size is 1000 base pairs and the predefined minimal insert coverage (DIC) is set at 50%, then at least 500 bp of that gene should be present on an insert to be considered as represented by the array. Genes smaller than the size of the predefined part of the insert, are considered as not represented on the array.
Gene Specific Information (GSI)-equation
Information on differential expression of a gene can only be quantitative and specific for that gene if it originates from a spot that only represents genes from a single transcription unit, assuming that all genes within one transcription unit are equally expressed. This was the requirement that was set for a gene to be considered represented according to the gene specific information (GSI) equation. The criteria for spots that could generate gene specific information are visualized in Fig. 1. One of the variables in the GSI-equation, the minimal overlap (Omf), allows one to set the minimal number of base pairs that are required for identification of a specific gene or transcript on an insert on the clone-based array.
Figure 1 Schematic representation of the criteria that were applied to determine whether gene specific information is generated by a specific insert. The upper line represents a genome fragment in which the block arrows represent genes. Arrows with a gray filling belong to the same transcription unit. The thinner lines represent possible locations of the inserts. The dashed lines represent inserts for which no gene specific information can be generated, since they contain genomic material that possibly belongs to another transcription unit.
Dataset preparation
Fifteen prokaryotes from various genera were selected as model species (Table 1). Genome data from these microorganisms were used for the generation of species-specific values for the expected fraction of represented genes as a function of the genome size (GS), number of clones (N), insert size (IS), and either DIC or Omf. Coordinates from all annotated genes from these organisms were obtained from GenBank, and were used to determine the gene sizes. In addition, information was obtained on the start and stop coordinates from the transcription unit to which the gene belongs, and the position of the gene in this transcription unit. It was assumed that transcription units start at the first base pair of the first gene and finish at the last base pair of the last gene. This information was generated by the combination of intergenic region based transcription unit predictions, generated by Moreno-Hagelsieb and Collado-Vides [6], with gene coordinates from GenBank.
The genome size (GS) could be included as a fictitious variable in the datasets, since not the species-specific genome size, but the species-specific gene size distribution and genome organization in genes and transcription units were relevant.
It was assumed that each possible genome fragment of the size of the insert size (IS) has an equal chance of being represented. To achieve this, fragments should be generated by physical fragmentation, and not by the use of endonucleases.
Dataset preparation for the fitting procedure for the MIC-equation
For each model organism, the fraction of the represented genes was determined for multiple combinations of the number of clones (N), fictitious genome size (GS), the insert size (IS), and the minimal insert coverage (DIC) in the ranges depicted in Table 2. In total, 140 different combinations of values for these variables were tested per strain. This was performed by first calculating the probability value of being represented per gene, and subsequently calculating the average of the probability values from all genes from the organism.
Table 2 Overview of the variables that were used for the model datasets on which the MIC- and the GSI-equation are based. Multiple combinations of the mentioned values were applied.
Variable Values
N 500; 1500; 2500; 3500; 4500; 5500; 6500; 7500; 8500; 9500
IS 100; 300; 500; 700; 900; 1100; 1300; 1500; 2100; 2700; 3000
GS 0.5; 1.5; 2.5; 3.5; 4.5; 5.5; 6.5; 7.5; 8.5; 9.5
Omf 50; 100; 150; 200; 250; 300; 350; 400; 450
DIC 10; 20; 30; 40; 50; 60; 70; 80; 90
The following formulas were developed for the calculation of the probability value per gene:
Omv=IS⋅DIC100 (1)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGpbWtdaWgaaWcbaGaemyBa0MaemODayhabeaakiabg2da9maalaaabaGaemysaKKaem4uamLaeyyXICTaemiraqKaemysaKKaem4qameabaGaeGymaeJaeGimaaJaeGimaadaaiaaxMaacaWLjaGaeiikaGIaeGymaeJaeiykaKcaaa@4077@
Gene>Omv⇒p=1−(1−Gene+1+IS−2⋅OmvGS)N (2)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGhbWrcqWGLbqzcqWGUbGBcqWGLbqzcqGH+aGpcqWGpbWtdaWgaaWcbaGaemyBa0MaemODayhabeaakiabgkDiElabdchaWjabg2da9iabigdaXiabgkHiTmaabmaabaGaeGymaeJaeyOeI0YaaSaaaeaacqWGhbWrcqWGLbqzcqWGUbGBcqWGLbqzcqGHRaWkcqaIXaqmcqGHRaWkcqWGjbqscqWGtbWucqGHsislcqaIYaGmcqGHflY1cqWGpbWtdaWgaaWcbaGaemyBa0MaemODayhabeaaaOqaaiabdEeahjabdofatbaaaiaawIcacaGLPaaadaahaaWcbeqaaiabd6eaobaakiaaxMaacaWLjaGaeiikaGIaeGOmaiJaeiykaKcaaa@5B31@
Gene <Omv ⇒ p = 0 (3)
Dataset preparation for the fittimg procedure for the GSI-equation
For each organism, the fraction of genes for which specific information could be generated was determined for 114 different combinations of the number of clones (N), fictitious genome size (GS), the insert size (IS), and the minimal required overlap (Omf) in the ranges depicted in Table 2. The represented fraction was determined by taking the average of the probability values per gene. Formulas were developed that describe different situations with respect to the localization and organization of the gene of interest on the insert (eq 4 – 15)
Formulas that were developed to determine the probability value for genes that are transcribed into a single gene transcript:
Gene≥IS⇒p=1−(1−Gene+1−ISGS)N (4)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGhbWrcqWGLbqzcqWGUbGBcqWGLbqzcqGHLjYScqWGjbqscqWGtbWucqGHshI3cqWGWbaCcqGH9aqpcqaIXaqmcqGHsisldaqadaqaaiabigdaXiabgkHiTmaalaaabaGaem4raCKaemyzauMaemOBa4MaemyzauMaey4kaSIaeGymaeJaeyOeI0IaemysaKKaem4uamfabaGaem4raCKaem4uamfaaaGaayjkaiaawMcaamaaCaaaleqabaGaemOta4eaaOGaaCzcaiaaxMaacqGGOaakcqaI0aancqGGPaqkaaa@51B5@
Gene <IS ⇒ p = 0 (5)
Formulas that were developed to determine the probability value for genes that are at the beginning of a transcription unit:
BPe ≤ IS - Omf ⇒ Oe = IS - BPe (6)
BPe > IS - Omf ⇒ Oe = Omf (7)
BPe+Gene>IS⇒p=1−(1−Gene+1−OeGS)N (8)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGcbGqcqWGqbaudaWgaaWcbaGaemyzaugabeaakiabgUcaRiabdEeahjabdwgaLjabd6gaUjabdwgaLjabg6da+iabdMeajjabdofatjabgkDiElabdchaWjabg2da9iabigdaXiabgkHiTmaabmaabaGaeGymaeJaeyOeI0YaaSaaaeaacqWGhbWrcqWGLbqzcqWGUbGBcqWGLbqzcqGHRaWkcqaIXaqmcqGHsislcqWGpbWtdaWgaaWcbaGaemyzaugabeaaaOqaaiabdEeahjabdofatbaaaiaawIcacaGLPaaadaahaaWcbeqaaiabd6eaobaakiaaxMaacaWLjaGaeiikaGIaeGioaGJaeiykaKcaaa@5606@
BPe + Gene <IS ⇒ p = 0 (9)
Formulas that were developed to determine the probability value for genes that are flanked at both sides by other genes that belong to the same transcription unit:
BPb ≤ IS - Omf ⇒ Ob = IS - BPb (10)
BPb > IS - Omf ⇒ Ob = Omf (11)
BPe ≤ IS - Omf ⇒ Oe = IS - BPe (12)
BPe > IS - Omf ⇒ Oe = Omf (13)
BPb+BPe+Gene>IS⇒p=1−(1−Gene+1+IS−Ob−OeGS)N (14)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@624C@
BPb + BPe + Gene > IS ⇒ p = 0 (15)
Models and fits
The datasets with the expected fractions of represented genes for the various combinations of parameters as presented in the previous section functioned as template for the fitting of the predictive equation for MIC and GSI.
MIC equation
From equation 2, which was used to determine the probability value per gene, it became apparent that organism-dependent gene size distribution influenced the expected number of represented genes on a clone based array. These organism dependent differences were neglected for the preparation of the MIC equation, which proofed to be justified when validating the MIC-equation (see validation section).
A polynome was developed as MIC model. In the polynome all variables were present in first and second order and in cross terms between two variables. Because of a high expected correlation between IS and DIC (based on equation 2), this relation was extended with a second order term composed of IS and DIC, resulting in:
pMIC = a + b1·DIC + b2·DIC2 + c1·N + c2·N2 + d1·GS + d2·GS2 + e1·IS + e2·IS2 + f·DIC·IS + g·DIC·N + h·DIC·GS + i·IS·N + j·IS·GS + k·GS·N + l·(IS·DIC)2 (16)
The model datasets for the 15 model species were used together in the regression procedure to estimate the parameters in the MIC model. Linear regression using a standard least squares algorithm (fminsearch) provided by Matlab (The MathWorks) was applied to search the parameters that minimize the sum of squares (SSQ) defined as:
SSQ = ∑(pMIC,exp - pMIC,mod)2 (17)
The resulting parameters are presented in Table 3. The average absolute deviation of the MIC equation from the model dataset was 0.0517.
Table 3 Values for the parameters in the MIC- and the GSI-equation.
parameter MIC equation GSI equation GSI equation
q r
a 4.85E-01 0.544 0
b1 2.54E-03 * *
b2 -1.51E-05 -4.26E-08 -3.05E-07
c1 1.27E-04 6.13E-05 1.46E-05
c2 -5.22E-09 0 -1.96E-09
d1 -1.22E-01 -7.84E-02 -1.06E-02
d2 3.42E-03 3.31E-03 3.23E-04
e1 3.95E-04 -5.36E-04 2.08E-04
e2 -9.57E-08 9.73E-08 -4.62E-08
f -9.85E-06 1.69E-08 3.42E-08
g -4.61E-07 * *
h 3.25E-04 2.55E-06 4.12E-06
I -1.69E-08 -2.22E-08 5.47E-09
j 2.01E-05 2.42E-05 -6.04E-06
k 2.26E-06 -1.76E-06 1.30E-06
l 2.60E-11 * *
ad *: this parameter is not present in the GSI equation
GSI equation
From the model datasets for the GSI equation it appeared that an organism dependent variable had a strong influence on the calculated number of represented genes (results not shown). Analysis revealed a positive correlation between the number of represented genes and the species-dependent average number of genes per transcription unit, R. R was determined by dividing the total number of genes (GenBank) by the total number of predicted transcription units [6] (Table 1).
Starting-point for the GSI model was a second order polynome for all variables, extended with the cross terms between two variables. A set of parameters was estimated for each individual model species (results not shown). Parts which appeared to contribute less than 1% to pGSI were not included, which resulted in the following relation:
pGSI = a + b2·Omf2 + c1·N + c2·N2 + d1·GS + d2·GS2 + e1·IS + e2·IS2 + f·Omf·IS + h·Omf·GS + i·IS·N + j·IS·GS + k·GS·N (18)
For each prokaryote a set of parameters was obtained by minimizing the SSQ, equivalent to equation 17. The average absolute deviation of the GSI equation from the model datasets was 0.0258.
In order to obtain one generic equation for the organism specific relations for pGSI, the species specific values for the parameters (a - k) in equation 18 were related to the species related variable R by a linear relation:
parameter(a - k) = q + r·R (19)
in which R is species specific (Table 1). Since no dependency of a with R could be established, a was set at the average of all individual a values: 0.544. With this value the polynome was fitted again, and the final relations between the other parameters and R were determined (Table 3).
Validations
In order to validate the MIC- and the GSI-equation, datasets were generated (as previously described in the "dataset preparation" section) for ten validation species (Table 1). Represented gene fractions were calculated per species for all possible combinations for the variables as presented in Table 4 and distracted from the values as they were predicted by MIC- and the GSI-equations 16 and 18, respectively. The distributions of the residuals, i.e. the difference between predicted and the calculated fraction, for both equations are presented as histograms in Figures 2a and 2b. The residual distributions of both the MIC- and the GSI-equation approach the normal distribution with a slight tendency to underestimate the fraction of represented genes (Fig. 2a and 2b). Moreover, in Table 5 the reliability of the formulas is depicted as the fraction of predictions that differ less than 0.01, 0.05 and 0.10 from the real values. It should be noted that the indicated reliabilities relate to the range of variables as depicted in Table 4.
Table 4 Overview of the variables and the values used for these variables that were used for the datasets that were used for the validation of the MIC- and the GSI-equation. All possible combinations of the mentioned values were tested.
Variable Values
N 1000; 4000; 7000; 10000
IS 100; 500; 1000; 1500; 2000
GS 1; 3; 5; 7
DIC 25; 50; 75
Omf 100; 200
Figure 2 Histogram representations of the residuals from the validation of the MIC-equation (A) and the GSI-equation (B).
Table 5 Reliability of the MIC- and the GSI-equation, depicted as the fraction of predictions that differ less than 0.01, 0.05 or 0.10 from the real values, for the validation sets defined in Table 4.
Abs (Δ predicted vs. real) Fraction for MIC-equation Fraction for GSI-equation
< 0.01 0.19 0.24
< 0.05 0.58 0.73
< 0.10 0.87 0.95
Deviations between the predicted fractions by the MIC-equation and the true values as they were determined for the validation species are mainly to be attributed to species-specific gene size distribution. In order to obtain one generic equation, and based on the accuracy of the equation in its current form (Table 5), it was decided not to include a species-specific variable.
Prediction of the optimum value for the insert size (IS)
Whereas an increase in N will always have a positive contribution to the fraction of represented genes, and an increase in GS, Omf, and MIC a negative contribution, there may be an optimum IS that depends on the values of the other variables. This optimum can be estimated by differentiation of equation 16 and 18 to IS (dp/dIS).
For the determination of the optimal value for IS for the MIC-approach this results in the following equation:
dpMICdIS=e1+2e2⋅IS+f⋅DIC+i⋅N+j⋅GS+2⋅l⋅DIC2⋅IS=0⇒ISMIC−opt=−e1−f⋅DIC−i⋅N−j⋅GS2e2+2l(DIC)2 (20)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@A1A7@
For the determination of the optimal value for IS for the GSI-approach the equation is as follows:
dpGSIdIS=e1+2e2⋅IS+f⋅O+i⋅N+j⋅GS=0⇒ISGSI−opt=−e1−f⋅O−i⋅N−j⋅GS2e2 (21)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@839D@
If the indicated values for ISopt are outside the range of 0 to 2000 bp (the range that was applied for validation of the models) no optimum can be identified within the boundaries of the model. In these cases small values of IS will give the best results.
Influence of the average number of genes per transcription unit (R) on the predicted values
From the input variables for the MIC and GSI formulas, N, IS, DIC and Omf are user-defined, while GS and R have to be estimated for the specific organism. Whereas current techniques allow for rapid and accurate estimations of GS [7-9], the organism specific value for R is difficult to determine for species from which little sequence information is available.
R was determined for 73 prokaryotes from multiple genera, as previously described in the "models and fits" section (Table 1). For 61 of the 73 strains in this list, R was within the narrow range from 1.5 – 2.0. Moreover these data indicate that accurate estimations of R can be made, based on the genus of the organism, with an exception for the mollicutes, the hyperthermophilic bacteria and the euryarchaeota.
The effect of false estimations of R was studied by the generation of validation sets as defined in Table 4 with the exception that higher or lower values for R were applied. The resulting values from the GSI-equation were compared with the true values (Table 6). It appeared that an over- or underestimation of 0.2 on R had limited effects on the fraction of predictions that differ less than 0.1 from the real values from the validation dataset (0.90 vs. 0.95 for the exact value of R). While an overestimation of 0.3 still results in 88% of the predictions that differ less than 0.1 from the real value from the validation dataset. This percentage was 80% in case of an underestimation of the same size.
Table 6 Effect of false estimations of R on the fraction of predictions that differ less than 0.01, 0.05 or 0.10 from the real values, for the validation set defined in Table 4.
Applied value for R Abs (Δ predicted vs. real) Fraction
R < 0.01 0.24
< 0.05 0.73
< 0.10 0.95
R - 0.1 < 0.01 0.17
< 0.05 0.65
< 0.10 0.95
R + 0.1 < 0.01 0.24
< 0.05 0.75
< 0.10 0.94
R - 0.2 < 0.01 0.12
< 0.05 0.55
< 0.10 0.90
R + 0.2 < 0.01 0.23
< 0.05 0.69
< 0.10 0.91
R - 0.3 < 0.01 0.10
< 0.05 0.38
< 0.10 0.81
R + 0.3 < 0.01 0.21
< 0.05 0.59
< 0.10 0.88
Application
As an example for the applicability of the developed equations, the effect of different combinations of the number of clones and insert size was determined for a prokaryote with a genome size of 4 Mbp and an estimated value for R of 1.8 using equations 16 and 18. The effect of multiple combination of N and IS on pMIC was determined for minimal insert coverage (DIC) values of 25%, 50% and 75%. The results are depicted in the contour plots in figure 3a–3c. The predicted fractions of represented genes for which gene specific information could be generated (pGSI) with a minimal overlap between the insert and the gene of 100 bp is depicted in figure 4.
Figure 3 Contour plots of the predicted fractions of represented genes with a minimal insert coverage of 25% (A), 50% (B), or 75% (C) as a function of the number of clones (N) and the insert size (IS) for a prokaryote with a genome size of 4 Mbp. The predicted fractions are depicted in the plot on top of the lines by which they are represented.
Figure 4 Contour plot of the predicted fraction of represented genes for which gene specific information could be generated as a function of the number of clones (N) and the insert size (IS) for a prokaryote with a genome size of 4 Mbp, an average number of genes per transcription unit (R) of 1.8, and a minimal overlap between the insert and the gene of 100 bp. The predicted fractions are depicted in the plot on top of the lines by which they are represented.
Plots like those depicted in figures 3a–c and 4 can be used to determine the preferred combination of the number of spots on the array and the insert size. If for instance the number of spots would be limited to 6000, an insert size of approximately 800 bp would be optimal with respect to the fraction of genes that are represented with a minimal insert coverage of 25% (Fig. 3a). From equation 20 this optimum appears to be 803 bp. With this combination of array parameters the predicted fraction of genes that cover at least 25% of the insert (which equals 803 × 0.25 = 201 bp) is 0.75 (eq. 16). Meanwhile the predicted fraction of genes for which gene specific information can be generated is 0.49 (eq. 18). If the specificity of the data is considered to be more important than the amount of represented genes, it is preferable to have an optimum value for pMIC for higher values of DIC (e.g. Fig. 3c) and a high value for pGSI (Fig. 4). These requirements are best fulfilled by combinations with low values for the insert size.
A Microsoft Excel fill in-spreadsheet that allows for calculations of pGSI, pMIC, and the optimal values for the insert size, is available as additional file with this paper [see Additional file 1].
Discussion
Classical approaches for the construction of DNA libraries form a suitable base for the construction of clone-based microarrays. However, as the construction of these libraries is a random process, it is beforehand uncertain whether a gene or transcription unit will be uniquely represented on a separate insert on the array. Genome coverage by a DNA library is usually determined by calculating the expectation that each single nucleotide from that gene is present [3,4]. These formulas will overestimate the number of clones required when the library is to be used for the construction of a microarray, since for this purpose partial representation of a gene is sufficient for hybridization.
To our knowledge, Akopyants et al. were the first to estimate genome coverage at the gene level [5]. They predicted the fraction of represented genes using equation 22:
pAkopyants=1−(1−(average transcript size+insert size−2×required overlapgenome size))number of clones (22)
MathType@MTEF@5@5@+=feaafeart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=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@B1F9@
An important variable in this formula is the average transcript size. However, use of this variable is not legitimate for this type of probability calculations since the average probability per gene (the required information) is not per se equal to the probability per average gene. When we validated the Akopyants formula on the same dataset that was applied for the validation of the MIC-equation, it appeared that 49% of the predictions deviated more than 0.1 from the real value (calculated as the average chance per gene), with a strong tendency to overestimation. The Akopyants formula therefore appears unreliable for calculating optimal library sizes
None of the previous formulas give insight in the fraction of genes for which gene specific information can be generated, while this is one of the most important features when one is interested in studying differential gene expression. The MIC-and GSI-equations that were developed in this study allow for good estimations of both the genome coverage at the gene level, and the fraction of genes for which gene specific transcription information can be generated.
Whereas the MIC-equation is rather straight-forward with respect to the input variables and interpretation, application of the GSI-equation requires the estimation of the average number of genes per transcription unit for an organism. Although a false estimation of this variable could lead to a wrong prediction of the represented fraction, Tables 1 and 6 indicate that this risk is limited.
The GSI-equation is partially based on operon predictions. For the development of the model and validation datasets we used log-likelihood based transcription unit predictions for adjacent pair of genes to be in the same operon [10]. This log-likelihood based prediction method is only applicable to organisms for which at least large parts of the genome have been sequenced, and will therefore not be useful when sequence data from array spots for which differential expression was identified, have to be interpreted. Nevertheless, good predictions can be made on whether or not genes that are co-represented in a single spot on the array belong to the same transcription unit. Strong indications can already be obtained from the physical organization of the DNA fragment of interest, like gene orientation and intergenic distance [6,11]. Other indications are the co-occurrence of genes with a joint function, and the conserved organization of homologous genes in other prokaryotes [11,12].
Conclusion
The MIC- and GSI-equations that were developed in this study were based on genomes from 15 prokaryotes from different genera, and validated on the genomes of 10 other prokaryotes. These validations show that these equations form reliable tools for optimal design of prokaryotic clone-based microarrays within the ranges that were tested (Table 4), and that they are applicable to a broad range of prokaryotes. Therefore, these equations form a good basis for the design of microarrays for prokaryotes from which the genome sequence is not available.
List of abbreviations
BPb number of base pairs within the operon in front of the specific gene [bp]
BPe number of base pairs within the operon behind the specific gene [bp]
DIC predefined minimal insert coverage, i.e. the minimal required representation of the gene on the insert [%]
Gene gene size [bp]
GS genome size [Mbp]
IS insert size [bp]
ISopt-MIC Optimal value of IS for the MIC equation [-]
ISopt-GSI Optimal value of IS for the GSI equation [-]
N number of clone [-]
Ob overlap of the fragment and the beginning of the gene [bp]
Oe overlap of fragment and the end of the gene [bp]
Omf minimal required overlap of the fragment and the gene (fixed) [bp]
Omv minimal required overlap of the fragment and the gene (variable) [bp]
p gene specific probability value [-]
pGSI predicted fraction of specifically represented genes [-]
pMIC predicted fraction of represented genes represented with a minimal insert coverage [-]
R average number of genes per transcription unit [-]
SSQ Residual Sum of Squares [-]
a-l parameters in MIC or GSI model [-]
Authors' contributions
Bart Pieterse is responsible for the original idea behind this work and performed the statistical and validation procedures. Elisabeth Quirijns developed the MIC- and GSI-equations and performed the fitting procedures. Frank Schuren provided input on the construction and application of clone based microarrays. Mariët van der Werf focused on the interpretability and applicability of the developed equations. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
The MIC- and GSI-equations (eq. 16 and 18), and the derived equations for prediction of the optimal values for IS (eq. 20 and 21), are available as a Microsoft Excel fill in spreadsheet. This spreadsheet can also be applied for the generation of contour plots in which the represented gene fractions are depicted as a function of the number of clones and the insert size.
Click here for file
Acknowledgements
We would like to thank Rolf Boesten, Martien Caspers, Nicole van Luijk and Karin Overkamp for their critical remarks and useful suggestions.
==== Refs
Cho JC Tiedje JM Bacterial species determination from DNA-DNA hybridization by using genome fragments and DNA microarrays Appl Environ Microbiol 2001 67 3677 3682 11472947 10.1128/AEM.67.8.3677-3682.2001
Askenazi M Driggers EM Holtzman DA Norman TC Iverson S Zimmer DP Boers ME Blomquist PR Martinez EJ Monreal AW Feibelman TP Mayorga ME Maxon ME Sykes K Tobin JV Cordero E Salama SR Trueheart J Royer JC Madden KT Integrating transcriptional and metabolite profiles to direct the engineering of lovastatin-producing fungal strains Nat Biotechnol 2003 21 150 156 12536215 10.1038/nbt781
Clark L Carbon J A colony bank containing synthetic Col E1 hybrids representative of the entire E. coli genome Cell 1976 9 91 99 788919 10.1016/0092-8674(76)90055-6
Lander ES Waterman MS Genomic mapping by fingerprinting random clones: a mathematical analysis Genomics 1988 2 231 239 3294162 10.1016/0888-7543(88)90007-9
Akopyants NS Clifton SW Martin J Pape D Wylie T Li L Kissinger JC Roos DS Beverley SM A survey of the Leishmania major Friedlin strain V1 genome by shotgun sequencing: a resource for DNA microarrays and expression profiling Mol Biochem Parasitol 2001 113 337 340 11295190 10.1016/S0166-6851(01)00227-4
Moreno-Hagelsieb G Collado-Vides J A powerful non-homology method for the prediction of operons in prokaryotes Bioinformatics 2002 18 S329 336 12169563
Chu G Vollrath D Davis RW Separation of large DNA molecules by contour-clamped homogeneous electric fields Science 1986 234 1582 1585 3538420
Sun LV Foster JM Tzertzinis G Ono M Bandi C Slatko BE O'Neill SL Determination of Wolbachia genome size by pulsed-field gel electrophoresis J Bacteriol 2001 183 2219 2225 11244060 10.1128/JB.183.7.2219-2225.2001
Wilhelm J Pingoud A Hahn M Real-time PCR-based method for the estimation of genome sizes Nucl Acids Res 2003 31 e56 12736322 10.1093/nar/gng056
Salgado H Moreno-Hagelsieb G Smith TF Collado-Vides J Operons in Escherichi coli: Genomic analyses and predictions Proc Nat Acad Sci 2000 97 6652 6657 10823905 10.1073/pnas.110147297
Westover BP Buhler JD Sonnenburg JL Gordon JI Operon prediction without a training set Bioinformatics 2005 21 880 888 15539453 10.1093/bioinformatics/bti123
Ermolaeva MD White O Salzberg SL Prediction of operons in microbial genomes Nucleic Acids Research 2001 29 1216 1221 11222772 10.1093/nar/29.5.1216
|
16191193
|
PMC1262695
|
CC BY
|
2021-01-04 16:27:27
|
no
|
BMC Bioinformatics. 2005 Sep 28; 6:238
|
utf-8
|
BMC Bioinformatics
| 2,005 |
10.1186/1471-2105-6-238
|
oa_comm
|
==== Front
BMC BiolBMC Biology1741-7007BioMed Central London 1741-7007-3-211620972110.1186/1741-7007-3-21Research ArticleHuman Lsg1 defines a family of essential GTPases that correlates with the evolution of compartmentalization Reynaud Emmanuel G [email protected] Miguel A [email protected] Fabien [email protected] Thi Bach Nga [email protected] Michael [email protected] Klaus [email protected] Rainer [email protected] Cell Biology and Cell Biophysics Programme, European Molecular Biology Laboratory (EMBL), Meyerhofstrasse 1, 69117 Heidelberg, Germany2 Ontario Genomics Innovation Centre, Ottawa Health Research Institute, 501 Smyth, Ottawa, ON K1H 8L6, Canada3 Structural and Computational Programme, European Molecular Biology Laboratory (EMBL), Meyerhofstrasse 1, 69117 Heidelberg, Germany2005 7 10 2005 3 21 21 7 9 2005 7 10 2005 Copyright © 2005 Reynaud et al; licensee BioMed Central Ltd.2005Reynaud et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Compartmentalization is a key feature of eukaryotic cells, but its evolution remains poorly understood. GTPases are the oldest enzymes that use nucleotides as substrates and they participate in a wide range of cellular processes. Therefore, they are ideal tools for comparative genomic studies aimed at understanding how aspects of biological complexity such as cellular compartmentalization evolved.
Results
We describe the identification and characterization of a unique family of circularly permuted GTPases represented by the human orthologue of yeast Lsg1p. We placed the members of this family in the phylogenetic context of the YlqF Related GTPase (YRG) family, which are present in Eukarya, Bacteria and Archea and include the stem cell regulator Nucleostemin. To extend the computational analysis, we showed that hLsg1 is an essential GTPase predominantly located in the endoplasmic reticulum and, in some cells, in Cajal bodies in the nucleus. Comparison of localization and siRNA datasets suggests that all members of the family are essential GTPases that have increased in number as the compartmentalization of the eukaryotic cell and the ribosome biogenesis pathway have evolved.
Conclusion
We propose a scenario, consistent with our data, for the evolution of this family: cytoplasmic components were first acquired, followed by nuclear components, and finally the mitochondrial and chloroplast elements were derived from different bacterial species, in parallel with the formation of the nucleolus and the specialization of nuclear components.
==== Body
Background
Comparative genomics is a powerful method for identifying the potential functions of previously uncharacterized genes, allowing their distribution among the kingdoms of life to be characterized, and the changes in sequence and regulation underpinning their conserved or divergent functions to be tracked [1]. Comparative genomics has been enormously facilitated by progress in bioinformatics tools, comprising the enormous amount of information available from databases concerning protein localization [2,3], viability [4,5], protein expression [6], genetic interactions [7] and protein-protein interactions [8]. These resources are usually focused on one particular organism (S. cerevisiae, C. elegans, D. melanogaster or B. subtilis) and are therefore mainly used by the small part of the scientific community working with this organism and able to handle the outcome and limitations. Attempts have been made to correlate large datasets across species, for example in the case of protein-protein interactions [9]. These cross-correlation analyses are based on the presumption that sequence and structural similarities between gene products can be used to assess functional similarities [10,11] and could in principle be extended to protein localization, viability or partners.
Genomics should be particularly powerful in the case of GTP binding proteins (or GTPases), which despite extraordinary functional diversity are all believed to have evolved from a single common ancestor [12]. As a result, all known GTPases have a conserved switch mechanism of action, core structure and sequence motifs. These proteins are found in all domains of life and are involved in such essential processes as vesicular trafficking, protein translation, intracellular signal transduction and cell cycle progression [12-14]. GTP binding proteins are often described as molecular switch proteins because of their particular mode of action. Binding and hydrolysis of GTP results in conformational changes in the so-called switch regions of the protein, which define the active GTP- and the inactive GDP-bound forms; these are used, for instance, for regulating receptor activation and cargo recruitment to membranes [12].
We have used comparative genomics to identify and characterize the human homologue of the yeast protein Lsg1. Here, we describe a novel family of GTP binding proteins, which we have named YRG (YlqF Related GTPases). Members of this family contain a central GTPase domain showing a unique circular permutation of the known G motifs of the GTP binding proteins. A phylogenetic analysis was used for cross-species comparisons, focusing on sub-cellular localization, cell viability and the known functions of each subfamily member. This analysis showed that YRG family members are essential, have increased in eukaryotes as cell compartmentalizationhas evolved, and show functional conservation in relation to rRNA maturation.
Results
Recently, we have localized more than 800 human proteins in living cells with the aim of gaining preliminary functional data [3]. Analysis of these proteins for sequences exhibiting characteristic GTPase motifs such as the P-loop [18] allowed us to identify a subset of proteins as putative GTPases.
Human Lsg1p defines a highly conserved GTPase protein family
One of these proteins possesses a central GTPase domain defined in the PFAM database of protein domains [19] as the MMR/HSR1 domain, a coiled coil, and a potential nuclear localization signal (NLS) (Figure 1A). No additional structural or enzymatic domains could be identified in the protein sequence using the SMART domain research server [20]. Interestingly, this GTPase domain is circularly permuted [21], in contrast to the canonical organization of GTPases based on the small GTPase Ras [12,22]. This circular permutation is unique, and it is surprising in view of the structure of the GTPase P-loop domain [22,23]. It implies that the four highly conserved elements of the GTPase that mediate interactions with the guanine nucleotides and effector proteins, known as the G1, G2, G3 and G4 motifs, are circularly permuted and reorganized, since G4 is followed by G1, G2 and G3 [12,22] (Figure 1A).
Figure 1 hLsg1 is member of a large circularly permuted GTPase family. (A) Schematic representation of hLsg1 – hLsg1 encodes a protein of 658 amino acids comprising a central MMR/HSR1 GTPase domain (black box, 151–500), a coiled-coil domain (Hatched box, 126–151) and a nuclear localization signal (NLS, grey box, 638–654). Domain organization of the GTPase is indicated as well as the insertion (white box inside the black box, 220–320) separating the G4 motif from the G1 motif. (B) Phylogenetic tree of the YRG superfamily. We constructed a multiple alignment of representative sequences from the YqeH, YjeQ, EngA, and YlqF families. The alignment was produced using ClustalW followed by manual editing [16]. The tree was generated from the alignment using MrBayes v3 [45] (100000 generations with parameter n chains = 4; convergence occurred after 33600 generations; the tree is the consensus of 664 trees computed using MrBayes. No molecular clock was assumed and therefore the branch lengths have no meaning. The numbers indicate the fraction of trees displaying the grouping given by the branch). The root of the tree is the one given by the MrBayes output. (C) Distribution of YRG members in cellular compartments in different organisms.
A BLAST search for similar protein sequences [15] shows that this unusual GTPase is present as a single copy per genome [22] (Figure 1B [see Additional file 1]). Only one member of the family has so far been experimentally defined, namely the Lsg1 protein in S. cerevisiae [24]. Accordingly, we named the human protein hLsg1 (human orthologue of Lsg1). All orthologues of hLsg1 possess a central MMR/HSR1 domain belonging to KOG1424 in the database of clusters of orthologous genes [25]. The identity between aligned sequences ranges from 31% to 88%. Interestingly, except in the E. cuniculi member, the GTPase domain contains an unusual insertion in comparison to the canonical GTPase structure. This insertion separates the G4 element from the remaining GTPase elements (G1, G2 and G3) (Figure 1A).
In order to elucidate the potential function of hLsg1, we extended our phylogenetic analysis. Owing to their unique structure, circularly permuted GTPases have previously been reported [22,23] and partially grouped into the Yawg/YlqF family (COG1160) [26], which is mainly restricted to prokaryotes and microbial eukaryotes (S. cerevisiae, E. cuniculi). This family contains five subfamilies: YjeQ (YloQ), MJ1464, YqeH, YlqF and Yawg. The latter three branches have eukaryotic members, YlqF representing the ancestor of hLsg1. Interestingly, while the YqeH subfamily is limited to only one member per species (also labeled as Euk-porin in sequence database), and the YjeQ subfamily is mainly restricted to bacteria [22], the YlqF subfamily shows a large expansion of this gene family in eukarya (Figure 1C). The YlqF subfamily can be further subdivided into five clades: YlqF (bacterial), MTG1 (KOG2485), LSG1 (KOG1424), NOG2 (Yawg, KOG2423), and NUG1 (KOG2484) according to the S. cerevisiae nomenclature. The YlqF family expands further in Coelomates [GNL1, 23] and in Deuterostomia (Nucleostemin [27]) (Figure 1B [see Additional file 2]).
Next, we exploited the experimental data from a comprehensive large-scale localization screen in yeast [2] and we conducted literature searches to deduce the possible cellular localizations of the different family members, ranging from the nucleolus to the mitochondria. The nucleolus is the compartment in which the large ribosomal RNA precursor (pre-rRNA) is synthesized, processed into the mature 18S, 5.8S, and 28S rRNAs and assembled with proteins to form ribosomal subunits that move to the nucleoplasm and are finally exported to the cytoplasm. Mitochondria and chloroplasts also possess a set of ribosomes. All yeast members (LSG1, NOG2, NUG1 and MTG1) are involved in ribosome biogenesis [24,28-30], and YjeQ binds to the ribosome in E. coli [31]. Finally, using ChloroP [32] to predict proteins localized to the chloroplast, we detected a sixth subfamily in YlqF, called ChYlqF (for Chloroplast YlqF), and a second subfamily in YqeH, called ChYqeH (for chloroplast YqeH). These are only found in plant genomes and group in the phylogenetic tree with the cyanobacteria YRG and YqeH members (Figure 1B [see Additional files 1 and 2]).
Nucleotide binding and GTPase activity of hLsg1
Lsg1-related proteins contain motifs that have been found to be important for guanine nucleotide binding and GTPase activity in a variety of cellular proteins [33]. Lsg1-related proteins contain the G1-4 motifs typical of GTPases (Figure 1A), suggesting that members of this family are likely to exhibit guanine nucleotide binding and GTPase activity. However, direct experimental evidence for this has been lacking so far, except for the distantly related bacterial homologues YjeQ [31], YlqF and YqeH [34]. To test this function in human Lsg1, we examined the binding of [32P] GTP to purified His tagged-hLsg1 using as control a His-tagged Sar1p, a well characterized GTPase regulating the vesicular coat complex COPII [35,36]. As shown in Figure 2A, hLsg1 binds to [32P] GTP, although more weakly than Sar1p. However, hLsg1 did not bind GDP under those experimental conditions (data not shown), which may reflect weak binding. To determine the GTPase activity of hLsg1, we performed a GTPase assay using an HPLC system as previously described [17]. In this assay, purified recombinant hLsg1 showed a low GTPase activity that proceeded to GMP and induced the further hydrolysis of GDP through GMP to guanosine. Such low GTPase activities have previously been observed in the distantly related bacterial homolog YjeQ [21], but also in GTPases in general, since their activities rely heavily on co-factors such as GAPs (GTPase Activating Proteins) [41] or GEFs (Guanine nucleotide Exchange Factors) [37]. Moreover, other GTPases such as the interferon-induced 67-kDa guanylate-binding protein (hGBP1) have been shown not to limit their hydrolysis to GDP [42]. To confirm our observations indicating that hLsg1 has GTPase activity, we immunoprecipitated endogenous hLsg1 from a HeLa cell extract using a polyclonal antibody raised against purified hLsg1, and analyzed the GTPase activity of the precipitate (Figure 2B). The GTPase activity was four times higher (incubation time 4 h compared to 18 h required for completion of GTP hydrolysis) than that of in vitro purified recombinant hLsg1 and GDP was the only final product (Figure 2B, lower panel). These data demonstrate GTPase activity in a eukaryotic member of the YlqF family for the first time.
Figure 2 Nucleotide binding and GTPase activity of hLsg1. (A) GTP binding of hLsg1. Nucleotide binding was measured as described in Materials and Methods. BSA was used as control, while Sar1p-WT and the GDP-restricted Sar1p mutant (Sar1p-T39N) were used as positive and negative GTP binding controls, respectively. The graph is the sum of three separate experiments. (B) GTPase activity of purified recombinant and immunoprecipitated hLsg1. Elution times of GDP and GTP standards are indicated (top panel). GTPase activities of purified recombinant Sar1-WT and hLsg1 are shown (middle panel) as well as GTPase activities of immunoprecipitated Sar1p and hLsg1 (lower panel). Incubation times were identical (18 h) except for the hLsg1 precipitate (4 h). (C) Hydrolysis of GTP by hLsg1. GTPase activities of purified recombinant hLsg1 were analyzed by HPLC as described in Materials and Methods. A solution containing 5 μM hLsg1 and 200 μM GTP was incubated at 37°C. Samples were taken at different time-points and analyzed for percentages of GTP and GDP.
hLsg1 is an essential protein, a characteristic of YRG family members
Yeast Lsg1, like the yeast YRG homologues NUG1, NOG2 and MTG1, is an essential protein [4]. To confirm the consequences of loss of hLsg1, we transfected siRNAs targeted against hLsg1 into HeLa cells, confirming the efficiency of the siRNA treatment by western blot analysis (Figure 3A). After 24 h, Lsg1 expression showed a drastic decrease in cells treated with Lsg1 siRNA compared to cells treated with a negative control siRNA. Moreover, hLsg1 expression in control cells or cells transfected with the negative control shows a band shift that increases with time, as observed in proteins post-translationally modified e.g. by phosphorylation. The increase in intensity could indicate that the polyclonal antibody has a higher affinity for the modified form. There was no significant change in actin expression in control cells, or in cells treated with either random siRNA or a specific hLsg1 siRNA (Figure 3A). However, microscopic observations during the course of the experiment showed that HeLa cell cultures exhibiting hLsg1 knockdown were less dense than control cells. In addition, hLsg1 knockdowns contained more apoptotic cells (not shown), suggesting a lethal effect. We confirmed this by immunostaining hLsg1-specific siRNA-treated cells with a polyclonal anti-hLsg1 antibody and staining the cell nuclei with DAPI at different times after transfection of the siRNA (Figure 3B). Cell numbers decreased rapidly after treatment with the specific hLsg1 siRNA in comparison to cells treated with oligofectamine alone or with control siRNA.
Figure 3 hLsg1 is an essential protein. (A) Rapid disappearance of hLsg1 in cells transfected with hLsg1 specific siRNA. HeLa cells were transfected with an hLsg1-specific siRNA (+) or with a scrambled siRNA as a negative control (-), and were harvested at 0, 12 and 24 h post-transfection. Extracts were prepared and 30 μg of each sample were separated on an 8% polyacrylamide gel and analyzed for hLsg1 by western blotting. Untreated cell lysate (30 μg) from confluent HeLa cells was run as a control (Co). To assess the specificity of the siRNAs, 30 μg of each extract was run on a 12% SDS-PAGE gel and analyzed for actin content by western blotting. (B) Cell count. HeLa cells plated on coverslips and transfected with no siRNA (black box), hLsg1-specific siRNA (white box), or a scrambled siRNA (grey box) were fixed at 0, 12, 24 and 36 h post-transfection. Cells were labeled with Dapi and anti-hLsg1 antibodies and the cell number was counted. The graph is the sum of three independent experiments (C) YRG family member lethality. Literature survey and database searches indicating that YRG family members are essential.
We used the large datasets from gene viability screens of bacteria, worms and flies to compare our observations with data about other YRG family members. YjeQ was shown to be indispensable for the growth of E. coli and B. subtilis [38]. In C. elegans, YRG orthologues are non-viable (t19a6.2a, t19a6.2b, k01c8.9, C53H9) (Figure 3C). Since large human RNAi screens are only now in progress, no data were available for other YRG human genes. However, interestingly, overexpression of nucleostemin was shown to be lethal [27].
According to our results, hLsg1 is essential, like its yeast counterpart, and this characteristic seems to be common to the YRG family members. This implies that each YRG protein fulfils essential functions.
hLsg1 localizes to the endoplasmic reticulum and to discrete nuclear structures
Compartmentalization of the human cell allows better control of function and reactions steps in many pathways, including ribosome assembly. Cellular localization is a key to defining protein function. Using large-scale localization screens, we previously identified hLsg1 as an endoplasmic reticulum localized protein [3], in contrast to yeast Lsg1, which is proposed to localize specifically to the cytosol [24].
We decided to confirm our preliminary data on hLsg1 localization in humans using GFP-fused constructs as well as specific polyclonal antibodies. When expressed as a C-terminally tagged YFP fusion protein, hLsg1 localized to the ER in most cells (Figure 4A, 1). In 10% of the transfected cells, however, discrete structures in the nucleus were observed and localization to the endoplasmic reticulum was decreased or even absent (Figure 4A, 2). An N-terminally tagged CFP-hLsg1 fusion protein was also localized to the ER and nuclear envelope, but more of the protein was cytosolic than in the case of the C-terminal hLsg1-YFP fusion (Figure 4A, 3). A truncated hLsg1 version (480 to 658aa) fused to the YFP, containing the potential NLS, accumulated in the nucleus and nucleolus (Figure 4A, 4). Collectively, these data indicate that the NLS present in the C-terminus of hLsg1 is functional, in contrast to the putative NLS in yeast Lsg1, which is reportedly restricted to the cytosol [24].
Figure 4 Subcellular localization analysis of hLsg1. (A) Subcellular localization of YFP and CFP tagged hLsg1 – HeLa cells were transiently transfected with hLsg1-YFP (1, 2), or CFP-hLsg1 (3) and visualized by fluorescence microscopy. (B) Localization of endogenous hLsg1. hLsg1 was visualized by staining with anti-hLsg1 antibodies, followed by Alexa 488-conjugated anti-rabbit antibodies. HeLa cells were transiently transfected with YFP-tagged lamin or Clontech ER-YFP marker. Coilin was visualized by monoclonal mouse anti-coilin antibodies, followed by rhodamine-conjugated anti-mouse antibodies. Bars = 10 μm
Immunostaining with an antibody against the entire protein showed that the endogenous protein also localized to reticular membranes, and in a fraction of the cells to a number of small punctuate nuclear structures. These results are very similar to those obtained with the hLsg1-YFP fusion protein (Figure 4A, 2). Double staining showed that hLsg1 partially co-localized with an ectopically expressed FP (fluorescent protein) used to mark the ER (Clontech ER marker) (Figure 4B, bottom row), as well as with the nuclear envelope marker lamin B1 (Figure 4B, top row). However, hLsg1 was largely absent from the Golgi complex, which was labeled with antibodies against the Golgi membrane protein golgin97, and from mitochondria, marked by antibodies against HSP60 (data not shown). Moreover, the small hLsg1-positive nuclear structures observed in a fraction of the cells co-localized with coilin, a typical marker of Cajal bodies (CBs) (indicated by arrowheads in Figure 4B, middle row). The CBs are functionally linked to the nucleolus and play a major role in the maturation of RNP, acting on the mRNA as well as the rRNA pathway [44].
These data demonstrate that in contrast to its yeast counterpart, hLsg1 localizes to the ER and to Cajal bodies in the nucleus.
hLsg1 shuttles between the nucleus and the cytosol
The dual localization of hLsg1 in the cytosol and nucleus suggests nucleocytoplasmic trafficking of the protein, possibly in relation to rRNA maturation. We constructed hLsg1 deletion mutants containing or excluding the putative the C-terminal NLS (YFP-hLsg1-1-600 and YFP-hLsg1-480-658) and transfected them into Hela cells (Figure 5A). While YFP-hLsg1-480-658 clearly localized in the nucleus, YFP-hLsg1-1-600, which contains no NLS, was excluded from the nucleus. Moreover, YFP-hLsg1-480-658 colocalized in the nucleus with SRP19-MRFP, a nucleolar marker [46], indicating that it sublocalizes to the nucleolus.
Figure 5 hLsg1 localized to the nucleus upon Leptomycin B treatment. (A) Subcellular localization of YFP tagged hLsg1 mutants – HeLa cells were transiently transfected with YFP-hLsg1-1-600 (1) or YFP-hLsg1-480-658 (2), or co-transfected with YFP-hLsg1-480-658 and SRP19-MRFP (3), and visualized by fluorescence microscopy. (B) Localization of hLsg1 and mutants upon Leptomycin B treatment. HeLa cells were transiently transfected with hLsg1-YFP (a), YFP-hLsg1-1-600 (b), or YFP-hLsg1-480-658 (c), treated with 15 nM Leptomycin B, fixed at 0 h, 3 h and 5 h, and visualized by fluorescence microscopy.
To determine whether hLsg1 shuttled between nucleus and cytosol via a CRM1-dependent nuclear export pathway, we transfected Vero cells with either hLsg1-YFP or hLsg1 deletion mutants and compared the localization of the fusion proteins after treatment with the CRM-1 nuclear export inhibitor Leptomycin B (LMB) (Figure 5B). Full-length hLsg1 (YFP-hLsg1) is LMB-sensitive (Figure 5B); so is its C-terminal counterpart hLsg1-CFP (data not shown). To confirm this, we performed the same experiment using the deletion mutants YFP-hLsg1-1-600 and YFP-hLsg1-480-658 as well as the full length YFP-hLsg1. We also took intermediate time points (3 h and 5 h) to obtain insights into the kinetics of hLsg1 shuttling. Interestingly, YFP-hLsg1 accumulates in the nucleus over an 8 h period, and at 5 h most of the transfected cells showed punctate labeling in the nucleus reminiscent of Cajal bodies. YFP-hLsg1-480-658 showed a permanent nuclear location and YFP-hLsg1-1-600 was constantly in the cytosol.
These data suggest that hLsg1 shuttles between the cytosol and Cajal bodies via a CRM1-dependent export mechanism.
Discussion
Using database sequence similarity searches coupled with phylogenetic analysis, we were able to unite the circularly permuted GTPases into a family that we have named YRG for YlqF Related GTPases [see Additional files 1 and 2]. The YlqF protein family represents the largest subfamily of YRG expansion in eukarya, which is potentially involved in ribosome biogenesis.
Phylogenetic analysis defines ten GTPase subfamilies with a global phyletic distribution compatible with their presence in the last universal common ancestor (LUCA) of extant life forms [22]. An emerging concept suggests that these universal GTPases are necessary either for ribosome function or for transmitting information from the ribosome to downstream targets to generate specific cellular responses. These are associated with translation and include four translation factors, two OBG-like GTPases, the two signal-recognition-associated GTPases, the MRP subfamily of MinD-like ATPases and the YRG family. Here we have defined the YRG family for the first time as a eukaryotic expansion of the original Yawg/YlqF family [22] tightly coupled to the evolution of compartmentalization.
The YRG family was originally defined as a particular class of GTPases showing a circularly permuted structure, with the four GTPase motifs reorganized as G4 followed by G1, G2 and G3 (Figure 1A). This circular permutation is unique in the GTPase superfamily. However, we have shown that this inverted structure does not seem to affect GTPase activity or folding, in agreement with other studies [31,39]. Moreover, regarding the potential function of this family, it has been pointed out that most YRG members bind to the ribosome [YjeQ, [31]], are involved in the maturation of ribosomes or mitoribosomes [24,28,29,2], localize to compartments related to rRNA maturation [NGP, [1,39]], and are essential proteins (see Figure 3C and Additional file 2). Altogether, this indicates that YRG members have an essential role in ribosomal assembly.
Strikingly, we could find a member of the YRG family for every cellular compartment linked to ribosomes, including the chloroplast (Figure 1C), correlating with the expansion of the eukaryotic cell (Figure 1C). According to the phylogenetic tree of the family, the cytosolic form block (LSG1) is distinct from the nuclear form blocks (NOG2, NGP1, YawG), which later expanded into a nucleolar form (NUG1), in parallel with the incorporation of members upon engulfment of the future mitochondria (MTG1) that cluster within the YlqF branch as well as the future chloroplast (ChYlqF). Other events within the YlqF family included the appearance of a second cytosolic form upon speciation of the coelomates (GNL1), which may have had an equivalent in the plant lineage, since we observed a form of Lsg1 in A. thaliana (Figure 1B). Moreover, we observed the appearance of a second nucleolar form (Nucleostemin) upon speciation of the deuterostomes. Since Nucleostemin is involved in cell-cycle regulation in stem cells, we can hypothesize a direct mechanism of rRNA maturation in those highly specialized animal cells. We propose the following scenario for the evolution of the YRG family. First, a cytosolic founding member was duplicated upon the formation of a proto-nucleus, allowing the rRNA maturation pathway to be maintained (Figure 6). The second step included the engulfment of mitochondria and chloroplasts containing specific YRG forms involved in rRNA maturation in these compartments. The final step(s) involved the evolution of the cytosolic and nuclear members upon the specialization of the eukaryotic cell (nucleolus etc). This scenario accords with the work of Mans et al. [47], which showed by comparative genomics that a large set of proteins was involved in the formation and structure of the nuclear envelope and the pore complex: the nucleus evolved from a primordial prekaryote compartment and a primitive nuclear pore complex dependent on Ran and on Nug1p/Nug2p, a nucleolar YRG member.
Figure 6 The YRG family expands in relation to compartmentalization. Scenario of the evolution of compartmentalization of the YRG members based on the phylogenetic tree in Figure 1.
Interestingly, hLsg1 is the only member of this family that shows a dual localization (cytosol/endoplasmic reticulum and Cajal Bodies). The cytosol contains huge numbers of ribosomes freely diffusing or bound to the endoplasmic reticulum, and is the main transit pathway for rRNA en route to the mitochondria or the chloroplast. Cajal Bodies are spherical nuclear bodies containing a variety of components including nucleolar proteins, snRNPs and SMN. They are dynamic structures functionally linked to the nucleolus, presumably involved in RNP maturation and related to gene expression [43,44]. Consistent with these data, one could hypothesize that hLsg1 is a regulator of the rRNA pathway that can relocate to Cajal Bodies and interact with specific factors such as nucleolar proteins. The observation that Leptomycin B treatment leads to accumulation of hLsg1 in the nucleus clearly indicates shuttling via a CRM1-dependent export pathway. We hypothesized that hLsg1 relocalizes from the cytosol to the nucleus in response to internal (e.g. cell cycle) or external (e.g. growth factor) stimuli. In this way, hLsg1 would act on the control of rRNA biosynthesis at its source: the nucleolus. In the future, these hypotheses will be tested for hLsg1 and for the other YRG family members to elucidate their role in rRNA biosynthesis and maturation.
Conclusion
Using comparative genomics, we defined the YRG family as a unique group of circularly permuted GTPases. We suggest a potential function for this family, as well as a potential pathway by which the family members may act sequentially, following an evolutionary process linked to compartimentalization (Figure 6). A future goal will be to test this hypothesis experimentally and to dissect the molecular mechanisms of action of each member of the pathway.
Methods
Analysis of Lsg1p protein family
Database similarity searches
The translated sequence of the Homo sapiens gene FLJ11301 (GenBank accession no. NP_060855) was used to search the non-redundant protein database at the National Center for Biotechnology Information using the PSI-BLASTP program (15). Homologues were identified in Homo sapiens (GenBank accession identifier BAA92116), Mus musculus (XP_148574), Danio rerio (AAH66695), Caenorhabditis elegans (NP_490904), Caenorhabditis briggsae (CAE74467), Drosophila melanogaster (NP_569915), Anopheles gambiae (EAA13064), Saccharomyces cerevisiae (NP_011416), Schizosaccaromyces pombe (NP_593948), Arabidopsis thaliana (NP_172317), Zea mays (AAD41267), Encephalitozoon cuniculi (CAD26329), Eremothecium gossypii (NP_985506) and Plasmodium falciparum (NP_702181). The sequence corresponding to Rattus norvegicus had to be reconstructed using an insertion from Mus musculus, probably owing to an incorrect gene prediction (XP_213604).
Phylogenetic analysis
The 14 orthologous sequences were aligned using the ClustalW program [16]. PSI-BLAST searches on the NCBI protein database were performed using different representatives of the YRG family as seed, according to the bibliography, and were iterated until members of the closest subfamily were found in the list of hits. The sets of orthologous sequences were manually checked for sequence integrity and to clarify subfamily definitions. Progressively larger multiple sequence alignments were built by constructing multiple sequence alignments of each subfamily, which were manually polished and added together stepwise. At each step, the parts outside the central GTPase domain, which often showed no homology across subfamilies (and therefore should not be aligned), were trimmed to facilitate the production of the next multiple sequence alignment. The final multiple sequence alignment was used to produce the corresponding phylogenetic tree (excluding the non-aligned regions) using ClustalW. The full list of sequences used for the tree and their database identifiers are given as supplementary material [see Additional file 1].
Cell culture, transfections, immunostaining and fluorescence microscopy
HeLa (ATCC CCL-2) and Vero (ATCC CCL-81) cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FCS and penicillin/streptomycin at 37°C in an atmosphere of 5% CO2. Cells were seeded on to glass coverslips, Nunc plates or LabTek dishes and were transfected using Fugene6 (Roche) according to the manufacturer's protocols. For immunocytochemistry, transiently transfected HeLa cells were grown on coverslips and fixed in ice-cold methanol for 5 min at -20°C. The cells were then washed again and incubated in PBS for 20 min. Primary and secondary antibodies were diluted in PBS. The cells were incubated with primary antibodies followed by secondary antibodies for intervals of 30 min with three washing steps in between. The coverslips were then mounted in Mowiol on glass slides. Images of the stained cells were acquired using either a Zeiss Cell Observer System or a Leica AOBS confocal laser-scanning microscope.
GTP binding and GTPase activity measurements
Nucleotide binding was measured by the filtration method. Recombinant proteins were incubated in 20 mM Tris-HCl pH 7.5, 1 mM DTT, 5 mM MgCl2, 10 mM EDTA, 0.5 g/l bovine serum albumin, (3H)GTP or (3H)GDP (7,7 Ci/mmol, Amersham-Pharmacia-Biotech) and cold 30 μM GTP or GDP. After incubation at 30°C for the indicated times, samples were diluted in 500 μl of ice-cold washing buffer (20 mM Tris-HCl pH 7.5, 25 mM MgCl2 and 100 mM NaCl) and applied to a nitrocellulose filter (0.45 μm, Millipore). The filters were rinsed with 4 × 4ml ice-cold washing buffer and the radioactivity retained on the filters was determined by scintillation counting.
GTPase activity measurement by HPLC was described by Ahmadian et al. 1999 [17].
siRNAs transfection and western blotting
siRNA sequences were BLAST searched against the human genome to ensure that they were specific for hLsg1. The hLsg1 siRNA sequence showed no exact or near exact matches to any other sequence in the human genome and are therefore hLsg1-specific. siRNAs were synthesized by EUROGENTEC. hLsg1 siRNA (5'-UGGAGAGAAACUGCAAGACTT-3') targets nucleotides 506–524 of human hLsg1 relative to the first nucleotide of the start codon.
Cells were seeded into 12-well plates. Twenty-four hours later, they were transfected with 1.68 μg of siRNA per well (unless otherwise noted). Transfections were as described [28] with the following modifications. Additional OptiMEM (Invitrogen) was not added, and medium was removed before transfection and replaced with 400 μl of OptiMEM. Full-serum medium (unless otherwise noted) was added 4 h post-transfection. At the indicated times post-transfection, the cells were washed twice with PBS and detached from the plate with PBS EDTA. Whole cell extract was obtained by lysing the cells with RIPA buffer containing protease inhibitors and DTT. Protein concentrations were measured using the Bradford assay. Extracts were run on 8% polyacrylamide gels (12% for actin) and transferred to nitrocellulose membranes. The membranes were blocked overnight at 4°C in 1% non-fat dry milk (1 h at room temperature in 5% non-fat dry milk for actin), then probed with either rabbit polyclonal anti-hLsg1 or anti-actin (Santa Cruz Biotechnology) antibody for 1 h at room temperature (overnight at 4°C for actin), washed, and probed with a horseradish peroxidase-conjugated secondary antibody for 1 h at room temperature. Signals were detected using the ECL-Plus reagent (Amersham Biosciences).
Authors' contributions
Reynaud E.G. carried out the molecular biology experiments, in collaboration with Ly T.B.N, the localization studies, the immunofluorescence and the cell biology studies. Andrade M.A. carried the complete phylogenetic studies of the YRG family and drafted part of the manuscript. Bonneau F. and Scheffzek K participated in the characterization of the Lsg1 GTPase activity. Knop M. participated in the design of, and performed part of, the yeast experiments. Pepperkok R. conceived the study and participated in its design and coordination. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Complete Phylogenetic Tree of the YRG family. A multiple alignment of sequences of the YqeH, YjeQ, EngA, and YlqF families. The alignment was produced using ClustalW followed by manual editing [16]. The tree was generated from the alignment using MrBayes v3 [45] (The root of the tree is the one given by the MrBayes output.
Click here for file
Additional File 2
The YRG family members. The table compiles all the data collected from several datasets available for all the YRG members known so far.
Click here for file
Acknowledgements
We thank Angus Lammond for sharing an aliquot of anti-coilin antibodies. The yeast Lsg1 construct was a kind gift from Arlen Johnson. We appreciate the help of Jeremy Simpson in siRNA design and preparation of this manuscript, and we are grateful for funding from the European Molecular Biology Organization (Long Term Fellowship) to E. Reynaud. M.A. Andrade is the recipient of a Canada Research Chair.
==== Refs
Hedges SB Kumar S Vertebrate genomes compared Science 2002 297 1283 1285 12193771 10.1126/science.1076231
Huh WK Falvo JV Gerke LC Carroll AS Howson RW Weissman JS O'Shea EK Global analysis of protein localization in budding yeast Nature 2003 425 686 691 14562095 10.1038/nature02026
Simpson JC Wellenreuther R Poustka A Pepperkok R Wiemann S Systematic subcellular localization of novel proteins identified by large-scale cDNA sequencing EMBO Rep 2000 1 287 292 11256614 10.1093/embo-reports/kvd058
Giaever G Chu AM Ni L Connelly C Riles L Veronneau S Dow S Lucau-Danila A Anderson K Andre B Arkin AP Astromoff A El-Bakkoury M Bangham R Benito R Brachat S Campanaro S Curtiss M Davis K Deutschbauer A Entian KD Flaherty P Foury F Garfinkel DJ Gerstein M Gotte D Guldener U Hegemann JH Hempel S Herman Z Jaramillo DF Kelly DE Kelly SL Kotter P LaBonte D Lamb DC Lan N Liang H Liao H Liu L Luo C Lussier M Mao R Menard P Ooi SL Revuelta JL Roberts CJ Rose M Ross-Macdonald P Scherens B Schimmack G Shafer B Shoemaker DD Sookhai-Mahadeo S Storms RK Strathern JN Valle G Voet M Volckaert G Wang CY Ward TR Wilhelmy J Winzeler EA Yang Y Yen G Youngman E Yu K Bussey H Boeke JD Snyder M Philippsen P Davis RW Johnston M Functional profiling of the Saccharomyces cerevisiae genome Nature 2002 418 387 391 12140549 10.1038/nature00935
Maeda I Kohara Y Yamamoto M Sugimoto A Large-scale analysis of gene function in Caenorhabditis elegans by high-throughput RNAi Curr Biol 2001 11 171 176 11231151 10.1016/S0960-9822(01)00052-5
Ghaemmaghami S Huh WK Bower K Howson RW Belle A Dephoure N O'Shea EK Weissman JS Global analysis of protein expression in yeast Nature 2003 425 737 741 14562106 10.1038/nature02046
Tong AH Lesage G Bader GD Ding H Xu H Xin X Young J Berriz GF Brost RL Chang M Chen Y Cheng X Chua G Friesen H Goldberg DS Haynes J Humphries C He G Hussein S Ke L Krogan N Li Z Levinson JN Lu H Menard P Munyana C Parsons AB Ryan O Tonikian R Roberts T Sdicu AM Shapiro J Sheikh B Suter B Wong SL Zhang LV Zhu H Burd CG Munro S Sander C Rine J Greenblatt J Peter M Bretscher A Bell G Roth FP Brown GW Andrews B Bussey H Boone C Global mapping of the yeast genetic interaction network Science 2004 303 808 813 14764870 10.1126/science.1091317
Ito T Tashiro K Muta S Ozawa R Chiba T Nishizawa M Yamamoto K Kuhara S Sakaki Y Toward a protein-protein interaction map of the budding yeast: A comprehensive system to examine two-hybrid interactions in all possible combinations between the yeast proteins Proc Natl Acad Sci U S A 2000 97 1143 1147 10655498 10.1073/pnas.97.3.1143
Yu H Luscombe NM Lu HX Zhu XX Yu H Jing-Dong J Bertin N Chung S Vidal M Gerstein M Annotation Transfer Between Genomes: Protein-Protein Interologs and Protein-DNA Regulogs Genome Res 2004 14 1107 1118 15173116 10.1101/gr.1774904
Wilson CA Kreychman J Gerstein M Assessing annotation transfer for genomics: Quantifying the relations between protein sequence, structure and function through traditional and probabilistic scores J Mol Biol 2000 297 233 249 10704319 10.1006/jmbi.2000.3550
Hegyi H Gerstein M Annotation transfer for genomics: Measuring functional divergence in multi-domain proteins Genome Res 2001 11 1632 1640 11591640 10.1101/gr. 183801
Bourne HR Sanders DA McCormick F The GTPase superfamily: conserved structure and molecular mechanism Nature 1991 349 117 127 1898771 10.1038/349117a0
Sprang SR G protein mechanisms: insights from structural analysis Annu Rev Biochem 1997 66 639 678 9242920 10.1146/annurev.biochem.66.1.639
Kjeldgaard M Nyborg J Clark BF The GTP binding motif: variations on a theme FASEB J 1996 10 1347 1368 8903506
Altschul SF Madden TL Schaffer AA Zhang J Zhang Z Miller W Lipman DJ Gapped BLAST and PSI-BLAST: a new generation of protein database search programs Nucleic Acids Res 1997 25 3389 3402 9254694 10.1093/nar/25.17.3389
Higgins DG Thompson JD Gibson TJ Using CLUSTAL for multiple sequence alignments Methods Enzymol 1996 266 383 402 8743695
Ahmadian MR Zor T Vogt D Kabsch W Selinger Z Wittinghofer A Scheffzek K Guanosine triphosphatase stimulation of oncogenic Ras mutants Proc Natl Acad Sci U S A 1999 96 7065 7070 10359839 10.1073/pnas.96.12.7065
Saraste M Sibbald PR Wittinghofer A The P-loop – a common motif in ATP- and GTP-binding proteins Trends Biochem Sci 1990 15 430 434 2126155 10.1016/0968-0004(90)90281-F
Bateman A Coin L Durbin R Finn RD Hollich V Griffiths-Jones S Khanna A Marshall M Moxon S Sonnhammer EL Studholme DJ Yeats C Eddy SR The Pfam protein families database Nucleic Acids Res 2004 1 D138 141 14681378 10.1093/nar/gkh121
Letunic I Copley RR Schmidt S Ciccarelli FD Doerks T Schultz J Ponting CP Bork P SMART 4.0: towards genomic data integration Nucleic Acids Res 2004 32 D142 D144 14681379 10.1093/nar/gkh088
Daigle DM Rossi L Berghuis AM Aravind L Koonin EV Brown ED YjeQ, an essential, conserved, uncharacterized protein from Escherichia coli, is an unusual GTPase with circularly permuted G-motifs and marked burst kinetics Biochemistry 2002 41 11109 11117 12220175 10.1021/bi020355q
Leipe DD Koonin EV Aravind L Evolution and classification of P-loop kinases and related proteins J Mol Biol 2003 333 781 815 14568537 10.1016/j.jmb.2003.08.040
Vernet C Ribouchon MT Chimini G Pontarotti P Structure and evolution of a member of a new subfamily of GTP-binding proteins mapping to the human MHC class I region Mamm Genome 1994 5 100 105 8180467 10.1007/BF00292335
Kallstrom G Hedges J Johnson A The putative GTPases Nog1p and Lsg1p are required for 60S ribosomal subunit biogenesis and are localized to the nucleus and cytoplasm, respectively Mol Cell Biol 2003 23 4344 4355 12773575 10.1128/MCB.23.12.4344-4355.2003
Koonin EV Fedorova ND Jackson JD Jacobs AR Krylov DM Makarova KS Mazumder R Mekhedov SL Nikolskaya AN Rao BS Rogozin IB Smirnov S Sorokin AV Sverdlov AV Vasudevan S Wolf YI Yin JJ Natale DA A comprehensive evolutionary classification of proteins encoded in complete eukaryotic genomes Genome Biol 2004 5 R7 14759257 10.1186/gb-2004-5-2-r7
Tatusov RL Fedorova ND Jackson JD Jacobs AR Kiryutin B Koonin EV Krylov DM Mazumder R Mekhedov SL Nikolskaya AN Rao BS Smirnov S Sverdlov AV Vasudevan S Wolf YI Yin JJ Natale DA The COG database: an updated version includes eukaryotes BMC Bioinformatics 2003 4 41 12969510 10.1186/1471-2105-4-41
Tsai RY McKay RD A nucleolar mechanism controlling cell proliferation in stem cells and cancer cells Genes Dev 2002 16 2991 3003 12464630 10.1101/gad.55671
Saveanu C Namane A Gleizes PE Lebreton A Rousselle JC Noaillac-Depeyre J Gas N Jacquier A Fromont-Racine M Sequential protein association with nascent 60S ribosomal particles Mol Cell Biol 2003 13 4449 4460 12808088 10.1128/MCB.23.13.4449-4460.2003
Bassler J Grandi P Gadal O Lessmann T Petfalski E Tollervey D Lechner J Hurt E Identification of a 60S preribosomal particle that is closely linked to nuclear export Mol Cell 2001 8 517 529 11583615 10.1016/S1097-2765(01)00342-2
Barrientos A Korr D Barwell KJ Sjulsen C Gajewski CD Manfredi G Ackerman S Tzagoloff A MTG1 codes for a conserved protein required for mitochondrial translation Mol Biol Cell 2003 14 2292 2302 12808030 10.1091/mbc.E02-10-0636
Daigle DM Brown ED Studies of the interaction of Escherichia coli YjeQ with the ribosome in vitro J Bacteriol 2004 186 1381 1387 14973029 10.1128/JB.186.5.1381-1387.2004
Emanuelsson O Nielsen H von Heijne G ChloroP, a neural network-based method for predicting chloroplast transit peptides and their cleavage sites Protein Science 1999 8 978 984 10338008
Bourne HR Sanders DA McCormick F The GTPase superfamily: a conserved switch for diverse cell functions Nature 1990 348 125 132 2122258 10.1038/348125a0
Morimoto T Loh PC Hirai T Asai K Kobayashi K Moriya S Ogasawara N Six GTP-binding proteins of the Era/Obg family are essential for cell growth in Bacillus subtilis Microbiology 2002 148 3539 3552 12427945
Kuge O Dascher C Orci L Rowe T Amherdt M Plutner H Ravazzola M Tanigawa G Rothman JE Balch WE Sar1 promotes vesicle budding from the endoplasmic reticulum but not Golgi compartments J Cell Biol 1994 125 51 65 8138575 10.1083/jcb.125.1.51
Barlowe C d'Enfert C Schekman R Purification and characterization of SAR1p, a small GTP-binding protein required for transport vesicle formation from the endoplasmic reticulum J Biol Chem 1993 268 873 879 8419365
Cherfils J Chardin P GEFs: structural basis for their activation of small GTP-binding proteins Trends Biochem Sci 1999 24 306 311 10431174 10.1016/S0968-0004(99)01429-2
Arigoni F Talabot F Peitsch M Edgerton MD Meldrum E Allet E Fish R Jamotte T Curchod ML Loferer H A genome-based approach for the identification of essential bacterial genes Nat Biotechnol 1998 16 851 856 9743119 10.1038/nbt0998-851
Levdikov VM Blagova EV Brannigan JA Cladiere L Antson AA Isupov MN Seror SJ Wilkinson AJ The crystal structure of YloQ, a circularly permuted GTPase essential for Bacillus subtilis viability J Mol Biol 2004 340 767 782 15223319 10.1016/j.jmb.2004.05.029
Racevskis J Dill A Stockert R Fineberg SA Cloning of a novel nucleolar guanosine 5'-triphosphate binding protein autoantigen from a breast tumor Cell Growth Differ 1996 7 271 280 8822211
Bernards A GAPs galore! A survey of putative Ras superfamily GTPase activating proteins in man and Drosophila Biochim Biophys Acta 2003 1603 47 82 12618308
Schlemiel M Staeheli P The interferon-induced 67-kDa guanylate-binding protein (hGBP1) is a GTPase that converts GTP to GMP J Biol Chem 1994 269 11299 11305 7512561
Gall JG Cajal bodies: the first 100 years Annu Rev Cell Dev Biol 2000 16 273 300 11031238 10.1146/annurev.cellbio.16.1.273
Ogg SC Lamond AI Cajal bodies and coilin-moving towards function J Cell Biol 2002 159 17 21 12379800 10.1083/jcb.200206111
Ronquist F Huelsenbeck JP MrBayes 3 Bayesian phylogenetic inference under mixed model Bioinformatics 2003 19 1572 1574 12912839 10.1093/bioinformatics/btg180
Politz JC Yarovoi S Kilroy SM Gowda K Zwieb C Pederson T Signal recognition particle components in the nucleolus Proc Natl Acad Sci U S A 2000 97 55 60 10618370 10.1073/pnas.97.1.55
Mans BJ Anantharaman V Aravind L Koonin EV Comparative genomics, evolution and origins of the nuclear envelope and nuclear pore complex Cell Cycle 2004 3 1612 1637 15611647
|
16209721
|
PMC1262696
|
CC BY
|
2021-01-04 16:39:16
|
no
|
BMC Biol. 2005 Oct 7; 3:21
|
utf-8
|
BMC Biol
| 2,005 |
10.1186/1741-7007-3-21
|
oa_comm
|
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-1231618802810.1186/1471-2407-5-123Research ArticlePhenotypic and functional analysis of lymphocytes infiltrating osteolytic tumors: use as a possible therapeutic approach of osteosarcoma Théoleyre S [email protected] K [email protected] B [email protected] N [email protected] F [email protected]édini F [email protected] D [email protected] INSERM ERI 7 – EA 3822; Physiopathologie de la Résorption Osseuse et Thérapie des Tumeurs Osseuses Primitives. Faculté de Médecine, 1 rue Gaston Veil, 44035 Nantes cedex 1, France2005 27 9 2005 5 123 123 6 6 2005 27 9 2005 Copyright © 2005 Théoleyre et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Osteosarcoma is the most common type of primary bone tumor. The use of aggressive chemotherapy has drastically improved the prognosis of the patients with non-metastatic osteosarcomas, however the prognosis of the patients with metastasis is still very poor. Then, new and more effective treatments for curing osteosarcoma, such as immunotherapy are needed. Tumor-infiltrating lymphocytes (TIL) have been involved in the control of tumor development and already assessed with success for the treatment of several cancers including melanoma. While TIL represent a fascinating therapeutic approach in numerous malignant pathologies, there is few report concerning adult bone-associated tumors including osteosarcoma.
Methods
Human TIL were isolated and characterized (phenotype, lytic activity) from twenty-seven patients with bone-associated tumors (osteosarcoma, Ewing's sarcoma, giant cell tumor, chondrosarcoma, plasmocytoma and bone metastases). Similar experiments were performed using rat osteosarcoma model.
Results
While TIL with a main CD4+ profile were easily isolated from most of the tumor samples, only TIL extracted from osteosarcoma were cytotoxic against allogeneic tumor cells. In all cases, TIL lytic activity was significantly higher compared to autologous peripheral blood leukocytes. Similar data were observed in rat osteosarcoma model where TIL were characterized by a main CD4+ profile and high lytic activity against allogeneic and autologous tumor cells. Moreover, rat TIL expansion was not accompanied by refractoriness to further activation stimulus mainly by tumor antigens.
Conclusion
These results demonstrated that TIL therapy could be a very efficient strategy for the treatment of adult osteosarcoma.
==== Body
Background
Primary or secondary malignant bone tumors represent a major therapeutic challenge in medical oncology. Despite efficiency of conventional treatments by chemo- and radiotherapy, long-term outcome of the patients suffering from malignant bone tumors remains poor. Among them, osteosarcoma is the most frequent primary bone tumor. Indeed, the current strategy for the treatment of high-grade osteosarcoma is based on neo-adjuvant chemotherapy, delayed en-bloc wide resection and adjuvant chemotherapy adapted to the histologic profile of the tumor tissues removed during surgery [1]. While marked improvements in surgery and the development of different regimens of multidrug chemotherapy over the past 25 years, the survival remains around 55 to 70% after 5 years [2,3]. Furthermore, the prognosis is worse in the patients with non-extremities localization, advancing age, radio-induced osteosarcoma and those arising from Paget's disease of bone, representing 40% of the entire osteosarcoma population. In addition, the patients with metastatic osteosarcoma at the time of diagnosis have poor survival statistics (30% at 5 years). All these findings suggest necessity of establishing new therapeutic strategies to improve the overall rate of survival, especially in high-risk sub-groups.
Among the potential new therapeutic strategies, immunotherapies are based on the up-regulation of the immune response in tumor-bearing host. Two therapeutic approaches can be distinguished: (i) active immunotherapies that elicit immune response against tumor cells in tumor-bearing host (including pulsed dendritic cells and cytokine treatments), (ii) passive or adoptive immunotherapies consist on the administration of ex vivo-expanded tumor-specific cytotoxic immune cells represented by T lymphocytes. The identification of tumor-specific lymphocytes has resulted in new therapeutic strategies based on mounting a sustained and effective anti-tumor immune response [4,5]. It is theorized that the infiltrating lymphoid represents a selected population of cells which have preferentially migrated to the tumor secondary to an immune response. These T lymphocytes termed tumor-infiltrating lymphocytes (TIL) are considered to be more specific in their immunological reactivity to tumor cells than the non-infiltrating lymphocytes [6]. However, if the role of TIL has not been clearly defined, TIL could recognized as a reactional mechanism against tumor development. Moreover, their reduced response in the tumor tissue may be due to the suppressive influence under the tumor microenvironment. In this context, TIL have been identified in numerous neoplasia, such as melanoma [5-9], various carcinomas [10-17], myeloma [18], pediatric tumors [19] and sarcomas [20-22] and the therapeutic relevance of these TIL has been also documented in both animal models and clinical trials [23-28]. Unfortunatly, up to date, very few data is available on the phenotypic and functional characterization of TIL isolated from adult bone-associated tumors.
Because, in vitro studies on bone tumor infiltrating T lymphocytes are lacking, the goal of the present study is to characterize the TIL derived from 27 adult human bone-associated tumors as well as autologous peripheral blood leukocytes (PBL) at the phenotypic and functional levels. The therapeutic potential of TIL was then discussed for osteosarcoma.
Methods
1) Patients
Twenty-seven bone-associated tumors [6 osteosarcomas, 2 Ewing's sarcomas, 2 chondrosarcomas, 7 giant cell tumors (GCTs), 2 plasmocytomas, 4 bone metastases (2 from kidney origin, 2 form unknown origin) and 4 other pathologies (1 chondromyxoid fibroma, 1 fibrous dysplasia, 1 chordoma, 1 undifferentiated sarcoma)] from 27 patients [12 women (38 + 17.8 years, range: 17–75) and 15 men (47.9 + 19.2) years, range: 16–75)] were included in the present study. All patients were treated in the Department of Orthopedic Surgery of Nantes University Hospital (France) between September 2004 and June 2005. Patients' characteristics were summarized in Table 1 ("Additional file 1"). Written informed consent was obtained before each patient was included in the study.
2) Cell lines
The rat osteosarcoma (UMR106), human osteosarcoma (MG63, SaOS2), B lymphoma (Daudi) and leukemia (K562) cell lines were purchased from the American Tissue Cell Collection (USA). The rat osteosarcoma ROS 17/2.8 cell line was a kind gift from Pr HJ Donahue (Penn State University, USA). The tumor cell lines were culturedin RPMI 1640 (Invitrogen, France) supplemented with 10% foetal calf serum (FCS) (Hyclone, Perbio, France). The primary osteosarcoma was isolated from small tumor fragment of rat osteosarcoma model described above. Briefly, small tumor fragments were digested by 1 mg/mL collagenase A (Boehringer Manheim) for 2 hours. The cell suspensions obtained were then washed three times with PBS and cultured in RPMI 1640 supplemented with 10% FCS. The medium was changed twice a week and the adherent osteosarcoma cells were harvested using a 0.05% trypsin- 0.02% EDTA solution and subjected to the study.
3) Osteosarcoma rat model
The osteosarcoma was initially induced by a local injection of colloidal radioactive 144Cerium in rats [29]. The tumor can be re-grafted and maintained in vivo for many months; otherwise fragments were frozen until re-utilization. This rat osteosarcoma model used in the present study mimics the human pathology in terms of tumor growth, bone involvements and lung metastases [30,31].
4) Peripheral blood leukocytes (PBL) preparation and tumor-infiltrating lymphocytes (TIL) extraction
Both human and rat PBL were isolated from heparinized peripheral blood after Ficoll-Hypaque (Eurobio, France) gradient centrifugation. The human and rat PBL were respectively cultured in RPMI 1640 supplemented with 10% FCS and 300 UI/ml IL-2 (Eurocetus, France) for 21 days and X-VIVO 15 supplemented with 900 UI/ml IL-2 for 14 days. Their phenotype and cytotoxic activity were then analyzed according to the procedure described below.
Human tumor specimens were directly received from the operating room under sterile condition, minced with scissors into 1 mm3 fragments and cultured immediately in RPMI 1640 (Invitrogen) supplemented with 10% FCS (Hyclone), 300 UI/ml IL-2 (Eurocetus) and antibiotics mixture (100 UI/mL penicillin and 100 μg/mL streptomycin). Half of medium was changed twice a week and TIL were cultured for 21 days.
Rat tumor specimens were obtained from the model described above, minced with scissors into 9 mm3 fragments and subjected enzymatic digestion using 2 mg/mL collagenase and 50 UI/mL hyaluronidase (Sigma-Aldrich, France) for 2 hours. The cell suspension (106 cell/mL) was filtrated using a 70 μm cell strainer. Then, enriched by Ficoll-Hypaque gradient centrifugation and cultured in X-VIVO 15 serum free-medium supplemented with 900 UI/mL IL-2 and antibiotics mixture (100 UI/mL penicillin and 100 μg/mL streptomycin). Half of medium was changed twice a week and TIL were cultured for 14 days. Both PBL and TIL were maintained at 37°C in a 5% CO2 humidified atmosphere.
5) Cytotoxicity assay
Target cells (UMR106, MG63, SaOS2, Daudi, K562, ROS17/2.8 and primary autologous rat osteosarcoma cells) were labeled with 10 μCi of Na2 51CrO4 (Amersham, UK) for 1 hour at 37°C. Then, target cells were washed three times with fresh culture medium. Lymphocytes (TIL or PBL) were added at a different target/effector ratio (T/E) of 1/12.5, 1/25 and 1/50. The plate was then incubated for 4 hours at 37°C, and 100 μL of supernatant were haversted for counting in a γ-counter. Maximal and spontenous 51Cr-releases were determined by incubating the target cells with 100 μL of detergent (Triton 1% in distillated water) or 100 μL of culture medium, respectively. Spontaneous release (SR) never exceeded 10% of maximal release (MR). The percentage of specific chromium release was calculated according to the following formula: [(experimental release-SR)/(MR-SR)] × 100%].
6) Flow cytometric analysis
For phenotyping, lymphocytes (TIL or PBL) were stained with conjugated monoclonal antibodies against CD3 (PE), CD4 (APC), CD8 (perCP), CD20 (FITC), CD45RA (FITC), CD161 (PE). Monoclonal antibodies were purchased from Becton Dickinson (France) (CD3, CD4, CD8 and CD20) and from Serotec (France) (CD161 and CD45RA). Cell suspensions were then analyzed on a Becton Dickinson fluorescence-activated cell sorter. Ten thousand scatter-gated cells were analyzed in each sample and fluorescence profiles of the cells were determined with logarithmic signal amplifiers.
7) Statistical analysis
All experiments were performed in triplicate. The mean and SD was calculated for all conditions and compared by an ANOVA test. Differences relative to a probability of two-tailed P < 0.05 were considered significant
Results
1) Human TIL can be extracted from various human tumor biopsies and cultured
TIL were able to isolate successfully from small tumor fragments (wet weight < 5 g) immediately after the biopsy except for chondrosarcoma, fibrous dysplasia and 50% of bone metastases (Table 1) and maintained for 3 weeks in the presence of 300 UI/ml IL-2 for flow cytometric analyses. Most of cultured TIL were mainly composed of CD3+ (range: 65%–99%) (Figure 1, hatched bars). TIL from plasmocytomas showed preferentially CD4+ profile (median: 64.4%) with minor CD8+ profile (median: 6.4%). Similarly, TIL from osteosarcomas were predominantly CD4+ (median: 49.9%) with second highest CD8+ sub-population (median: 15.5%). TIL isolated from Ewing's sarcomas, GCTs and bone metastases are characterized by almost same amount of CD4+ and CD8+ populations (around 50%) (Figure 1). A minor amount of CD20+ and CD161+ cells were observed in all TIL without correlation with tumor histology.
To compare PBL and TIL expanded from the same patients, phenotypic analyses of PBL were also performed in all included cases (Figure 1, dark bars). The amount of CD8+ TIL was lower in plasmocytoma, Ewing's sarcoma and osteosarcoma TIL compared to that of PBL (respectively 88%, 58%, and 57%). A high amount of natural killer (NK) CD161+ (43%) cells were detected in PBL from Ewing's sarcoma-bearing patients in contrast to the others.
2) Human TIL isolated from osteosarcomas exert cytotoxic activity against allogeneic cell lines
To determine the cytotoxic activity of expanded TIL against allogeneic cell lines, 51Cr-released assays were carried out. The lytic patterns of TIL against Daudi and K562 cell lines that reflect LAK and NK activity, and against two allogeneic osteosarcoma cell lines SaOS2 and MG63, which exhibit respectively high and low osteoblastic differentiation state were assessed. TIL amplified from plastocytomas, Ewing's sarcomas and GCTs did not show any lytic activity against all allogeneic cell lines tested (data not shown). However, TIL extracted from osteosarcoma biopsies were significantly cytotoxic against all cell lines analyzed compared to matched autologous PBL (p < 0.01) (Figure 2). The cytotoxic profiles against K562 and the other cell lines were different along with T/E ratio. Indeed, the magnitude of TIL cytotoxicity against K562 cells varied from 50% to 70% respectively for a ratio T/E of 1/12.5 to 1/50. The magnitude of TIL lytic activity against Daudi, SaOS2 and MG63 reached 20% for a ratio T/E of 1/12.5 to 65% for a ratio T/E of 1/50. In all osteosarcomas analyzed, autologous PBL were weakly cytotoxic against the four allogeneic cell lines, and no dose-response was observed in autologous PBL cytotoxic activity (Figure 2).
3) TIL isolated from rat osteosarcoma showed a similar phenotype and lytic pattern to human TIL
TIL expanded from rat tumor biopsies revealed same phenotype as those of human osteosarcoma. Namely, rat TIL were 90% CD3+, 50% CD4+ and 18% CD8+ in median value (Figure 3, hatched bars). Autologous PBL had an inverse CD4+/CD8+ ratio with 16% CD4+ and 32% CD8+ cells in median value (Figure 3, dark bars).
51Cr-released assays also disclosed the similar lytic pattern of rat TIL against autologous and allogeneic tumor cell lines (UMR 106, ROS 17/2.8) to human TIL (Figure 4). Moreover, this lytic activity was higher against autologous tumor cells and UMR 106 (63% T/E of 1/50) compared to ROS 17/2.8 cells (48% T/E of 1/50) (Figure 4).
To elucidate the capability of autologous tumor cells to stimulate the proliferation of rat TIL, expanded TIL were cultured with irradiated tumor cells. Figure 5 showed the kinetic expansion of rat TIL without tumor antigens during the first 23 days and with irradiated tumor cells in the following term. TIL proliferation was characterized by gradual expansion (more than 4-fold) in the first 3 weeks culture and rapid proliferation reached 10-fold number within following five days culture (Figure 5). TIL expanded from rat osteosarcoma were very sensitive to the tumor antigens expressed by autologous tumor cells. These data demonstrated that TIL expanded from osteosarcoma were very sensitive to the tumor antigens expressed by autologous tumor cells.
Discussion
The present data demonstrated that TIL could be isolated from the most of adult bone-associated tumors studied (osteosarcomas, Ewing's sarcomas, GCTs, plasmocytomas) except for chondrosarcoma, fibrous dysplasia and 50% of bone metastases; while only TIL extracted from osteosarcomas demonstrated a lytic activity against allogeneic tumor cell lines. Moreover, TIL extracted from osteosarcoma biopsies were significantly cytotoxic against all cell lines analyzed compared to matched autologous PBL. These findings demonstrating the interest of expanded TIL to target osteosarcoma cells.
Lymphocytes localized in a tumor foci could represent a tumor-specific response of T lymphocytes associated for one part to a non-specific component due to the inflammatory infiltrate. In this context, the adoptive transfer of specific-cytotoxic T cells has been already shown to induce tumor rejection in several animal tumor models [32,33]. Although isolation and characterization of TIL have been successfully performed in human malignant tumors [5-21], few data is available in the literature about bone-associated tumors except for pediatric tumors [34,35]. The presence of T lymphocytes in human osteosarcoma tissues was previously studied by Trieb et al. by immunohistochemical studies [36]. Phenotypic analyses have revealed that these infiltrating lymphocytes into osteosarcoma were 95% CD3+ and 68% CD8+ [36]. Rivoltini L et al. have also analyzed the phenotype of TIL in 37 pediatric tumors, including 12 osteosarcomas and demonstrated CD8+ predominancy [34]. The latter also revealed that the lytic pattern of TIL against both allogeneic and autologous tumor cells were variable depend on the histology. Rivoltini L et al. concluded, in 1992, that TIL extracted from pediatric tumors were difficult to expand at levels required for immunotherapy, however they could be envisaged [34]. In our study, the phenotype of TIL isolated from adult osteosarcomas are very similar in all cases (two third of CD4+ and one third of CD8+) and autologous PBL were weakly cytotoxic against the four allogeneic cell lines tested. This discrepancy between TIL and PBL could be explained by the homing processus of the tumor-specific T lymphocytes to the tumor foci. Indeed, recently, Haanen et al investigated the presence of TIL and circulating Tumor Associated antigens(TAA)-T cells before and after autologous tumor cell vaccination in metastatic melanoma patients [35]. They demonstrated that in none of the patients, TAA-specific T cells were found both in tumor tissue and circulating blood at the same time. No significant changes in the frequency and specificity circulating TAA-specific T were found during the treatment period in all patients while inside melanoma tissue, TAA-specific TIL could be detected in 75% of tumor samples analyzed. These data suggested that a possible homing phenomenon of the tumor-specific T cell population to the tumor site could contribute to the effectiveness of antitumor immunity.
An experiment rat osteosarcoma model mimicking the human pathology was used to elucidate the availability of TIL in an autologous system. TIL extracted from rat tumor biopsies demonstrated similar phenotype to human osteosarcoma TIL (predominantly CD4+) and exhibit a highly cytotoxic activity against autologous tumor cells. The predominance of CD4+ TIL obtained from human and rat biopsies contrasts partially to the data reported by Topalian et al. [37]. Indeed, these authors showed a correlation between CD8+ lymphocytes and high cytotoxic activity. However, the role of the CD4+ T lymphocyte sub-populations on the CD8+ TIL activity remains to be clearly defined. The CD8+ T cell survival and activity against the most of the tumor is increased by CD4+ T helper cells (CD4+CD25-) and downregulated by CD4+ T regulator cells (CD4+CD25+) [38-41]. The CD4+CD25+ T cells maintain an microenvironment in the tumor sites that conceal the immunogenicity of tumor to permit progressive growth of antigenic tumors. In such cases, the suppression (in vivo or ex vivo from TIL extracted) of the T regulator cells represents a possible design of immunotherapeutic approach. In contrast to these data, the intestinal tumors which are strongly associated with an inflammatory microenvironment regress after injection of CD4+CD25+ T regulator cells [42]. The overall published papers raise the possibility of broader functions for regulatory lymphocytes in prevention and treatment of human cancers. Thus, CD4+ predominance in TIL population could not be prejudicial to the success of T lymphocyte-based immunotherapy but could be closely related to their CD25+ (low or high), CD25- status.
The lytic activity exhibited by rat osteosarcoma TIL is lower against ROS 17/2.8 than against UMR 106 and autologous tumor cells, suggesting some common tumor antigens shared by tumor cells. In contrast to the results of Rivoltini et al. [34], the low proliferation rate of rat TIL could be easily re-stimulated by autologous tumor antigens (Figure 5). These data therefore revealed that stimulation by tumor antigens could be a more appropriate technique to expand TIL for their use immunotherapy.
In the past decade, several reports underlined the capability of cellular therapy for osteosarcoma. The identification of human tumor antigens has opened new areas of antigen-specific cancer immunotherapy specifically targeting these antigens [6]. Indeed, several tumor antigens such as melanoma-associated antigen (MAGE) [43], squamous cell carcinoma antigen recognized by T cells (SART) 1 [44], SART3 [45] and papillomavirus binding factor [46] are expressed in osteosarcoma and provided the rationale to develop cellular therapies for this pathology. SART3 is highly expressed in osteosarcoma [45]. Moreover, the SART3-derived peptides were able to induce HLA-A2-restricted and tumor-specific cytotoxic T lymphocytes [45]. These data strengthened the potential use of the SART3-derived peptides for specific immunotherapy of HLA-A2+ patients suffering from osteosarcoma. Taken together with the prevalence of HLA-A24 [46], this strategy could be applicable for approximately 60% of patients with osteosarcoma. Furthermore, no severe adverse response associated with peptides administration, and a significant up-modulation of the cellular immune response against tumor cells in clinical trial using SART3-derived peptides in HLA-A24+ patients with colon cancer [47] encourages further application of this strategy for osteosarcoma. More recently, Trieb et al. demonstrated that Hsp 72 is involved in the interaction between T lymphocytes and osteosarcoma cells expressing Hsp72 [48]. The cytotoxic potential of these lymphocytes might lead to an increased immune response leading to the rejection of the osteosarcoma [48]. In light of the complementary animal experiments, cytotoxic lymphocytes recognizing specific tumor antigens appear to be a more effective therapeutic strategy than polyclonal TIL-based immunotherapy. Combined TIL with above strategies therefore might be more effective to achieve desired results.
Conclusion
The results of the present report reveal that TIL exhibiting a high cytotoxic activity can be easily isolated from adult osteosarcomas. The poor prognosis of osteosarcoma must lead to develop new therapeutic approaches, especially in high sub-group such as adult osteosarcomas. TIL therapy is one of the hopeful strategies contribute to answer this intention.
Abbreviations
Tumor-infiltrating lymphocytes (TIL)
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
TS and M K contributed equally to this work. TS, MK and CB and carried out the in vitro experiments (TIL extraction, phenotypic and functional analyses); PN and GF coordinated the patient healthcare included in the study and collected human samples. FR carried out and coordinated the animal studies. HD conceived of the study, participated in its design and coordination, and has written the manuscript. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional File 1
Table 1: Panel of patients studied. 27 patients treated in the Department of Orthopedic Surgery of Nantes University Hospital (France) between September 2004 and June 2005, were included in the present study. The Table summarize the patients' characteristics included.
Click here for file
Acknowledgements
This work was supported by a grant from the French Ministry of Research and Technology (ACI n°TS/0220044) and by a grant from the Pays-de-Loire Committee of the Ligue Contre le Cancer. S. Théoleyre and K. Mori, respectively, received a fellowship from the Pays-de-Loire Committee of the Ligue Contre le Cancer and from the Ligue Nationale Contre le Cancer.
Figures and Tables
Figure 1 Phenotypic analysis of TIL isolated from human bone associated tumors compared with the phenotype of autologous PBL. TIL were isolated from tumor biopsies and cultured for 21 days in RPMI 1640 supplemented with 10% FCS and 300 UI/ml IL-2. Autologous PBL were isolated from peripheral blood samples after gradient density centrifugation and cultured in the same conditions for 21 days. After the indicated culture period, the phenotype of TIL (hatched bars) and PBL (dark bars) were analyzed by flow cytometry. The percentages of positive cells are given as median value. Plasmocytoma, n = 2; Ewing's sarcoma, n = 2; giant cell tumor, n = 7; bone metastase, n = 4; osteosarcoma, n= 6.
Figure 2 Comparison of cytotoxic activity between TIL isolated from human osteosarcomas and autologous PBL. TIL and autologous PBL were isolated from human osteosarcomas (n = 6). After 21 days of culture in RPMI 1640 supplemented with 10% FCS and 300 UI/ml IL-2, TIL and autologous PBL cytotoxic activities were determined using 51Cr-released assays. The cytotoxicity was measured against allogeneic cell lines including Daudi, K562 cell lines and two osteosarcoma cell lines, high (SaOS2) and low (MG63) osteoblastic differentiated osteosarcoma. ** p < 0.01; *** p < 0.001, TIL versus PBL cytotoxicity.
Figure 3 Comparative analysis of TIL and autologous PBL phenotype isolated from rat osteosarcoma. TIL were isolated from rat osteosarcoma biopsies (n = 8) by enzymatic treatment and cultured for 14 days in X-VIVO 15 supplemented with 900 UI/ml IL-2. Autologous PBL were isolated from peripheral blood samples after gradient density centrifugation and cultured in the same conditions for 14 days. After the culture period, the phenotype of TIL (hatched bars) and PBL (dark bars) were analyzed by flow cytometry. The percentages of positive cells are given as median value.
Figure 4 Comparison of cytotoxic activity between TIL isolated from rat osteosarcomas and autologous PBL. TIL were isolated from rat osteosarcoma biopsies (n = 8). After 14 days of culture in X-VIVO 15 supplemented with 900 UI/ml IL-2, TIL and autologous PBL cytotoxic activities were determined using 51Cr-released assays. The cytotoxicity was measured against autologous rat osteosarcoma cells and allogeneic rat osteosarcoma cell lines including UMR 106 and ROS 17/2.8. * p < 0.05, TIL versus PBL cytotoxicity.
Figure 5 Proliferation profile of TIL isolated from rat osteosarcoma. These data are representative of three independent tumors. TIL isolated from rat osteosarcoma biopsies by enzymatic treatment were cultured at 5 × 105 TIL per well (6-multiwell plate) for 28 days in X-VIVO 15 supplemented with 900 UI/ml IL-2. At day 23 (arrow), 20 × 103 irradiated (75 grays) autologous osteosarcoma cells were added in each well. TIL were manually numbered using the trypan blue staining.
==== Refs
Rosen G Murphy ML Huvos AG Gutierrez M Marcove RC Chemotherapy, en bloc resection, and prosthetic bone replacement in treatment of osteogenic sarcoma Cancer 1976 37 1 11 1082364
Link MP Goorin AM Horowitz M Meyer WH Belasco J Baker A Ayala A Shuster J Adjuvant chemotherapy of high-grade osteosarcoma of the extremity. Updated results of the multi-institutional osteosarcoma study Clin Orthop 1991 270 8 14 1884563
Provisor AJ Ettinger LJ Nachman JB Krailo MD Makley JT Yunis EJ Huvos AG Betcher DL Baum ES Kisker CT Miser JS Treatment of non-metastatic osteosarcoma of the extremity with pre-operative chemotherapy: a report from the Children's Cancer Group J Clin Oncol 1997 15 76 84 8996127
Rosenberg SA Spiess P Lafreniere R A new approach to the adoptive immunotherapy of cancer with tumor-infiltrating lymphocytes Science 1986 233 1318 1321 3489291
Aebersold P Hyatt C Johnson S Hines K Korcak L Sanders M Lotze M Topalian S Yang J Rosenberg SA Lysis of autologous melanoma cells by tumor-infiltrating lymphocytes: association with clinical response J Natl Cancer Inst 1991 83 932 937 2067036
Rosenberg SA Progress in human tumour immunology and immunotherapy Nature 2001 411 380 384 11357146 10.1038/35077246
Dudley ME Wunderlich JR Shelton TE Even J Rosenberg SA Generation of tumor-infiltrating lymphocyte cultures for use in adoptive transfer therapy for melanoma patients J Immunother 2003 26 332 342 12843795 10.1097/00002371-200307000-00005
Fabbri M Ridolfi R Maltino R Ridolfi L Riccobon A Flamini E De Paola F Verdecchia GM Amadori D Tumor infiltrating lymphocytes and continous infusion interleukin-2 after metastasectomy in 61 patients with melanoma, colorectal and renal carcinoma Tumori 2000 86 46 52 10778766
Huang J Khong HT Dudley ME El-Gamil M Li YF Rosenberg SA Robbins PF Survival, persistence, and progressive differentiation of adoptively transferred tumor-reactive T cells associated with tumor regression J Immunother 2005 28 258 267 15838383 10.1097/01.cji.0000158855.92792.7a
El Hage F Vergnon I Grunenwald D Soria JC Chouaib S Mami-Chouaib F Generation of diverse mutated tumor antigen-specific cytotoxic T lymphocytes in a lung cancer patient with long survival Oncol Rep 2005 14 763 769 16077989
Dillma R Schiltz P DePriest C Barth N Beutel L de Leon C O'Connor A Nayak S Tumor-infiltrating lymphocytes and interleukin-2: dose and schedules of administration in the treatment of metastatic cancer Cancer Biother Radiopharm 2004 19 730 737 15665620 10.1089/cbr.2004.19.730
Kowalczyk D Skorupski W Kwias Z Nowak J Flow cytometric analysis of tumour-infiltrating lymphocytes in patients with renal cell carcinoma Br J Urol 1997 80 543 547 9352689
Dierderichsen AC Zeuthen J Christensen PB Kristensen T Characterization of tumour infiltrating lymphocytes and correlations with immunological surface molecules in colorectal cancer Eur J Cancer 1999 35 721 726 10505031 10.1016/S0959-8049(99)00033-7
Ikeguchi M Oi K Hirooka Y Kaibara N CD8+ lymphocyte infiltration and apoptosis in hepatocellular carcinoma Eur J Surg Oncol 2004 30 53 57 14736523 10.1016/j.ejso.2003.10.001
Yakirevich E Izhak OB Rennert G Kovacs ZG Resnick MB Cytotoxic phenotype of tumor infiltrating lymphocytes in medullary carcinoma of the breast Mod Pathol 1999 12 1050 1056 10574602
Mercader M Bodner BK Moser MT Kwon PS Park ES Manecke RG Ellis TM Wojcik EM Yang D Flanigan RC Waters WB Kast WM Kwon ED T cell infiltration of the prostate induced by androgen withdrawal in patients with prostate cancer Proc Natl Acad Sci USA 2001 98 14565 14570 11734652 10.1073/pnas.251140998
Kobayashi M Ito M Sano K Honda T Nakayama J Pulmonary lymphoepithelioma-like carcinoma: predominant infiltration of tumor-associated cytotoxic T lymphocytes might represent the enhanced tumor immunity Intern Med 2004 43 323 326 15168777 10.2169/internalmedicine.43.323
Noonan K Matsui W Serafini P Carbley R Tan G Khalifi J Bonyhadi M Levitsky H Whartenby K Borrelo I Activated marrow-infiltrating lymphocytes effectively target plasma cells and their clonogenic precursors Cancer Res 2005 65 2026 2034 15753403 10.1158/0008-5472.CAN-04-3337
Rivoltini L Arienti F Orazi A Cefalo G Gasparini M Gambacorti-Passerini C Fossati-Bellani F Parmiani G Phenotypic and functional analysis of lymphocytes infiltrating paediatric tumours, with a characterization of the tumour phenotype Cancer Immunol Immunother 1992 34 241 251 1311218 10.1007/BF01741792
Zietz C Rumpler U Sturzl M Lohrs U Inverse relation of Fas-ligand and tumor-infiltrating lymphocytes in angiosarcoma: indications of apoptotic tumor counterattack Am J Pathol 2001 159 963 970 11549589
Barber MR Yang TJ Tumor infiltrating lymphocytes: CD8+ lymphocytes in canine transmissible venereal sarcomas at different stages of tumor growth Anticancer Res 1999 19 1137 1142 10368665
Yu R Fujio K Tahara H Araki Y Yamamoto K Clonal dynamics of tumor-infiltrating lymphocytes Eur J Immunol 2005 35 1754 1763 15902685 10.1002/eji.200425866
Rosenberg SA Yannelli JR Yang JC Topalian SL Schwartzentruber DJ Weber JS Parkinson DR Seipp CA Einhorn JH White DE Treatment of patients with metastatic melanoma with autologous tumor-infiltrating lymphocytes and interleukin 2 J Natl Cancer Inst 1994 86 1159 1166 8028037
Dreno B Nguyen JM Khammari A Pandolfino MC Tessier MH Bercegeay S Cassidanius A Lemarre P Billaudel S Labarriere N Jotereau F Randomized trial of adoptive transfer of melanoma tumor-infiltrating lymphocytes as adjuvant therapy for stage III melanoma Cancer Immunol Immunother 2002 51 539 546 12384805 10.1007/s00262-002-0315-1
Burger UL Chang MP Nagoshi M Goedegebuure PS Eberlein TJ Improved in vivo efficacy of TIL after restimulation with irradiated tumor cells in vitro Ann Surg Oncol 1996 3 580 587 8915492
Harada M Matsuzaki G Shinomiya Y Kurosawa S Ito O Okamoto T Takenoyama M Sumitika H Nishimura Y Nomoto K Generation of tumor-specific cytotoxic T lymphocytes in vivo combined treatment with inactivated tumor cells and recombinant interleukin-2 Cancer Immunol Immunother 1994 38 332 338 8162615
Yang JC Perry-Lalley D Rosenberg SA An improved method for growing murine tumor-infiltrating lymphocytes with in vivo antitumor activity J Biol Response Mod 1990 9 149 159 1971302
Rosenberg SA Spiess P Lafreniere R A new approach to the adoptive immunotherapy of cancer with infiltrating lymphocytes Science 1986 233 1318 1321 3489291
Klein B Pals S Masse R Lafuma J Morin M Binart N Jasmin JR Jasmin C Studies of bone and soft-tissue tumours induced in rats with radioactive cerium chloride Int J Cancer 1977 20 112 119 269099
Thiéry J-P Perdereau B Gongora R Gongora G Mazabraud A Un modèle experimental d'ostéosarcome chez le rat. II L'ostéosarcome greffable du rat Sem Hôp Paris 1982 58 1686 1689 8424255
Heymann D Ory B Blanchard F Heymann MF Coipeau P Charrier C Couillaud S Thiéry JP Gouin F Rédini F Enhanced tumor regression and tissue repair when zoledronic acid is combined with ifosfamide in rat osteosarcoma Bone 2005 37 74 86 15894525 10.1016/j.bone.2005.02.020
Lynch DH Miller RE Immunotherapeutic elimination of syngenic tumors in vivo by cytotoxic T lymphocytes generated in vitro from lymphocytes from the draining lymph nodes of tumor-bearing mice Eur J Immunol 1991 21 1403 1410 1904359
Melief CJ Kast WM T-cell immunotherapy of tumors by adoptive transfer of cytotoxic T lymphocytes and by vaccination with minimal essential epitopes Immunol Rev 1995 145 167 177 7590825
Rivoltini L Arienti F Orazi A Cefalo G Gasparini M Gambacorti-Passerini C Fossati-Bellani F Parmiani G Phenotypic and functional analysis of lymphocytes infiltrating paediatric tumours, with a characterization of the tumour phenotype Cancer Immunol Immunother 1992 34 241 251 1311218 10.1007/BF01741792
Haanen JB Baars A Gomez R Weber P Smits M de Gruijl TD von Blomberg BM Bloemena E Scheper RJ van Ham SM Pinedo HM van den Eertwegh AJ Melanoma-specific tumor-infiltrating lymphocytes but not circulating melanoma-specific T cells may predict survival in resected advanced-stage melanoma patients Cancer Immunol Immunother
Trieb K Lechleitner T Lang S Windhager R Kotz R Dirnhofer S Evaluation of HLA-DR expression and T-lymphocyte infiltration in osteosarcoma Pathol Res Pract 1998 194 679 684 9820863
Topalian SL Muul LM Rosenberg SA Growth and immunological characteristics of lymphocytes infiltrating human tumors [abstract] Surg Forum 1987 37 390
Antony PA Piccirillo CA Akpinarli A Finkelstein SE Speiss PJ Surman DR Palmer DC Chan CC Klebanoff CA Overwijk WW Rosenberg SA Resifo NP CD8+ T cell immunity against a tumor/self-antigen is augmented by CD4+ T helper cells and hindered by naturally occurring T regulatory cells J Immunol 2005 174 2591 601 15728465
Yu P Lee Y Liu W Krusz T Chong A Schreiber H Fu YX Intratumor depletion of CD4+ cells unmasks tumor immunogenicity leading to the rejection of late-stage tumors J Exp Med 2005 201 779 791 15753211 10.1084/jem.20041684
Ormandy LA Hillemann T Wedemeyer H Manns MP Greten TF Korangy F Increased populations of regulatory T cells in peripheral blood of patients with hepatocellular carcinoma Cancer Res 2005 65 2457 2464 15781662 10.1158/0008-5472.CAN-04-3232
Kojima H Kanno Y Hase H Kobata T CD4+CD25+ regulatory T cells attenuate the phosphatidylinositol 3-kinase/Akt pathway in antigen-primed immature CD8+ CTLs during functional maturation J Immunol 2005 174 5959 5967 15879088
Erman SE Sohn JJ Rao VP Nambiar PR Ge Z Fox JG Schauer DB CD4+CD25+ regulatory lymphocytes induce regression of intestinal tumors in ApcMin/+ mice Cancer Res 2005 65 3998 4004 15899788 10.1158/0008-5472.CAN-04-3104
Sudo T Kuramoto T Komiya S Inoue A Itoh K Expression of MAGE genes in osteosarcoma J Orthop Res 1997 15 128 132 9066537 10.1002/jor.1100150119
Ishida H Komiya S Inoue Y Yutani S Inoue A Itoh K Expression of the SART1 tumor-rejection antigen in human osteosarcomas Int J Oncol 2000 17 29 32 10853014
Tsuda N Murayama K Ishida H Matsunaga K Komiya S Itoh K Yamada A Expression of a newly defined tumor-rejection antigen SART3 in musculoskeletal tumors and induction of HLA class I-restricted cytotoxic T lymphocytes by SART3-derived peptides J Orthop Res 2001 19 346 351 11398844 10.1016/S0736-0266(00)90031-7
Tsukahara T Nabeta Y Kawaguchi S Ikeda H Sato Y Shimozawa K Ida K Asanuma H Hirohashi Y Torigoe T Hiraga H Nagoya S Wada T Yamashita T Sato N Identification of human autologous cytotoxic T-lymphocyte-defined osteosarcoma gene that encodes a transcriptional regulator, papillomavirus binding factor Cancer Res 2004 64 5442 5448 15289353 10.1158/0008-5472.CAN-04-0522
Miyagi Y Imai N Sasatomi T Yamada A Mine T Katagiri K Nakagawa M Muto A Okouchi S Isomoto H Shirouzu K Yamana H Itoh K Induction of cellular immune responses to tumor cells and peptides in colorectal cancer patients by vaccination with SART3 peptides Clin Cancer Res 2001 7 3950 3962 11751487
Trieb K Lang S Kotz R Heat-shock protein 72 in human osteosarcoma: T-lymphocytes reactivity and cytotoxicity Pediatr Hematol Oncol 2000 17 355 364 10914045 10.1080/08880010050034283
|
16188028
|
PMC1262697
|
CC BY
|
2021-01-04 16:03:05
|
no
|
BMC Cancer. 2005 Sep 27; 5:123
|
utf-8
|
BMC Cancer
| 2,005 |
10.1186/1471-2407-5-123
|
oa_comm
|
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-1251620213210.1186/1471-2407-5-125Research ArticleCombining RNA interference and kinase inhibitors against cell signalling components involved in cancer O'Grady Michael [email protected] Debasish [email protected] Bonnie J [email protected] Michaeline [email protected] George T [email protected] Invitrogen Corporation, 501 Charmany Drive, Madison, WI 53719 USA2 Invitrogen Corporation, 1600 Faraday Avenue, Carlsbad, CA 92008 USA2005 3 10 2005 5 125 125 15 7 2005 3 10 2005 Copyright © 2005 O'Grady et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The transcription factor activator protein-1 (AP-1) has been implicated in a large variety of biological processes including oncogenic transformation. The tyrosine kinases of the epidermal growth factor receptor (EGFR) constitute the beginning of one signal transduction cascade leading to AP-1 activation and are known to control cell proliferation and differentiation. Drug discovery efforts targeting this receptor and other pathway components have centred on monoclonal antibodies and small molecule inhibitors. Resistance to such inhibitors has already been observed, guiding the prediction of their use in combination therapies with other targeted agents such as RNA interference (RNAi). This study examines the use of RNAi and kinase inhibitors for qualification of components involved in the EGFR/AP-1 pathway of ME180 cells, and their inhibitory effects when evaluated individually or in tandem against multiple components of this important disease-related pathway.
Methods
AP-1 activation was assessed using an ME180 cell line stably transfected with a beta-lactamase reporter gene under the control of AP-1 response element following epidermal growth factor (EGF) stimulation. Immunocytochemistry allowed for further quantification of small molecule inhibition on a cellular protein level. RNAi and RT-qPCR experiments were performed to assess the amount of knockdown on an mRNA level, and immunocytochemistry was used to reveal cellular protein levels for the targeted pathway components.
Results
Increased potency of kinase inhibitors was shown by combining RNAi directed towards EGFR and small molecule inhibitors acting at proximal or distal points in the pathway. After cellular stimulation with EGF and analysis at the level of AP-1 activation using a β-lactamase reporter gene, a 10–12 fold shift or 2.5–3 fold shift toward greater potency in the IC50 was observed for EGFR and MEK-1 inhibitors, respectively, in the presence of RNAi targeting EGFR.
Conclusion
EGFR pathway components were qualified as targets for inhibition of AP-1 activation using RNAi and small molecule inhibitors. The combination of these two targeted agents was shown to increase the efficacy of EGFR and MEK-1 kinase inhibitors, leading to possible implications for overcoming or preventing drug resistance, lowering effective drug doses, and providing new strategies for interrogating cellular signalling pathways.
==== Body
Background
Cellular processes such as proliferation, differentiation, and death are regulated by signal transduction pathways which commonly exert their function through receptor mediated activation. The discovery in 1978 that the v-Src oncogene was a protein kinase led to a "cascade" of research into the role of kinases in cell-signalling pathways, and the subsequent finding that human cancer can result from the activity of nonviral, endogenous oncogenes, a major portion of which code for protein tyrosine kinases (PTKs) [1,2]. The epidermal growth factor receptor (EGFR) is a tyrosine kinase which acts as a master switch leading to activation of the transcription factor, activator protein-1 (AP-1), and other related pathways. The receptor itself is composed of extracellular, transmembrane, and tyrosine kinase domains. Ligand binding elicits a conformational change of the extracellular domain leading to receptor dimerization and subsequent transphosphorylation of intracellular domain tyrosines. The phosphorylated tyrosines act as binding sites for signal transducers initiating a series of kinase actions resulting in cellular proliferation and differentiation [3-5]. Aberrant signalling occurring from EGFR results in its conversion into an oncoprotein, and the consequent malfunction of cellular signalling networks leads to the development of cancers and other proliferative diseases. EGFR and its ligands are involved in over 70% of all cancers [[4,6], and [7]].
Hidaki, et.al. in the early 1980's discovered the first protein-kinase inhibitors, and established the principle of changing chemical structure to elicit different kinase inhibition specificity [8]. Drug development has followed the lead of the academic community in developing novel inhibitory compounds at points along these disease-related pathways. The protein kinase target class is now the second largest group of drug targets behind G-protein-coupled-receptors [3]. Kinases of the Tyrosine and Serine/Threonine family have been targeted successfully by small-molecule inhibitors and monoclonal antibodies, with many undergoing human clinical trials or successfully launched as therapeutic entities [9-13].
Acquired resistance to kinase-targeted anticancer therapy has been documented, and most extensively studied with imatinib (Gleevec™), an inhibitor of the aberrant BCR-ABL kinase, in chronic myelogenous leukemia [14]. Resistance has also occurred in EGFR-targeted inhibitor therapy using gefitinib (Iressa™) and erlotinib (Tarceva™). Mutations occurring in the catalytic domain of the receptor have been implicated in this resistance, but cannot account for all resistance seen to these small molecule inhibitors, indicating other mechanisms are involved in the resistance seen to date [15,16]. Therefore, multiple strategies will be necessary to overcome the observed resistance to these new molecularly targeted therapies, as well as methods to predict their efficacy.
Most kinase inhibitors target the ATP-binding site common to all kinases, and can bind multiple kinases [17]. This generates an inability to predict compound specificity for a particular kinase, and the subsequent need to analyze large numbers of kinases through a screening or profiling approach. Data from these in vitro assays allow the researcher to predict clinical uses for inhibitors and possible offsite target effects. Studies using purified kinase and substrate are dependent on ATP concentration used, and the apparent Km for ATP can differ between kinases. This can lead to problems in the development of small molecule inhibitors based on competition at the ATP-binding site of a kinase, as the ATP concentration in vivo may differ greatly from that used in vitro. In addition, kinase activity studies in a purified setting may use domains of proteins and peptide substrates, which can lead to erroneous interpretation of the true nature of kinase activity and/or inhibition. The in vivo studies using Western blot analysis also can be difficult to interpret due to the need to use a protein preparation from a cellular lysate, and inherent variability when using antibodies for Western blot analysis. Small changes in any step of the protocol could lead to differences in interpretation of the results. For these reasons, and the need for strategies to prevent or overcome resistance formation in malignancies, we have used an in vitro and functional cellular assay approach to study the EGFR/AP-1 signal transduction pathway. AP-1 activation through EGFR was assessed using a β-lactamase reporter gene assay, and served as a model for inhibition of pathway components on a functional cellular level. Kinase profiling using full length EGFR and peptide substrates was used in parallel for confirmation and specificity of inhibition. The knockdown of qualified targets in the EGFR/AP-1 pathway was further studied by immunocytochemistry, allowing for assessment on a cellular protein level. Following qualification of targets in this cancer-related pathway, a functional cellular assay was used to analyze the potential therapeutic benefit of combining RNAi and kinase inhibitors against EGFR and MEK-1 in the AP-1 activation pathway of a human cervical cancer cell line.
Methods
AP-1-bla ME180 CellSensor™cell line
An AP-1 response element (TGACTAA, 7X) was inserted into the MCS of the β-lactamase lentiviral reporter vector using Gateway® technology according to the manufacturer's instructions (Invitrogen, Carlsbad, CA). The virus created from this vector was used to transduce ME180 cells [18] according to manufacturer's instructions (Invitrogen, Carlsbad, CA). Following a one day incubation, the cells were split and placed in media with Blasticidin (5 ug/ml) for 3–4 days. Cells were then washed with phosphate buffered saline (PBS), refed, and further selected for seven to ten days in media containing Blasticidin. Flow cytometry was used to select for β-lactamase expressing cells according to previously published protocols [19].
EGF ligand stimulation of AP-1-bla ME180
AP-1-bla ME180 cells were plated in 96 well plates at a density of 20,000–25,000 cells/well in DMEM with 1 mM sodium pyruvate, 25 mM HEPES, 0.1 mM non-essential amino acids (Invitrogen Gibco, Grand Island, NY) plus 1% fetal bovine serum (FBS) and incubated overnight at 37°C with 5% CO2. The following day rhEGF (Calbiochem, San Diego, CA) was diluted in media with 1% FBS at desired concentrations, and cells incubated at 37°C with 5% CO2 for 5 hours. EGF treatment was performed at n = 8.
β-lactamase reporter analysis
β-lactamase reporter gene expression was determined using the LiveBLAzer™ FRET B/G assay kit (Invitrogen Drug Discovery, Madison, WI) according to manufacturer's instructions.
Inhibitor treatment of AP-1-bla ME180
Small molecule inhibitors (Calbiochem, San Diego, CA) were diluted in cell culture media at the desired concentration and preincubated with the cells for 30 minutes at 37°C with 5% CO2. EGF ligand stimulation was performed as indicated above followed by β-lactamase reporter analysis. DMSO was used as the negative control, per its use as the compound reconstitution medium. Final DMSO concentration in the medium was 0.05% for both compounds and negative controls.
RNAi design and transfection of AP-1-bla ME180
dsRNAi Stealth™ oligos were designed against EGFR. The following sequences were used for the oligos:
EGFR: Sense 5' UUA GAU AAG ACU GCU AAG GCA UAG G 3'
Anti-Sense 5' CCU AUG CCU UAG CAG UCU UAU CUA A 3'
AP-1-bla ME180 cells were plated at a concentration of 10,000 cells/well in a 96 well plate and incubated overnight at 37°C with 5% CO2. AP-1-bla ME180 cells were transfected with 50 nM dsRNAi Stealth™ oligos and 2 μg/ml Lipofectamine 2000 according to manufacturer's suggestions (Invitrogen, Carlsbad, CA). mRNA was extracted from transfected cells 24 hours post transfection and EGFR expression levels determined by RT-qPCR using light upon extension (LUX™) primer sets (Invitrogen, Carlsbad, CA) for the target of interest and cyclophilin as a normalization control. Percent knockdown of the targeted message was determined as a ratio of target versus cyclophilin control.
Functional cellular analysis of the effect of RNAi knockdown was studied using the AP-1-bla ME180 CellSensor. The cells were transfected with ds RNAi oligos as above and incubated for various lengths of time prior to EGF stimulation for 5 hours and subsequent quantitation of β-lactamase using the LiveBLAzer FRET B/G assay kit (Invitrogen).
Immunocytochemistry for cellular protein knockdown and phosphorylation status of EGFR
AP-1-bla ME180 cells were analyzed for EGFR knockdown at a cellular protein level following RNAi treatment. Cells were transfected with dsRNAi oligos as in the functional cellular studies above and incubated for ~60 hrs. The cells were then fixed with 4% Paraformaldehyde for 30 minutes at 22°C, followed by permeabilization in PBS + 0.25% Triton X-100 for 5 mins at 22°C. Blocking of antigen binding sites was performed with PBS + 2% FBS for 2 hours at 22°C with rocking. Cells were washed twice with PBS + 2% FBS, then incubated with rabbit anti-phospho-EGFR (Tyr1086) or mouse anti-EGFR (31G7) primary antibody (Invitrogen Zymed, San Francisco, CA) at 1:100 concentration and 1:25 concentration respectively in PBS + 1% FBS for 1 hour at 22°C with rocking. Following primary antibody incubation, the cells were washed twice with PBS + 1% FBS at 22°C with rocking. The secondary antibodies goat anti-rabbit IgG (H+L) Alexa Fluor® 488 and goat anti-mouse IgG (H+L) Alexa Fluor® 594 (Invitrogen Molecular Probes, Eugene, OR) were used at 1 μg/mL concentration in PBS + 1% FBS, and incubated at 22°C with rocking. Cells were washed three times with PBS + 1% FBS followed by addition of PBS + 10% glycerol for assay and storage. A Zeiss Axiovert 25CFL microscope with a FITC filter set (excitation D480/30X, emission D535/40 M, Chroma Technology Corporation, Rockingham VT) for Alexa 488, and a Texas Red filter set (excitation D560/40X, emission D630/60 M, Chroma Technology Corporation, Rockingham VT) for Alexa 594 was used for imaging of cells. Quantitation of knockdown was done using a Tecan Safire® instrument using excitation wavelength of 485/7.5 nm and emission wavelength of 510/7.5 nm for the Alexa 488 labelled secondary antibody and excitation wavelength of 590/7.5 nm and emission wavelength of 615/7.5 nm for the Alexa 594 labelled secondary antibody.
Immunocytochemistry for EGFR phosphorylation following kinase inhibitor treatment
AP-1-bla ME180 cells were treated with kinase inhibitors as outlined above. Immunocytochemistry using appropriate phospho-specific antibodies was performed as in the RNAi experiments to analyze EGFR autophosphorylation.
Combining RNAi and kinase inhibitors on the AP-1-bla ME180 CellSensor
AP-1-bla ME180 cells were treated with 50 nM RNAi against EGFR as in experiments outlined above. Only one-half of a 96 well plate was treated, while the other half was mock treated with Lipofectamine™ 2000 (Invitrogen, Carlsbad, CA) alone. At 60 hours the cells were treated with a dilution series of AG1478 or U0126 as outlined above. Cells were then stimulated with EGF and the resulting β-lactamase readout was analyzed using GraphPad Prism® to determine IC50 values. Controls for RNAi transfection (Med GC; 40–50% GC nucleotide content), EGFR knockdown (EGFR RNAi), and β-Lactamase knockdown (β-Lac) were used to set standards for transfection, single treatment, and maximal effects respectively.
Kinase profiling of compounds used in functional cellular assays
Small molecule compounds used in the cellular assays were profiled in the SelectScreen™ Kinase Profiling Service (Invitrogen Drug Discovery Solutions, Madison, WI). This service utilizes the Z'-Lyte technology assay platform [21] where the biochemical assays were performed at a final concentration of 1 μM compound in 0.1% DMSO and an ATP concentration of Km, app for the EGFR protein. The assays were analyzed on a Tecan Safire2® detection instrument. The percent inhibition values were calculated using XL fit 4.0 by ID-BS.
Results
ME180 EGFR/AP-1 signal transduction pathway is responsive to EGF stimulation
In order to use the AP-1-bla ME180 CellSensor as a model system for combining targeted agents against components of a cancer-related pathway, we determined the dose-dependent response of the pathway to EGF stimulation. The EC50 was determined to be 0.31 ng/mL from the blue:green ratiometric readout of the β-lactamase assay (Figure 1A). Imaging capabilities made possible with the β-lactamase reporter system allow for microscopic visualization of reporter enzyme activity due to cleavage of the FRET-based substrate (Figure 1B). To maximize the assay window while maintaining sensitivity to inhibition by putative inhibitors, an EC80 concentration of 1 ng/ml EGF was used to stimulate pathway response in inhibition assays.
EGFR/AP-1 signal transduction can be inhibited at multiple points in the pathway
The commercial availability of known small molecule kinase inhibitors allowed for the study of their potency in our cellular system. Prior to analyzing efficacy in a functional cellular assay, we tested their inhibition of EGFR in a biochemical assay using Z'-LYTE, a FRET-based platform.
The small molecule compounds are structurally related with some targeting the ATP binding site of the kinase (AG1478, CL387785, PD153035, SB202190) [20], and others targeting the tyrosine substrate binding site (AG490, AG183) [4,22]. U0126 acts in the unique manner of blocking activation of MEK through a mechanism independent of the ATP binding site [12,23]. The kinase inhibitors tested show a spectrum of inhibition against EGFR (Figure 2), and correlate well with published results of inhibitor specificity [17].
The inhibitors were next tested for their effectiveness in inhibiting the EGFR/AP-1 pathway in a functional cellular assay. Initial analysis was performed with an inhibitor concentration of 500 nM. Results indicate the CL387785, PD153035, and AG1478 compounds strongly inhibit AP-1 gene activation (Figure 3A), indicating that targeting of EGFR leads to a potent reduction in AP-1 gene expression in ME180 EGF stimulated cells. AG183 shows no inhibition at 500 nM and also showed no inhibition of EGFR in the biochemical assay (Figure 2). The U0126 compound, which targets MEK-1, also shows inhibition of the EGFR/AP-1 pathway (Figure 3A), an expected result when targeting another component of the EGFR/AP-1 pathway. The p38 MAP kinase inhibitor SB202190 shows no inhibition of the pathway in the AP-1-bla ME180 CellSensor as could be predicted based on known pathway components. The live cell assay format allows for imaging and further confirmation of the inhibition shown in the quantitative panel (Figure 3B).
The inhibitors that tested positive for inhibition in the functional cellular assay at a 500 nM concentration were further studied in a dose response manner. Inhibitor potency could be assigned through IC50 values determined on a cellular pathway level. The EGFR inhibitors CL387785, PD153035, and AG1478 show increased efficacy of pathway inhibition when compared to U0126, as might be expected when targeting the initial component of this receptor linked pathway (Figure 4A, B, and 4C).
Cellular phosphorylation of EGFR on a cellular protein level matches reporter gene results
Confirmation on a cellular protein level of the results obtained using kinase inhibitors on the EGFR/AP-1 pathway in a cell-based reporter readout was obtained through immunocytochemistry. Inhibition of autophosphorylation was shown by using antibodies specific for the phosphorylated tyrosine-1086 of EGFR. PD153035 showed strong inhibition of EGFR autophosporylation, while AG183 did not (Figure 5A and 5B) validating the results seen in the functional cellular assay (figure 4).
RNAi qualifies EGFR as a target for AP-1 pathway inhibition
dsRNAi towards EGFR was shown to knockdown the total amount of EGFR present as well as the autophosphorylation of the receptor. Analysis of knockdown was analyzed by RT-PCR demonstrating an ~80% knockdown on an mRNA level (Figure 6A). Immunocytochemistry performed on RNAi treated AP-1-bla ME180 cells shows a knockdown of EGFR on a cellular protein level with a concomitant loss of autophosphorylation of the receptor (Figure 6B and 6C). Analysis of the effect of RNAi knockdown in a functional cellular assay confirmed the essential role of EGFR in the activation of AP-1 (Figure 6D). In addition, lower receptor levels result in less pathway response, and the stronger effect on AP-1 activation correlates with RNAi incubation time. Additional controls using a low GC (30–40% GC content) dsRNAi oligo and dsRNAi oligos targeting genes not involved in the EGFR/AP-1 pathway (NF-κB, IKK-α) show no effect on a cellular protein level or in a functional cellular assay (data not shown).
Combining EGFR RNAi with AG1478 or U0126 increases the potency of the kinase inhibitors against AP-1 activation
To determine whether there were additive effects, which could be relevant in therapeutic efficacy and prevention of drug resistance (15,16, and 24), we performed experiments using small molecule inhibitors in combination with RNAi. U0126, a known MEK-1 inhibitor, exhibits an IC50 of 543 nM when used in isolation on the AP-1 cell line. Upon combination with RNAi against EGFR, the IC50 for the MEK-1 inhibitor shifts 2.5–3 fold more potent to 192 nM (Figure 7A, B, and 7C). AG1478, an EGFR inhibitor, exhibits an IC50 of 18 nM in isolation, but shifts to 1.5 nM when used in tandem with RNAi targeting EGFR (Figure 7D, E, and 7F). The results indicate the validity of using a functional cellular assay for the preclinical analysis of putative benefits when combining kinase inhibitors with other targeted agents of signal transduction pathways related to cancer.
Discussion
Molecular targeted therapies against signalling pathway components involved in cancer have arrived in the clinic, and drug discovery efforts are increasing in this evolving area [[3,9,11,13,24], and [25]]. The growth of screening for small molecule compounds which act as kinase inhibitors has led to their becoming the second most targeted group of druggable entities after G-protein-coupled receptors [26].
Modulation of kinase activity can be accomplished by strategies other than inhibition of phosphorylation activity through the blocking of ATP binding. Such methods include disruption of protein-protein interactions and the knockdown or downregulation of kinase gene expression by antisense or RNA interference approaches [16,24]. The need for multiple approaches for therapies targeting kinases can be seen in the reports of resistance towards the recently launched kinase inhibitors gefitinib (Iressa™) and imatinib (Gleevec™), which inhibit the EGFR and Bcr-Abl kinase respectively [14,27]. Strategies implemented for overcoming or preventing this resistance have included chemical modifications of the inhibitor compound using a rational drug design strategy to increase the potency against the targeted kinase. The ATP binding site of kinases has proven to be a "hot spot" for kinase mutations and includes a "gatekeeper" region shown to be difficult to block with inhibitors and their new more potent derivatives. Recent work has been published on inhibitors targeting the substrate binding site of the Bcr-Abl kinase [28], which seem to inhibit wild-type and all imatinib-resistant kinase domain mutations, including the "gatekeeper" mutation [29]. These approaches have proven successful for imatinib and will undoubtedly work in the near term for other small molecule inhibitors, but experience suggests the possibility of further mutations leading to increased resistance [[15,16], and [24]].
Blockade of one kinase alone might not be sufficient to achieve needed pathway inhibition, and so targeting of multiple kinases could be more promising in terms of efficacy and prevention of resistance. This type of combinatorial therapeutic approach has become the standard for HIV treatment to maximize potency, minimize toxicity, and diminish the risk for resistance development [30]. Targeted compounds could be used together or in combination with toxic agents such as chemotherapy or ionizing radiation, as well as with other novel agents. This approach has proven successful in recent studies using chemotherapeutic agents or ionizing radiation in combination with kinase inhibitors [[31-33] and [34]]. However, given the rapidly growing number of inhibitory agents and an exponential number of possible combinations, it will not be possible to test all such groupings in a clinical trial setting [16,24]. The need for predictive preclinical models allowing for the choice of which studies to advance through the drug discovery process led to the experiments outlined in this paper with RNAi and the kinase inhibitors U0126 and AG1478.
First, in order to establish a functional cellular model for investigation of the EGFR/AP-1 pathway we built a stable cell line responsive to EGF stimulation. Kinase inhibitors and RNAi analysis established components that were involved in the functional response of the EGFR/AP-1 pathway that could be inhibited by molecularly targeted agents.
Biochemical analysis of EGFR inhibition using the same small molecule inhibitors further supported our cellular results and demonstrated the importance of complimentary approaches when analyzing signal transduction pathways. Inhibitors such as U0126 inhibit the functional response of the AP-1 pathway as shown in the cellular reporter readout, but work at a level independent of EGFR inhibition determined by in vitro assays.
RNAi analysis and immunocytochemistry further demonstrated the essential role EGFR plays at the beginning of the AP-1 activation cascade. Showing that inhibition or knockdown of the receptor leads to the functional results observed in the cellular assay model, led us to analyze combinatorial effects of using RNAi and kinase inhibitors in tandem. When the targeted agents were used together, an increased potency of the inhibitors was observed. This finding demonstrates the use of a cellular reporter system to predict the cocktail effects of a growing number of targeted agents against cell signalling components involved in cancer.
Conclusion
Our results demonstrate the essential role for EGFR in AP-1 activation when analyzed using an ME180 cervical cancer cell line. Confirmation of "druggable" targets was shown using the parallel approach of known kinase inhibitors and RNA interference. An immunocytochemistry approach using phospho- and pan-specific EGFR antibodies strengthened the argument for the use of the AP-1-bla ME180 cell line as a viable model for the analysis of a combinatorial targeted agent approach to cancer-related pathways. Results obtained combining RNAi toward EGFR and small molecule inhibitors of MEK-1 and EGFR indicate a beneficial effect on EGFR/AP-1 pathway inhibition. This cellular model approach could lead to further studies combining kinase inhibitors with other targeted agents, and suggests possible implications for screening of compound libraries to uncover novel pathway inhibitors.
Abbreviations
AP-1 = Activator Protein 1, bla = beta-lactamase, EGFR = Epidermal Growth Factor Receptor, MEK1 = Mitogen Activated Protein Kinase Kinase 1, Phosphate Buffered Saline PBS, PTKs = Protein Tyrosine Kinases, RNAi = RNA interference.
Competing interests
The authors are all employees of Invitrogen Corporation whose products are used in this article for research purposes.
Authors' contributions
MO performed all experiments involving the ME180 AP-1 CellSensor with kinase inhibitors, RNAi, and immunocytochemistry, as well as coordinating and executing the studies performed. BJH assisted with immunocytochemistry as well as coordination and execution of the study. MB and DR assisted and coordinated RNAi experiments included in the study. GTH assisted in all experiments, as well as in the coordination and execution of the study. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional File 1
EGFR/AP-1 pathway Schematic representation of EGFR/AP-1 pathway indicating components targeted by inhibitors or RNAi.
Click here for file
Acknowledgements
Thanks to Greg Parker for figure preparation. Andy Kopp, Tom Zielinski, Jessica Honer, Heidi Braun, Jeff Beauchaine, Dave Lasky, Leah Aston, and Jenny Fronczak at Invitrogen-Madison, and Kristin Wiederholt and Mason Brooks at Invitrogen-Carlsbad, for research support. Also thanks to Brian Pollok, Kurt Vogel, Tammy Turek-Etienne, and Peter Welch for critical reading of the manuscript.
Figures and Tables
Figure 1 EGF mediated activation in ME180AP-1 CellSensor. EGF effectively stimulated the EGFR/AP-1 pathway in ME180 cervical cancer cells. Cells expressing β-lactamase were detected using a fluorescent plate reader (Panel A) or under fluorescent microscopy (Panel B). β-lactamase was measured using Invitrogen's LiveBLAzer™ FRET B/G assay. Ratios represent the mean of (± SEM) eight independent data points (n = 8).
Figure 2 Biochemical assay results with kinase inhibitors. Percent inhibition values determined by using the Z'-LYTE technology platform (Invitrogen), and single point concentration testing (1 μM) with two independent data points (n = 2). Small molecule compounds CL387785, PD153035, and AG1478 show strong inhibition of the EGFR, while others tested exhibit none.
Figure 3 Kinase inhibitors block activation at multiple points in EGFR/AP-1 pathway. Inhibitors decrease the response of the AP-1-bla ME180 CellSensor when stimulated with EGF. β-lactamase was measured using Invitrogen's LiveBLAzer™ assay (Panel A). Percent inhibition is determined in comparison to untreated sample and represents the mean of (± SEM) eight independent data points (n = 8). Photomicrographs of cells were obtained under fluorescent microscopy (Panel B).
Figure 4 Potency of individual inhibitors towards kinases in the AP-1 pathway determined in a dose response manner. Inhibitors were used in a dose response manner on the AP-1-bla ME180 CellSensor, and β-lactamase measured using Invitrogen's liveBLAzer™ FRET B/G assay (Panels A and B). IC50 determinations were performed using GraphPad Prism® and values shown with images of inhibitors at 0.2 μM concentration (Panel B).
Figure 5 Autophosphorylation of EGFR shows similar inhibition profile as the AP-1-bla ME180 functional cellular assay. Immunocytochemistry with phospho-specific antibodies demonstrates the same inhibition profile as the AP-1 activation assay performed using the ME180 CellSensor. Fluorescence was measured from the Alexa fluor 488 and 594 labelled secondary antibodies used for detection (Panel A). Percent inhibition was determined in comparison to untreated sample and represents the mean of (± SEM) eight independent data points (n = 8). Images were obtained from the Alexa fluor 488 labelled secondary antibody under fluorescent microscopy (Panel B).
Figure 6 Knockdown by RNAi qualifies EGFR involvement in ME180 AP-1 CellSensor pathway. AP-1-bla ME180 cells transfected with RNAi targeting EGFR were analyzed by RT-qPCR (Panel A). EGFR mRNA expression is knocked down ~80% as compared to a non-specific control (Med GC). Results are represented as a ratio using cyclophilin as a control. Cells RNAi treated with oligos specific for EGFR show a reduced level of EGFR protein on the cell surface including a loss of autophosphorylation (Panels B and C). RFU represent measurement of the Alexa fluor 488 and 594 labelled secondary antibodies used for detection of the phospho and pan EGFR antibodies and represent the mean of (± SEM) eight independent data points (n = 8). RNAi treated AP-1-bla ME180 cells also exhibit a knockdown of AP-1 activation when EGF stimulated and analyzed at different timepoints (Panel D). Normalized response represents the amount of AP-1 gene activation present when compared with a non-specific control (Med GC), and represent the mean of (± SEM) eight independent data points (n = 8).
Figure 7 Combining RNAi toward EGFR and the kinase inhibitors U0126 and AG1478 increases the potency of the small molecule compounds. Small molecule kinase inhibitors shown to have inhibitory properties in the AP-1 CellSensor assay were used in tandem with RNAi. RNAi targeting the EGFR was kept constant while U0126 targeting MEK-1 kinase (Panel A) and AG1478 targeting EGFR (Panel C) were used in a dose response manner to examine potency effects of the combined treatment. The shift in IC50 indicates the "cocktail" effect seen with the tandem treatment.
==== Refs
Collett M Erikson R Protein kinase activity associated with the avian sarcoma virus src gene product Proc Natl Acad Sci 1978 75 2021 2024 205879
Atalay G Cardaso F Awada A Piccart MJ Novel therapeutic strategies targeting the epidermal growth factor receptor (EGFR) family and its downstream effectors in breast cancer Annals of Oncology 2003 14 1346 1363 12954573 10.1093/annonc/mdg365
Cohen P Protein kinases-the major drug targets of the twenty-first century? Nat Rev Drug Discov 2002 1 309 315 12120282 10.1038/nrd773
Levitzki A Tyrosine kinases as targets for cancer therapy European Journal of Cancer 2002 38 S11 S18 12528768 10.1016/S0959-8049(02)80598-6
Tiseo M Loprevite M Ardizzoni A Epidermal growth factor receptor inhibitors: a new prospective in the treatment of lung cancer Curr Med Chem Anti-Canc Agents 2004 4 139 148 15032719 10.2174/1568011043482106
Perez-Soler R HER1/EGFR Targeting: Refining the Strategy The Oncologist 2004 9 58 67 14755015 10.1634/theoncologist.9-1-58
Levitzki A EGF receptor as a therapeutic target Lung Cancer 2003 41 S9 S14 12867057 10.1016/S0169-5002(03)00134-X
Hidaki H Inagaki M Kawamoto S Sasaki Y Isoquinolinesulfonamides, novel and potent inhibitors of cyclic nucleotide dependent protein kinase and protein kinase C Biochemistry 1984 23 5036 5041 6238627 10.1021/bi00316a032
Muhsin M Graham J Kirkpatrick P Hot Drugs 2004: Cancer Nature Rev Drug Discov 2004 3 S6 S10 10.1038/nrd1406
Gough NR Adler EM Ray B Focus Issue: Targeting Signaling Pathways for Drug Discovery Science's STKE 2004 eg5 10.1126/stke.2252004eg5
Smith JK Mamoon NM Duhe RJ Emerging Roles of Targeted Small Molecule Protein-Tyrosine Kinase Inhibitors in Cancer Therapy Oncology Research 2004 14 175 225 14977353
Sebolt-Leopold JS MEK inhibitors: a therapeutic approach to targeting the Ras-MAP kinase pathway in tumors Curr Pharm Des 2004 10 1907 14 15180527 10.2174/1381612043384439
Garcia-Echeverria C Fabbro D Therapeutically targeted anticancer agents: inhibitors of receptor tyrosine kinases Mini Rev Med Chem 2004 4 273 283 15032674 10.2174/1389557043487349
Azam M Latek R Daley G Mechanisms of autoinhibition and STI-571/Imatinib resistance revealed by mutagenesis of BCR-ABL Cell 2003 112 831 843 12654249 10.1016/S0092-8674(03)00190-9
Pao W Miller V Politi K Riely G Somwar R Zakowski M Kris M Varmus H Acquired resistance of lung adenocarcinomas to gefitinib or elotinib is associated with a second mutation in the EGFR kinase domain PLoS Medicine 2005 2 225 234 10.1371/journal.pmed.0020225
Daub H Specht S Ullrich A Strategies to overcome resistance to targeted protein kinase inhibitors Nature Rev Drug Discov 2005 3 1001 1010 10.1038/nrd1579
Bain J McLauchlan H Elliott M Cohen P The specificities of protein kinase inhibitors: an update Biochem J 2003 371 199 204 12534346 10.1042/BJ20021535
Sykes J Some properties of a new epithelial cell line of human origin J Nat Cancer Inst 1970 66 107 112
Zlokarnik G Knapp TE Mere L Burres N Feng L Whitney M Roemer K Tsien RY Quantitation of transcription and clonal selection of single living cells with beta-lactamase as reporter Science 1998 279 84 88 9417030 10.1126/science.279.5347.84
Smith JK Mamoon NM Duhe RJ Emerging Roles of Targeted Small Molecule Protein-Tyrosine Kinase Inhibitors in Cancer Therapy Oncology Research 2004 14 175 225 14977353
Rodems SM Hamman BD Lin C Zhao J Shah S Heidary D Makings L Stack JH Pollok BA A FRET-based assay platform for ultra-high density drug screening of protein kinases and phosphatases Assay Drug Dev Technol 2002 1 9 19 15090152 10.1089/154065802761001266
Gazit A Yaish P Gilon C Levitzki A Tyrphostins I: synthesis and biological activity of protein tyrosine kinase inhibitors J Med Chem 1989 32 2344 52 2552117 10.1021/jm00130a020
Sebolt-Leopold JS Herrera R Targeting the mitogen-activated protein kinase cascade to treat cancer Nat Rev Cancer 2004 4 937 47 15573115 10.1038/nrc1503
Melnikova I Golden J Targeted protein kinases Nature Rev Drug Disc 2004 3 993 994 10.1038/nrd1600
Kim JA Targeted therapies for the treatment of cancer Am J Surg 2003 186 264 268 12946830 10.1016/S0002-9610(03)00212-5
Vlahovic G Crawford J Activation of tyrosine kinases in cancer The Oncologist 2003 8 531 538 14657531 10.1634/theoncologist.8-6-531
Kobayashi S Boggon TJ Dayaram T Janne PA Kocher O Meyerson M Johnson BE Eck MJ Tenen DG Halmos B EGFR mutation and resistance of non-small-cell lung cancer to gefitinib N Engl J Med 2005 352 786 92 15728811 10.1056/NEJMoa044238
Gumireddy K Baker SJ Cosenza SC John P Kang AD Robell KA Reddy MV Reddy EP A non-ATP-competitive inhibitor of BCR-ABL overrides imatinib resistance Proc Natl Acad Sci U S A 2005 102 1992 7 15677719 10.1073/pnas.0408283102
O'Hare T Walters DK Deininger MW Druker BJ AMN107: tightening the grip of imatinib Cancer Cell 2005 7 117 9 15710324 10.1016/j.ccr.2005.01.020
De Clercq E Emerging anti-HIV drugs Expert Opin Emerg Drugs 2005 10 241 73 15934866 10.1517/14728214.10.2.241
Study JJ HER1/EGFR tyrosine kinase inhibitors for the treatment of glioblastoma multiforme J Neurooncol 2005 74 77 86 16078112 10.1007/s11060-005-0603-7
Bell IM Stirdivant SM Ahern J Culberson JC Darke PL Dinsmore CJ Drakas RA Gallicchio SN Graham SL Heimbrook DC Hall DL Hau J Kett NR Kim AS Kornienko M Kuo LC Munshi SK Quigley AG Reid JC Trotter BW Waxman LH Williams TM Zartman CB Biochemical and structural characterization of a novel class of inhibitors of the type 1 insulin-like growth factor and insulin receptor kinases Biochemistry 2005 44 9430 40 15996097 10.1021/bi0500628
Poh TW Pervaiz S LY294002 and LY303511 sensitize tumor cells to drug-induced apoptosis via intracellular hydrogen peroxide production independent of the phosphoinositide 3-kinase-Akt pathway Cancer Res 2005 65 6264 74 16024628 10.1158/0008-5472.CAN-05-0152
Bianco R Troiani T Tortora G Ciardiello F Intrinsic and acquired resistance to EGFR inhibitors in human cancer therapy Endocr Relat Cancer 2005 12 S159 S171 16113092 10.1677/erc.1.00999
|
16202132
|
PMC1262698
|
CC BY
|
2021-01-04 16:03:03
|
no
|
BMC Cancer. 2005 Oct 3; 5:125
|
utf-8
|
BMC Cancer
| 2,005 |
10.1186/1471-2407-5-125
|
oa_comm
|
==== Front
BMC CancerBMC Cancer1471-2407BioMed Central London 1471-2407-5-1281620215010.1186/1471-2407-5-128Research ArticleAlteration of protein expression pattern of vascular endothelial growth factor (VEGF) from soluble to cell-associated isoform during tumourigenesis Cressey Ratchada [email protected] Onusa [email protected] Nirush [email protected] Usanee [email protected] Department of Associated Medical Science, Chiang Mai University, Chiang Mai, Thailand2 Department of Pathology, Chiang Mai University, Chiang Mai, Thailand3 Department of Biochemistry, Chiang Mai University, Chiang Mai, Thailand2005 4 10 2005 5 128 128 20 6 2005 4 10 2005 Copyright © 2005 Cressey et al; licensee BioMed Central Ltd.2005Cressey et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Vascular endothelial growth factor (VEGF) is a potent mitogen for endothelial cells, and its expression has been correlated with increased tumour angiogenesis. Although numerous publications dealing with the measurement of circulating VEGF for diagnostic and therapeutic monitoring have been published, the relationship between the production of tissue VEGF and its concentration in blood is still unclear. The aims of this study were to determine: 1) The expression pattern of VEGF isoforms at the protein level in colorectal and lung adenocarcinoma in comparison to the pattern in corresponding adjacent normal tissues 2) The relationship between the expression pattern of VEGF and total level of circulating VEGF in the blood to clarify whether the results of measuring circulating VEGF can be used to predict VEGF expression in tumour tissues.
Methods
Ninety-four tissue samples were obtained from patients, 76 colorectal tumour tissues and 18 lung tumour tissues. VEGF protein expression pattern and total circulating VEGF were examined using western blot and capture ELISA, respectively.
Results
Three major protein bands were predominately detected in tumour samples with an apparent molecular mass under reducing conditions of 18, 23 and 26 kDa. The 18 kDa VEGF protein was expressed equally in both normal and colorectal tumour tissues and predominately expressed in normal tissues of lung, whereas the 23 and 26 kDa protein was only detected at higher levels in tumour tissues. The 18, 23 and 26 kDa proteins are believed to represent the VEGF121, the VEGF165 and the VEGF189, respectively. There was a significant correlation of the expression of VEGF165 with a smaller tumour size maximum diameter <5 cm (p < 0.05), and there was a significant correlation of VEGF189 with advanced clinical stage of colorectal tumours. The measurement of total circulating VEGF in serum revealed that cancer patients significantly (p < 0.001) possessed a higher level of circulating VEGF (1081 ± 652 pg/ml in colorectal and 1,251 ± 568 pg/ml in lung) than a healthy volunteer group (543 ± 344 pg/ml). No correlation between the level of circulating VEGF and the pathologic features of tumours was observed.
Conclusion
Our findings indicate that the expression patterns of VEGF isoforms are altered during tumourigenesis as certain isoform overexpression in tumour tissues correlated with tumour progression indicating their important role in tumour development. However, measurement of VEGF in the circulation as a prognostic marker needs to be carefully evaluated as the cell-associated isoform (VEGF189), but not the soluble isoform (VEGF121 and VEGF165) appears to play important role in tumour progression.
==== Body
Background
VEGF plays a crucial role in tumour expansion by initiating permeabilization of blood vessels, by extravasation of plasma proteins, by invasion of stromal cells, and by causing the sprouting of new blood vessels that supply the tumour with oxygen and nutrients [1]. As a result of alternative splicing, 6 VEGF isoforms of 121, 145, 165, 183, 189 and 206 amino acids are produced from a single gene [2]. Due to differential incorporation of basic residues encoded by exon 6 and 7, VEGF isoforms differ in their heparin-binding properties, membrane association, and secretion [3]. VEGF121, which lacks the basic residues of both exons, does not bind heparin-containing cell surface proteoglycan [4], and is freely soluble. VEGF165 is also secreted. However, cationic residues in exon 7 enable VEGF165 to bind heparin, thus, some remains bound to the cell surface or to extracellular matrix. VEGF189 which retain both exons, has the highest affinity for heparin and therefore, remains tightly cell associated.
Detection of circulating VEGF has been investigated as a potential serum diagnostic marker for malignant disease and for inflammation [5]. Increased serum concentrations of free VEGF have been measured in various types of cancer, including brain, lung, gastrointestinal, hepatobiliary, renal, and ovarian cancers [6]. However, the relationship between the pattern of the production of VEGF protein isoforms in tumour tissues and their concentration in the circulation is still unclear.
A number of studies have shown that expression of certain VEGF transcripts are correlated with tumour progression. Increased mRNA expression of VEGF189 is correlated with poor prognosis in osteosarcoma [7] and non-small cell lung cancer [8,9], whereas expression of VEGF121 was correlated with lymph node metastasis in primary lung tumours [10]. Although increases of certain VEGF transcripts have been demonstrated to correlate with the progression of various tumours, the actual protein levels of the different VEGF isoforms and their significance during cellular transformation are unknown. Moreover, it has been suggested that elevated protein expression in tumour tissues was mediated by both enhanced transcription [11] and translation [12]. Thus, in order to understand the role of VEGF in tumour progression, it is important to investigate expression of different VEGF isoforms at the protein level during tumourigenesis. To our knowledge, no studies focusing on the VEGF isoform pattern at the protein level and their relationship with respect to total VEGF in the circulation have been reported.
Therefore, the aims of this study were to determine: 1) The protein expression pattern of VEGF isoforms in colorectal and lung tumours in comparison to the corresponding adjacent normal tissues in order to understand whether specific VEGF protein isoforms play an important role during tumourigenesis. 2) The relationship between the expression pattern of VEGF and the level of total circulating VEGF in the blood.
Methods
Selection of patients and sample
Between April 2002 and June 2004, samples were collected from cancer patients at Maharaj Nakorn Chiang Mai Hospital, which comprised 76 colorectal tumours (averaged age was 59 ± 15.2 (mean ± SD), 46 females and 30 males) and 18 non-small cell lung tumours (averaged age was 55 ± 14.6, 10 females and 8 males, 9 adenocarcinomas and 9 squamous cell carcinomas). In each case, adjacent normal tissue was collected. These specimens were immediately placed in vials, frozen in embedded medium for the preservation of cell integrity, and stored at -80°C until analyzed. Samples were graded by a pathologist according to the pathological features of the tumours, which included tumour size in maximal diameter, histological grading, lymph node metastasis, distant metastasis, and tumour staging (the AJCC TNM classification).
To avoid pre-analytical sample-to-sample variation due to blood collecting procedures, each blood sample was allowed to clot for at least 4 hrs before collecting serum as it has been reported that the release of VEGF during clotting period would have reached a plateau by this time [13]. Of 94 patients recruited in this study, serums were obtained from 56 cancer patients prior to the operation (38 from colorectal cancer patients, 18 from lung cancer patients). The age range of cancer patients was 58 ± 12.5 years and composed of 32 females and 24 males.
Serums were also collected from 47 healthy volunteers with no history of rheumatoid arthritis or recent pregnancy, trauma, surgery (within 1 month) or menstruation (within 1 week) using the same procedure as for the cancer patients so that a comparison could be made. The age range of healthy volunteers was 51 ± 10.9 (mean ± SD) years, composed of 20 female and 27 males. All serums were stored at -70°C until analyzed. The study was approved by the ethical committee of the Faculty of Medicine, Chiang Mai University (document number 56/2545).
Western blotting
Western blotting was performed to evaluate the expression of VEGF in each tissue. Frozen tissues were thawed, cut into small pieces and homogenized in SDS lysis buffer (0.5 M Tris-HCl pH 6.8, 2% SDS (w/v) and 10% glycerol (v/v)) containing a protease inhibitors cocktail (104 mM AEBSF, 0.08 mM aprotinin, 2.2 mM leupeptin, 3.6 mM bestatin, 1.5 mM pepstatin A, 1.4 mM E-64; Sigma, U.S.A). The tissue homogenate was then centrifuged at 10,000 g for 15 minutes at 4°C, after which the supernatant was removed and the protein concentration of the supernatant was estimated using the BCA protein assay kit (PIERCE, U.S.A). Twenty-five micrograms of protein from the tumour tissue and normal tissue from each patient was resolved on a 10% SDS polyacrylamide gel under reducing conditions and electrotransferred onto a nitrocellulose membrane (Biorad, U.S.A). After blocking with 5% non-fat milk in TBS containing 0.05% Tween-20 (TBS-Tween) for 1 hour, the membrane was incubated with anti-VEGF antibodies (Santa Cruz Biotechnology, Inc., USA, Cat. no. SC-152, dilution 1:1000) for 1 hour. After washing with TBS-Tween, the membrane was incubated for 1 hour at RT with horseradish peroxidase-conjugated goat anti-mouse IgG (Dako, U.S.A). After washing with TBS-Tween, immunoreactive protein was visualized with a chemiluminescence-based procedure using the ECL Plus detection kit according to the manufacturer's protocol (Amersham, U.S.A). In order to examine the equality of protein loaded, the amount of total protein loaded into each lane was examined by staining with coomassie blue.
Measurement of total VEGF in serum
For the detection of circulating VEGF in serum, enzyme-linked immunosorbent assay (ELISA) was performed using two different anti-VEGF antibodies purchased from R&D system, USA. Briefly, capture antibodies specific for VEGF (R&D System, cat no. AF293 at concentration 200 ng/ml) was immobilized onto-96-well microtiter plates. Unbound antibody was removed by washing the plate and a blocking reagent was added. Following a wash, recombinant VEGF protein standard (VEGF165) diluted in PBS containing 5% BSA to various concentrations (75–2,500 pg/ml), unknown serum and control serum were then incubated with the solid phase antibodies, which capture VEGF. After washing away unbound molecules, a detection antibody specific for VEGF (R&D System, cat no. MAB293 at concentration 500 ng/ml) was added. After incubation and washing, HRP-conjugated anti-mouse immunoglobulin was added. The plate was washed and a TMB substrate solution (Zymed, USA) was added. After 20 minutes, the color development was stopped and the intensity of color was measured using a microtiter plate reader (450 nm). The color developed in proportion to the amount of bound VEGF. When we measured 20 serum samples twice in two separated assays, the inter-assay variation ranged between 5–10% within the same concentration range. The average recovery of the added recombinant VEGF165 ranged between 85–115%, indicating an acceptable level of specificity of the assay.
Statistical analysis
Total VEGF levels are expressed as mean ± standard deviation. Differences in the circulating VEGF level of two independent groups were evaluated using the Mann-Whitney test. Correlation between VEGF isoform expression and the pathological features were evaluated using chi-square test. All the statistical evaluations were performed by using the SPSS for Window version 10.0 (SPSS, Inc., Chicago, IL, USA).
Results
Pattern of VEGF protein expression in normal and tumour tissues of colon and lung
The expression pattern of VEGF isoforms in tumour tissues in comparison to normal tissues determined by western blot analysis are shown in Figure 1. Three major protein bands were predominately detected in colorectal and lung tumour samples with an apparent molecular mass under reducing conditions of 18 kDa, 23 kDa, and 26 kDa. The 18 kDa VEGF was equally expressed in both normal and tumour tissues of colorectal and predominately expressed in normal tissue of lung, whereas the 23 and 26 kDa were only detected at higher levels in tumour tissues of both organs. Expression of the 23 kDa VEGF isoform was observed in 55.3% (42 of 76 patients) of colorectal tumours and 88.9% (16 of 18 patients) of lung tumour tissues. Whereas, expression of 26 kDa VEGF isoform was detected in 69.7% (53 of 76 patients) and 88.9% (16 of 18 patients) of colorectal and lung tumour tissues, respectively.
Figure 1 Representative Western blots showing protein expression pattern of VEGF isoform in (a) colorectal and (b) non-small cell lung tumour tissues and their corresponding adjacent normal tissues (T, tumor tissues; N, normal tissues).
Protein expression patterns of VEGF isoforms in tumour tissues of colon and lung in relation to pathological features
The two types of cancer were classified according to the pathologic features, which included tumour sizes in maximum diameter, depth of invasion, lymph node metastasis, distant metastasis and histological differentiation. Expression of the VEGF isoforms in relation to the pathological features of colorectal tumours and lung tumours are summarized in Table 1.
Table 1 Summary of relationship between VEGF isoform expression and pathologic features in colorectal and non-small cell lung cancers (p < 0.05 was considered significant)
Pathologic features VEGF isoform (kDa)
VEGF 23 kDa p valuea VEGF 26 kDa p valuea
Colorectal cancer (total 76 cases) 42 (55.3%) 53 (69.7%)
Gender
Female (46 cases) 28 (60.8%) 0.224 32 (69.5%) 0.968
Male (30 cases) 14 (46.7%) 21 (70.0%)
Tumour size
≤ 5 cm (45 cases) 30 (66.7%)b < 0.05 30 (66.7%) 0.483
> 5 cm (31 cases) 12 (38.7%) 23 (74.1%)
Histological differentiation
Well (42 cases) 22 (52.4%) 0.574 28 (66.7%) 0.517
Moderate or Poor (34 cases) 20 (58.8%) 25 (73.5%)
Tumour stage grouping
Early stage (I or II) (34 cases) 23 (67.6%) 0.09 19 (55.9%) < 0.01
Late stage (III or IV) (42 cases) 19 (45.2%) 34 (80.9%)
Metastasis
No (49 cases) 31 (63.3%) 0.059 29 (59.2%) < 0.01
Yes (27 cases) 11 (40.7%) 24 (88.9%)
Lung cancer (total 18 cases) 16 (88.9%) 16 (88.9%)
Gender
Female (10 cases) 8 (80.0%) 0.477 8 (80.0%) 0.477
Male (8 cases) 8 (100.0%) 8(100%)
Tumour size
≤ 5 cm (7 cases) 7 (100.0%) 0.231 7 (100.0%) 0.231
> 5 cm (11 cases) 9 (81.8%) 9 (81.8%)
Histological differentiation
Well (5 cases) 4 (80.0%) 0.490 4 (80.0%) 0.490
Moderate or Poor (13 cases) 12 (92.3%) 12 (93.3%)
Tumour stage grouping
Early stage (I or II) (2 cases) 1 (50.0%) 0.210 0 (0.0%) < 0.01
Late stage (III or IV) (16 cases) 15 (93.8%) 16 (100%)
Metastasis
No (7 cases) 6 (85.7%) 1.00 5 (71.4%) 0.137
Yes (11 cases) 10 (90.9%) 11 (100.9%)
a chi-square test, b percentage in relation to total number of cases in each pathological features
No significant difference between gender of the VEGF expression pattern was observed in both types of cancer. In colorectal cancer, it was found that expression of VEGF isoform with molecular weight 23 kDa was significantly correlated with a smaller tumour size (maximum diameter < 5 cm, p < 0.05), whereas the 26 kDa VEGF isoform was significantly correlated with advanced clinical stage and metastasis of the tumour (p < 0.01). Expression of the 26 kDa VEGF isoform was also significantly correlated with advanced clinical stage of non-small cell lung cancer (p < 0.001). Sixteen (out of 18) lung tumour tissues which overexpressed 26 kDa VEGF were late stage tumours (Table 1). No significant difference of the expression pattern of VEGF between different histology type (adenocarcinoma and squamous cell carcinoma) was observed (data not shown).
Levels of circulating VEGF in cancer patients compared to healthy volunteers and their relationship to pathological features
Preoperative serum was collected from 56 cancer patients; these included 38 patients with colorectal cancer and 18 with lung cancer. Serum from 47 healthy volunteers was also collected for comparison. The result showed that cancer patients possessed significantly higher level of circulating VEGF than those in healthy volunteers (Figure 2). While level of total circulating VEGF in healthy volunteer was only 543 ± 344 pg/ml, it was 1081 ± 652 pg/ml and 1,251 ± 568 pg/ml in patients with colorectal and lung cancer, respectively (Table 2). No significant relationship between the level of circulating VEGF and the pathological features was observed (Table 3). Gender also did not show any impact on circulating level of VEGF (Table 3). In addition, none of the VEGF isoforms showed a significant relationship with the serum level of VEGF. Although colorectal cancer patients with overexpression of VEGF 23 kDa, which is believed to be VEGF165 (one of a secretable form of VEGF) in tumour tissues, possessed higher levels of circulating VEGF in serum (1190 ± 752 pg/ml) than those possessing undetectable level of VEGF165 (875 ± 330 pg/ml), it was not statistically significant (p = 0.207, Mann-Whitney test).
Figure 2 Serum level of circulating VEGF of colorectal and non-small cell lung cancer patients in comparison to healthy volunteers.
Table 2 Serum level of VEGF in colorectal and non-small cell lung cancer patients in comparison to healthy volunteers (p < 0.05 was considered significant)
Type of sample No. of cases VEGF concentration (pg/ml) p valuea
Colorectal cancer patients 38 1,081 ± 652b <0.001
Lung cancer patients 18 1,251 ± 568 <0.001
Healthy volunteers 47 543 ± 344
aMann-Whitney test, bmean ± SD
Table 3 Serum level of VEGF in relation to the clinicopathologic features of colorectal and non-small cell lung cancers (p < 0.05 was considered significant)
Pathological features No. of cases VEGF concentration (pg/ml) p valuea
Colorectal tumours 38 1,081 ± 652b
Gender
Female 22 987 ± 470 0.433
Male 16 1212 ± 841
Tumour size
≤ 5 cm 23 1134 ± 727 0.411
> 5 cm 15 1002 ± 531
Tumour stage grouping
Early stage 23 1166 ± 799 0.777
Late stage 15 954 ± 303
Metastasis
No 27 1132 ± 749 0.910
Yes 11 961 ± 304
VEGF 23 kDa
Positive 25 1190 ± 752 0.207
Negative 13 875 ± 330
VEGF 26 kDa
Positive 26 1125 ± 720 0.742
Negative 12 989 ± 486
Lung tumours 18 1,251 ± 568
Gender
Female 10 1297 ± 509 0.374
Male 8 1194 ± 666
Tumour size
≤ 5 cm 7 1304 ± 661 0.762
> 5 cm 11 1217 ± 532
Tumour stage grouping
Early stage 2 716 ± 666 0.206
Late stage 16 1318 ± 541
Metastasis
Yes 8 1125 ± 655 0.556
No 10 1331 ± 522
VEGF 23 kDa
Positive 16 1190 ± 574 0.160
Negative 2 1741 ± 9.8
VEGF 26 kDa
Positive 16 1224 ± 590 0.482
Negative 2 1468 ± 395
a Mann-Whitney test, bmean ± SD
Discussion
It has become clear that the growth of solid tumours is dependent on the process of angiogenesis and that VEGF is a central positive regulator of this process. Most VEGF-producing cells appear preferentially to express VEGF121, VEGF165 and VEGF189. In this study, we investigated the expression pattern of the VEGF protein isoform in colorectal tumour tissues and in lung tumour tissues and compared them with the expression pattern of normal tissues from each organ, respectively. Three major protein bands with molecular weight 18, 23 and 26 kDa were predominately detected. The 23-kDa protein band is believed to be the VEGF165 as this band was at the same position as the human recombinant VEGF165 protein standard (R&D System, USA) used in this study. Expression of VEGF145 and VEGF206 is comparatively rare seemingly restricted to cells of placental origin [14,15]. Therefore, protein bands with molecular weight of 18 and 26 kDa are assumed to be VEGF121 [16,17] and VEGF189, respectively.
In colorectal tumours, it was found that VEGF121 was expressed equally in both tumour and normal tissues, whereas the VEGF165 and VEGF189 were only detected at higher level in tumour tissues. However in lung tumour, VEGF121 appeared to be predominately expressed in normal tissues, whereas VEGF165 and VEGF189 were predominately expressed in tumours tissues. Protein expression of VEGF165 correlated significantly with a smaller tumour size, whereas VEGF189 correlated significantly with advanced clinical stage and metastasis of the tumours. Although only 18 lung tumours were investigated in this study, the 26-kDa VEGF isoform was also overexpressed significantly in advanced stage of the tumour (Table 1).
Although the regulation of VEGF expression is becoming well understood, its mode of action, particularly the regulation of expression and distribution of the three primary isoform (VEGF121, VEGF165 and VEGF189), remains unclear. It has been demonstrated that overexpression of smaller isoforms resulted in hemorrhagic events, but the expression of VEGF189 resulted in increased vessel density [18]. In this study we found that overexpression of VEGF189 protein isoform, but not VEGF121 or VEGF165, was associated with advanced tumour stage. Our data is consistent with the previous reports where expression of VEGF189 transcripts was correlated with poor prognosis in non-small cell lung, osteosarcoma, renal, colorectal and esophageal cancer [7-9,19,20]. It was also suggested that up-regulation of VEGF189 might result in increased angiogenesis, tumour growth and metastasis in a colon cancer cell line [21]. Moreover, VEGF189 has been demonstrated to be a potent permeability factor in vivo [22], supporting the role of this isoform in the control of angiogenesis.
VEGF165 has also been demonstrated to play an important role in tumourigenesis. When different isoforms of VEGF were transfected into the VEGF-null cells in isolation and the transfected cells were implanted into nude mice, it was found that VEGF165 was the most prominent isoform that can fully rescue expansion of the angiogenesis-deficient tumour, while VEGF121 and VEGF189 only partially or failed completely to rescue tumour growth, respectively [23]. However, these authors suggested that VEGF isoforms work in a coordinated fashion to recruit and expand tumour vasculature. In our study we found that although VEGF165 was predominately expressed in colorectal tumour tissues, its expression was significantly correlated with smaller tumour size (maximum diameter less than 5 cm.). Although expression of VEGF121 mRNA has been previously reported to be correlated with lymph node metastasis [10] of primary lung cancer and the invasiveness of bladder cancer [24], in our study we found that level of the 18 kDa VEGF protein, which believed to be VEGF121 [16], was equally expressed in both normal and tumour tissues of colorectal, and predominately expressed in normal tissues of the lung.
Detection of VEGF has long been known as a potential serum diagnostic marker for malignant diseases. Increased serum VEGF concentrations have been measured in various types of cancer, including, brain, lung, gastrointestinal, hepatobiliary, renal and ovarian cancer [25]. However, the relationship between the pattern of the production of VEGF protein isoforms in tumours and its concentration in the circulation is still unclear. In this study, we determined the expression pattern of VEGF isoforms in tumour tissues in relation to the level of total VEGF in a patient's serum. The comparison of the VEGF level in serum of cancer patients with that of normal volunteers revealed that cancer patients possessed significantly (p < 0.001) higher levels of VEGF in serum. However, some normal volunteers also possessed quite a high level of VEGF, which may due to the possibility that normal tissues, like lung tissue (Figure 1) can also produce VEGF121 that is secretable into the circulation. In addition, no significant relationship between level of circulating VEGF and pathologic features was observed.
Conclusion
Our findings indicate that the expression patterns of VEGF isoforms are altered during tumourigenesis as certain isoform overexpression in tumour tissues correlated with tumour progression indicating their important role in tumour development. However, measurement of circulating VEGF in serum may have limited use as a tumour marker. This may be due to the following reasons: 1) The VEGF isoform that appeared to be significantly correlated with tumour progression is VEGF189, which is the cell-associated isoform, is not soluble. 2) Some normal tissues, i.e. lung (as shown in Figure 1), expressed high-level VEGF isoforms (VEGF121) that secreted into the circulation. 3) Expression of some secretable VEGF isoforms (VEGF165) was negatively correlated with the progression of tumour size, thus its level may not positively indicate the stage of the tumour. 4) As has been previously reported, other physiologic and pathologic condition, i.e., pregnancy, RA and cardiovascular diseases can also cause the induction the circulating level of VEGF [26,27].
Competing interests
The author(s) declare that they have no competing interest.
Authors' contributions
RC designed and sought funding for the study, initiated coordination, performed statistical analysis and drafted the manuscript. OW recruited healthy volunteers and performed the ELISA and western blot analysis. NL recruited cancer patients and carried out the pathological feature examination. UV participated in design of the study and coordination. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This work was sponsored by the Thailand Research Fund (Grant no. MRG4580013) and partly supported by the Thailand National Center for Genetic Engineering and Biotechnology (BIOTEC), National Science and Technology Development Agency (NSTDA).
==== Refs
Ruegg C Meuwly JY Driscoll R Werffeli P Zaman K Stupp R The quest for surrogate markers of angiogenesis: a paradigm for translational research in tumor angiogenesis and anti-angiogenesis trials Curr Mol Med 2003 3 673 691 14682490 10.2174/1566524033479410
Ferrara N Houck KA Jakeman LB Winer J Leung DW The vascular endothelial growth factor family of polypeptides J Cell Biochem 1991 47 211 218 1791185 10.1002/jcb.240470305
Park JE Keller GA Ferrara N The vascular endothelial growth factor (VEGF) isoforms: differential deposition into the subepithelial extracellular matrix and bioactivity of extracellular matrix-bound VEGF Mol Biol Cell 1993 4 1317 1326 8167412
Cohen T Gitay-Goren H Sharon R Shibuya M Halaban R Levi BZ Neufeld G VEGF121, a vascular endothelial growth factor (VEGF) isoform lacking heparin binding ability, requires cell-surface heparan sulfates for efficient binding to the VEGF receptors of human melanoma cells J Biol Chem 1995 270 11322 11326 7744769 10.1074/jbc.270.19.11322
Bottomley MJ Webb NJ Watson CJ Holt L Bukhari M Denton J al Placenta growth factor (PIGF) induces vascular endothelial growth factor (VEGF) secretion from mononuclear cells and is co-expressed with VEGF in synovial fluid. Clin Exp Immunol 2000 119 182 188 10606981 10.1046/j.1365-2249.2000.01097.x
Jelkmann W Pitfalls in the measurement of circulating vascular endothelial growth factor Clin Chem 2001 47 617 623 11274009
Lee YH Tokunaga T Oshika Y Suto R Yanagisawa K Tomisawa M Fukuda H Nakano H Abe S Tateishi A Kijima H Yamazaki H Tamaoki N Ueyama Y Nakamura M Cell-retained isoforms of vascular endothelial growth factor (VEGF) are correlated with poor prognosis in osteosarcoma Eur J Cancer 1999 35 1089 1093 10533453 10.1016/S0959-8049(99)00073-8
Oshika Y Nakamura M Tokunaga T Ozeki Y Fukushima Y Hatanaka H Abe Y Yamazaki H Kijima H Tamaoki N Ueyama Y Expression of cell-associated isoform of vascular endothelial growth factor 189 and its prognostic relevance in non-small cell lung cancer Int J Oncol 1998 12 541 544 9472090
Yuan A Yu CJ Kuo SH Chen WJ Lin FY Luh KT Yang PC Lee YC Vascular endothelial growth factor 189 mRNA isoform expression specifically correlates with tumor angiogenesis, patient survival, and postoperative relapse in non-small-cell lung cancer J Clin Oncol 2001 19 432 441 11208836
Ohta Y Watanabe Y Murakami S Oda M Hayashi Y Nonomura A Endo Y Sasaki T Vascular endothelial growth factor and lymph node metastasis in primary lung cancer Br J Cancer 1997 76 1041 1045 9376264
Shweiki D Itin A Soffer D Keshet E Vascular endothelial growth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis Nature 1992 359 843 845 1279431 10.1038/359843a0
Kelvis CH De Benedetti A Payne DK Coe LL Laroux FX Alexandra JS Translational regulation of vascular permeability factor by eukaryotic initiation factor 4E: implication for tumor angiogenesis Int J Cancer 1996 65 785 790 8631593 10.1002/(SICI)1097-0215(19960315)65:6<785::AID-IJC14>3.0.CO;2-3
Dittadi R Meo S Fabris F Gasparini G Contri D Medici M Gion M Validation of blood collection procedures for the determination of circulating vascular endothelial growth factor (VEGF) in different blood compartments Int J Biol Markers 2001 16 87 96 11471901
Anthony FW Wheeler T Elcock CI Pickett M Thomas EJ Short report: identification of a specific pattern of vascular endothelial growth factor mRNA expression in human placenta and cultured fibroblasts Placenta 1994 15 557 561 7997455
Cheung N Singh M Ebaugh MJ Brace RA Vascular endothelial growth factor gene expression in ovine placenta and fetal membrane Am J Obstet Gynecol 1995 173 753 759 7573238 10.1016/0002-9378(95)90335-6
Mohanraj D Olson T Ramakrishnan S A novel method to purify recombinant vascular endothelial growth factor (VEGF121) expressed in yeast Biochem Biophys Res Commun 1995 215 750 756 7488018 10.1006/bbrc.1995.2527
Nishimura S Maeno N Matsuo K Nakajima T Kitajima I Saito H Maruyama I Human lactiferous mammary gland cells produce vascular endothelial growth factor (VEGF) and express the VEGF receptors, Flt-1 AND KDR/Flk-1 Cytokine 2002 18 191 198 12126641 10.1006/cyto.2002.1032
Cheng SY Nagane M Huang HS Cavenee WK Intracerebral tumor-associated hemorrhage caused by overexpression of the vascular endothelial growth factor isoforms VEGF121 and VEGF165 but not VEGF189 Proc Natl Acad Sci U S A 1997 94 12081 12087 9342366 10.1073/pnas.94.22.12081
Tokunaga T Oshika Y Abe Y Ozeki Y Sadahiro S Kijima H Tsuchida T Yamazaki H Ueyama Y Tamaoki N Nakamura M Vascular endothelial growth factor (VEGF) mRNA isoform expression pattern is correlated with liver metastasis and poor prognosis in colon cancer Br J Cancer 1998 77 998 1002 9528847
Tomisawa M Tokunaga T Oshika Y Tsuchida T Fukushima Y Sato H Kijima H Yamazaki H Ueyama Y Tamaoki N Nakamura M Expression pattern of vascular endothelial growth factor isoform is closely correlated with tumour stage and vascularisation in renal cell carcinoma Eur J Cancer 1999 35 133 137 10211101 10.1016/S0959-8049(98)00278-0
Koura AN Liu W Kitadai Y Singh RK Radinsky R Ellis LM Regulation of vascular endothelial growth factor expression in human colon carcinoma cells by cell density Cancer Res 1996 56 3891 3894 8752153
Ancelin M Buteau-Lozano H Meduri G Osborne-Pellegrin M Sordello S Plouet J Perrot-Applanat M A dynamic shift of VEGF isoforms with a transient and selective progesterone-induced expression of VEGF189 regulates angiogenesis and vascular permeability in human uterus Proc Natl Acad Sci U S A 2002 99 6023 6028 11972026 10.1073/pnas.082110999
Grunstein J Masbad JJ Hickey R Giordano F Johnson RS Isoforms of vascular endothelial growth factor act in a coordinate fashion To recruit and expand tumor vasculature Mol Cell Biol 2000 20 7282 7291 10982845 10.1128/MCB.20.19.7282-7291.2000
Li NC Kanda K Fukumori T Inoue Y Nishitani M Kanayama H Kagawa S Expression of vascular endothelial growth factor isoforms and platelet-derived endothelial cell growth factor in bladder cancer Urologic Oncology 2000 6 10 15 11113367
Kondo S Asano M Matsuo K Ohmori I Suzuki H Vascular endothelial growth factor/vascular permeability factor is detectable in the sera of tumor-bearing mice and cancer patients Biochim Biophys Acta 1994 1221 211 214 8148401 10.1016/0167-4889(94)90016-7
Afuwape AO Kiriakidis S Paleolog EM The role of the angiogenic molecule VEGF in the pathogenesis of rheumatoid arthritis Histol Histopathol 2002 17 961 972 12168808
Zygmunt M Herr F Munstedt K Lang U Liang OD Angiogenesis and vasculogenesis in pregnancy Eur J Obstet Gynecol Reprod Biol 2003 110 Suppl 1 S10 8 12965086 10.1016/S0301-2115(03)00168-4
|
16202150
|
PMC1262699
|
CC BY
|
2021-01-04 16:39:12
|
no
|
BMC Cancer. 2005 Oct 4; 5:128
|
utf-8
|
BMC Cancer
| 2,005 |
10.1186/1471-2407-5-128
|
oa_comm
|
==== Front
BMC Cardiovasc DisordBMC Cardiovascular Disorders1471-2261BioMed Central London 1471-2261-5-291618536110.1186/1471-2261-5-29Research ArticleAccuracy of popular automatic QT Interval algorithms assessed by a 'Gold Standard' and comparison with a Novel method: computer simulation study Hunt Anthony Charles [email protected] PSI HeartSignals Ltd, Institute of Medical Technology, Glasgow Technology Park, PO Box 7043, Glasgow G44 9AB. UK2005 26 9 2005 5 29 29 12 5 2005 26 9 2005 Copyright © 2005 Hunt; licensee BioMed Central Ltd.2005Hunt; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Accurate measurement of the QT interval is very important from a clinical and pharmaceutical drug safety screening perspective. Expert manual measurement is both imprecise and imperfectly reproducible, yet it is used as the reference standard to assess the accuracy of current automatic computer algorithms, which thus produce reproducible but incorrect measurements of the QT interval. There is a scientific imperative to evaluate the most commonly used algorithms with an accurate and objective 'gold standard' and investigate novel automatic algorithms if the commonly used algorithms are found to be deficient.
Methods
This study uses a validated computer simulation of 8 different noise contaminated ECG waveforms (with known QT intervals of 461 and 495 ms), generated from a cell array using Luo-Rudy membrane kinetics and the Crank-Nicholson method, as a reference standard to assess the accuracy of commonly used QT measurement algorithms. Each ECG contaminated with 39 mixtures of noise at 3 levels of intensity was first filtered then subjected to three threshold methods (T1, T2, T3), two T wave slope methods (S1, S2) and a Novel method. The reproducibility and accuracy of each algorithm was compared for each ECG.
Results
The coefficient of variation for methods T1, T2, T3, S1, S2 and Novel were 0.36, 0.23, 1.9, 0.93, 0.92 and 0.62 respectively. For ECGs of real QT interval 461 ms the methods T1, T2, T3, S1, S2 and Novel calculated the mean QT intervals(standard deviations) to be 379.4(1.29), 368.5(0.8), 401.3(8.4), 358.9(4.8), 381.5(4.6) and 464(4.9) ms respectively. For ECGs of real QT interval 495 ms the methods T1, T2, T3, S1, S2 and Novel calculated the mean QT intervals(standard deviations) to be 396.9(1.7), 387.2(0.97), 424.9(8.7), 386.7(2.2), 396.8(2.8) and 493(0.97) ms respectively. These results showed significant differences between means at >95% confidence level. Shifting ECG baselines caused large errors of QT interval with T1 and T2 but no error with Novel.
Conclusion
The algorithms T2, T1 and Novel gave low coefficients of variation for QT measurement. The Novel technique gave the most accurate measurement of QT interval, T3 (a differential threshold method) was the next most accurate by a large margin. The objective and accurate 'gold standard' presented in this paper may be useful to assess new QT measurement algorithms. The Novel algorithm may prove to be more accurate and reliable method to measure the QT interval.
==== Body
Background
The QT interval is measured as the time interval between the onset of the QRS complex and the end of the T wave, the end of the T wave being the time at which repolarisation is completed and the T wave voltage amplitude returns to the baseline [1]. The QT interval is thus a measure of the duration of the ventricular depolarisation and repolarisation. Inaccuracies occur in the measurement of the QT interval due to the low frequency content of the T wave offset, which has a low signal to noise ratio. Also the presence of a U wave merging with the end of the T wave lead to inaccuracies [2].
Accurate measurement of the QT interval is very important from clinical and pharmaceutical drug safety screening perspective, as prolongation of repolarisation, manifested by prolongation of this interval, increases susceptibility to potentially fatal torsade de pointes ventricular arrhythmia [3,4]. A statistically significant increase in the mean QT interval (corrected for heart rate) as small as 6 milliseconds between baseline and maximal drug effect may be important as a signal of repolarisation abnormality [5]. Expert manual measurement of this interval is both imprecise and poorly reproducible, inter-operator differences of up to 28 milliseconds have been reported [6]. Automatic QT interval measurements have been shown to be more stable and reproducible than manual measurement [7]. It is not surprising that manual assessment of QT dispersion has also been shown to have very poor reproducibility [8], QT dispersion being derived from the difference between maximum and minimum QT interval in a 12 lead ECG. Yet manual measurement continues to be used as a meter to assess the accuracy of much needed reliable automatic computer algorithms to measure this interval reproducibly. For a given ECG, the very fact that the manual QT interval measurement presents a varying, dynamic reference standard means that manual measurements are imprecise and non-reproducible. The degree of manual measurement imprecision is currently unknown since there has not been an accurate reproducible reference standard with which to make an assessment.
Currently available automatic computer algorithms which measure the QT interval on a given ECG will have a reasonably good reproducibility but the QT interval measured may be reproducibily incorrect. Earlier efforts by Willems et al to evaluate the performance of ECG computer measurement programs and provide a common standard were based on the stability of results or reproducibilty [9]. In the report by Willems et al, the median result of measurements made by 5 Cardiologist referees was used to assess the median results of combined computer programme measurements. The referees determined the end of the T wave a mean 5–15 ms later than the combined computer algorithms but because a subjective manual reference system was used it is unknown as to whether the combined referee of combined computer programs was more accurate in determining the real T wave offset.
Presented in this research is an objective reference standard against which the accuracy and reproducibility of commonly used QT measurement algorithms can be assessed. This suggested new 'gold standard' uses a validated computer simulated electrocardiographic waveform [10] generated from a cell matrix using validated Luo-Rudy membrane kinetics [11]. The simulated ECG allows the exact timing of the cessation of repolarisation or end-point of the T wave to the nearest 0.005 milliseconds. ECGs will be constructed to simulate the varying intramyocardial conductivity and myocardial volume effects, in addition to extramyocardial conductance effects. Superadded to the simulated ECGs will be mixtures of three simulated different types of noise at different intensities [12,13]. Six different algorithms will be used to measure T wave offset [14], including a Novel algorithm. All algorithms will be statistically compared for reproducibility and accuracy to determine the underlying real QT interval for each ECG.
Methods
Initially four ECGs were constructed to simulate lumped variations in the electrophysiological properties of myocardial type and myocardial fibre orientation. The four ECGs constructed represented all the permutations of a high myocardial resistivity (100,000 Ohm cm), low resistivity (10,000 Ohm cm), time-dependent potassium channel conductance of 584 mS/cm2 and time dependent potassium channel conductance 262 mS/cm2. Variations in the ratio of density of the rapid and delayed potassium rectifier channels exist within different strata of the myocardium and a change in the ratio of these channels will therefore produce a different total conductance [15]. Variation in fibre orientation will produce a variation in resistivity.
Each of the four models of myocardial fibres with combinations of different resistivities and conductances consisted of 100 individually calculated cardiac cell membrane potentials, interconnected via resistors which represent gap junctions. For each myocardial fibre model the membrane ion kinetics were computed at an extracellular potassium concentration of 5.4 mmols using the meticulous Luo-Rudy physiological model which has been validated by experimental data [11]. The current propagation in one direction along each of the homogenous myocardial fibre types was evaluated by solving numerically the parabolic partial differential equation shown below (equation 1) for the transmembrane potential Vm, using the Crank-Nicholson implicit method [16], as was used in the insightful computer model by Virag et al [10]. The Crank-Nicholson implicit method has the advantage of stability irrespective of the time increments used in the iterations and provides improved accuracy at the expense of requiring the solution of a set of simultaneous equations at each time step.
Equation 1.
This equations describes current flow only intracellularly and the extracellular current is grounded [17] Cm is the membrane capacitance (1 μF/cm2), Sv is the membrane surface area to intracellular volume ratio (0.24 μm-1), Iion is the sum of the fast sodium, slow inward calcium, time dependent potassium, time independent potassium, plateau potassium and background membrane ionic currents in (μAmps/cm2), Istim is the stimulus current in (μAmps/cm2) and px is the resistivity in the x direction (10,000 and 100,000 Ohm cm). Equation 1 is discretised to equation 2 shown below.
Equation 2.
Vmi,t+Δt - Vmi,t + (Iioni,t - Istimi,t)Δt/Cmi = r/2(Vmi-1,t+Δt - 2 Vmi,t+Δt + Vmi+1,t+Δt + Vmi-1,t - 2 Vmi,t + Vmi+1,t)
In equation 2, Δt is the iteration time step of 0.005 milliseconds and Δx is the space discretisation of the grid (100 μm). The partial derivative of Vm being written as the finite difference (Vmi,t+Δt - Vmi,t)/Δt.
The symbol r is the coupling factor = Δt/Cmi Sv (Δx)2px. The superscript i is the cell position in a grid of 100 cells. The superscript t is the time elapsed from onset of the simulation, Δt is the iteration time increment of 0.005 milliseconds.
The computational method is iterative in time following two steps, firstly the total ionic current is computed for each cell for each Vmi,t using the Luo- Rudy membrane kinetics. Calculation of the ionic current for each cell for each iteration is a major time limiting step, this process was accelerated at the time Vmi,t+Δt reached each individual cell threshold potential of greater than -60 mV, by the use of a pre-calculated action potential (customised to the electrophysiological parameters) lookup table which gives the ionic currents for each corresponding voltage at each 0.005 millisecond time step. The set of simultaneous equations produced by each iteration of Equation 2 of for each cell was solved by the implicit method using tridiagonal matrices. Von Neumann boundary conditions are adopted which means no current flows out of either end of the cell array.
The ECG voltage potentials for the four electrophysiological conditions of varying resistivities and time dependent potassium current conductances was calculated by representing each pair of adjacent cell elements within the array as constituting an electric dipole of length Δx, the current density depending on the difference of potential (Vmi,t - Vmi+1,t) between two adjacent cells and the resistivity of their gap junction. Given a constant electrode orientation along the x axis, the instantaneous potential recorded from a remote electrode is proportional to the sum of the dipole moments or dipole potential differences over the cell array distance calculated for each iteration time [18].
Each of the four different ECGs generated along a single vector can be used to simulate orthogonal XYZ vector ECGs through a volume of myocardium with inhomogeneous electrophysiological characteristics within each orthogonal plane of myocardium. By the addition of each XYZ potential voltage generated every 0.005 milliseconds for each of the XYZ vector ECGs it is possible to derive a resultant vector ECG for the volume of myocardium simulated. Using three combinations of ECGs derived under different electrophysiological conditions and positive or negative orientations it is possible to produce complex resultant ECG T waves. Two predominantly positive T waves (Res1 and Res2) and two biphasic T waves (Res3 and Res4) were simulated.
Extracellular unipolar potentials Φ generated by a myocardial fibre orientated along a vector axis x in an extensive medium of extracellular conductivity σe are computed from the transmembrane potential Vm using equation 3 [19].
Equation 3 Φ = a2 σi (∫ (-grad(Vm))(grad(1/r))dx)/4 σe
Where a is the cross sectional area of the fibre, σi is the intracellular conductivity, r is the distance from the point source to the field point and grad is the grad vector operator. It can be seen that none of the variables and operators in equation 3 contain the variable of time. Therefore a change in magnitude of σe or other dimensions of length for a given σi and Vm will have a linear effect and directly proportional effect on the amplitude of Φ without any effect on the time duration. This has been experimentally borne out by Mirvis et al [18].
Therefore to simulate the effects of a doubling of σe for the four resultant ECGs generated the ECG voltage potential was doubled every 0.005 milliseconds of the ECG complex duration. All simulated ECGs voltages below 0.00001 millivolts were reduced to zero.
Noise simulations
Mixtures of three types of noise at three intensity levels, producing 39 different combinations of noise, were added to each of the eight simulated ECGs. Creating a total of 312 ECGs analysed. The three types of noise contaminating the ECG with their respective frequencies at signal to noise ratios (SNR) of 30, 40 and 50 dBs, comprised: Mains noise (50 Hz), electromyographic noise (Gaussian white noise) and respiratory noise (0.25 and 0.5 Hz). Tikkanen et al, using simulated noise to define optimal QT intervals of noisy ambulatory ECGs, used an SNR range of 5–50 dB. Levels of SNR in the lower half of this range were used to simulate the extremely noisy tracings of active ambulatory patients [12]. As it was intended to simulate noise on resting ECGs in this study, an SNR level starting at 30 dB more accurately simulate the noise levels found on resting ECGs.
Baseline wander due to respiration at rates of 30 and 15 breaths per minute were simulated by a sine wave of 0.5 and 0.25 Hz with lag phases of pi/2, pi and 3pi/2 radians and respiratory modulation of 15% [12]. The highest rate of respiration was the same in the Tikkanen study to simulate a mildly breathless patient.
The SNR was calculated as 20log(standard deviation of the baseline signal/standard deviation of added noise). Abrupt step effects due to motion artefacts were not modelled as resting ECGs were simulated in this study. The mains noise was simulated using a 50 Hz sine wave, harmonics of the powerline frequencies were not modelled as the 50 Hz frequency would be dominant [13]. Gaussian white noise was simulated using a random noise generator, the respiratory effects of sinusoidal baseline wander and amplitude modulation were combined and simulated from the following function: (1 + A(sine(2(pi)R + ϕ))y(t).
Where A is the modulation index at 15%, R is the respiratory frequency (15 or 30) breaths per minute, ϕ is the phase lag of pi/2, pi and 3pi/2 radians and y(t) is the function of uncontaminated simulated ECG as a function of time. The abbreviations which describe different combinations of the three noise types at various intensities are shown below.
N0 = The baseline simulated ECG without noise
N1 = Baseline + 50 dB SNR mains noise.
N2 = Baseline + 40 dB SNR mains noise.
N3 = Baseline + 30 dB SNR mains noise.
N4 = Baseline + 50 dB SNR white noise.
N5 = Baseline + 40 db SNR white noise.
N6 = Baseline + 30 db SNR white noise.
N7 = Baseline + Respiration 15/min + pi/2 phase.
N8 = Baseline + Respiration 15/min + pi phase.
N9 = Baseline + Respiration 15/min + 3pi/2 phase.
N10 = Baseline + Respiration 30/min + pi/2 phase.
N11 = Baseline + Respiration 30/min + pi phase.
N12 = Baseline + Respiration 30/min + 3pi/2 phase.
N13 = Baseline + 50 dB SNR white noise + 30 dB SNR mains noise.
N14 = Baseline + 50 dB SNR white noise + 30 dB SNR mains noise.
N15 = Baseline + 50 dB SNR white noise + 30 dB SNR mains noise.
N16 = Baseline + Respiration 30/min + pi/2 phase + 30 dB SNRwhite noise.
N17 = Baseline + Respiration 30/min + pi/2 phase + 30 dB SNR mains noise.
N18 = Baseline + Respiration 30/min + pi/2 phase + 30 dB SNR white noise +30 dB SNR mains noise.
N19 = Baseline + Respiration 30/min + pi phase + 30 dB SNR white noise.
N20 = Baseline + Respiration 30/min + pi phase + 30 dB SNR mains noise.
N21 = Baseline + Respiration 30/min + pi phase + 30 dB SNR mains noise + 30 dB SNR white noise.
N22 = Baseline + Respiration 30/min + 3pi/2 phase + 30 dB SNR white noise.
N23 = Baseline + Respiration 30/min + 3pi/2 phase + 30 dB SNR mains noise.
N24 = Baseline + Respiration 30/min + 3pi/2 phase + 30 dB SNR mains noise + 30 dB SNR white noise.
N25 = Baseline + Respiration 15/min + 0 phase + 30 dB SNR white noise.
N26 = Baseline + Respiration 15/min + 0 phase + 30 dB SNR mains noise.
N27 = Baseline + Respiration 15/min + 0 phase + 30 dB SNR mains noise + 30 dB SNR white noise.
N28 = Baseline + Respiration 15/min + pi/2 phase + 30 dB SNR white noise.
N29 = Baseline + Respiration 15/min + pi/2 phase + 30 dB SNR mains noise.
N30 = Baseline + Respiration 15/min + pi/2 phase + 30 dB SNR mains noise + 30 dB SNR white noise.
N31 = Baseline + Respiration 15/min + pi phase + 30 dB SNR white noise.
N32 = Baseline + Respiration 15/min + pi phase + 30 dB SNR mains noise.
N33 = Baseline + Respiration 15/min + pi phase + 30 dB SNR mains noise + 30 dB SNR white noise.
N34 = Baseline + Respiration 15/min + 3pi/2 phase + 30 dB SNR white noise.
N35 = Baseline + Respiration 15/min + 3pi/2 phase + 30 dB SNR mains noise.
N36 = Baseline + Respiration 15/min + 3pi/2 phase + 30 dB SNR mains noise + 30 dB SNR white noise.
N37 = Baseline + Respiration 30/min + 0 phase + 30 dB SNR white noise.
N38 = Baseline + Respiration 30/min + 0 phase + 30 dB SNR mains noise.
N39 = Baseline + Respiration 30/min + 0 phase + 30 dB SNR white noise + 30 dB SNR mains noise.
Automatic computer algorithm
Prior to the algorithm quantifying the QT intervals, the signals N1 to N39 were smoothed using a moving median smoother over 21 consecutive time points. The residuals from the first smoothing operation were then smoothed in the same fashion as the original smoothed vector and the residual smoothed vector were added. This smoothing operation enabled low pass filtering without phase distortion. The uncontaminated T wave offset contains only low frequency signals and therefore removal of the high frequency noisy signals and absence of phase distortion would theoretically have minimal effect on the real T wave offset hidden within the noise contaminated ECG [20].
Four automatic computer algorithms were used to determine the ECG T wave offsets and have been described previously [14,20]. The threshold technique determines the end of the T wave as that time point when the ECG signal crosses the threshold at 5% the amplitude of the peak T wave (T1) or when the ECG crosses the threshold at 15% the amplitude of the peak T wave (T2). These ranges were used in the study by McLaughlin et al.
The differential threshold technique, is a method whereby the T wave end is determined as the interception of the first differential of the T wave with respect to time with the zero isoelectric line (T3). Two other commonly used algorithms based on the slope features of the T wave were also used: The slope intercept technique identifies the end of the T wave as the intercept of the line tangential to the point of maximum T wave down-slope with the isoelectric line and the least slope intercept method calculates a least squares fitted line of 8 milliseconds duration around the region of the maximum slope point and the time of intersection of this line with the zero isoelectric line is deemed the end of the T wave.
The Novel algorithm first involves a filtering process (fully described below). Following the filtering process which removes noise and produces a filtered signal the down-slope of the T wave ends when it becomes an isoelectric (that is line of zero voltage gradient ie constant millivoltage) baseline. The algorithm detects the first four milliseconds of the filtered signal which becomes a constant voltage (isoelectric) and then this four seconds of baseline plus any ensuing baseline up to the next P wave, are best least squares fit to the same duration of filtered inverted image baseline. The Novel algorithm is based on the axiomatic principle that the T wave end point is that first point of intersection (overlap) of the T wave with a superimposed inverted image of itself i.e. when both the T wave and its inverted image first coincide and return to a common baseline (the isoelectric line). This occurs when there is best least squares fit between both the T wave and its inverted image along the isoelectric line within the TP segment. This method is equivalent to using the T wave as a template which measures itself.
As is standard practice in commercial ECG machines the raw waveform was first pre-processed by a low pass filter, before any of the six algorithms were applied to the noisy ECG. This was performed by median smoothing as described previously.
The signal was then further smoothed by applying a zero-phase Butterworth 125th order low pass filter with an adaptive iterative low pass algorithm which expands the low pass threshold between a range of 30–40 Hertz by increments of 0.1 Hertz. Each iteration of filtering can be described as repetitive looping of the signal through a low pass filter, the low pass range incrementing by 0.1 Hertz during each loop. After each iteration, fitting coefficients were calculated between the 30 millisecond segments (beginning at the time of maximum negative gradient for each T wave down-slope) of the pre-filtered signal and the filtered ECG waveform produced at that iteration. The filtered waveform exhibiting the best fit between the segments of maximum down-slopes of the unfiltered and filtered T waves were then thresholded by reducing all values below 0.00001 millivolts to zero, consensual with the thresholding applied to all the computer generated ECGs. The zero phase filtering was performed to smooth the ECG isoelectric baseline maximally whilst preserving the correspondence between the T waves of the pre-filtered and filtered signal as accurately as possible. Following the low pass filtering, the image of the filtered signal demonstrating best fit with the pre-filtered signal (over the 30 millisecond segment beginning at time of maximum negative T wave downslope) was inverted and this inverted image of the filtered signal was then vertically shifted as necessary towards the upright filtered signal to the position of best least squares fit between their respective TP isoelectric (that is line of zero voltage gradient ie constant millivoltage) baselines. The TP baseline is detected by the algorithm as the first four milliseconds of the filtered signal which becomes a constant voltage and then this four milliseconds of baseline plus any ensuing baseline are best least squares fit to the same duration of filtered inverted image baseline The end of the T wave was deemed as the first time point of intersection (overlap) between the two respective isoelectric lines.
Figure 1 is a block diagram illustrating the process of the algorithm and Figure 2 illustrates step D in Figure 1 applied to an in-vivo signal.
Figure 1 Block diagram showing the Novel algorithm process. The Max grad segment refers to the 30 millisecond segment starting from the time of maximum gradient measured on the pre-filtered T wave downslope. Step A describes smoothing of the raw signal and calculation of Max grad. Step B is the iteration or looping of the algorithm as described in the text. LP is low Pass filter. Step C is the generation of a filtered signal with best fit between its Max grad and that of the smoothed signal. Step D is the overlapping between the isoelectric TP baseline segments of the signal generated in C and its inverted counterpart.
Figure 2 Example of raw in-vivo signal undergoing Novel algorithm process. The upper figure shows a raw human digitised ECG signal undergoing processing by the Novel algorithm. The first vertical line (T85) is sited at a point 85% the duration of the QT interval measured using the T wave end calculated by the Novel algorithm. The second vertical line (EndT) shows the end of the T wave as calculated by the Novel algorithm. The lower figure shows the upright and inverted T wave signals (generated in step C) undergoing step D in Figure 1. The vertical lines T85 and EndT have the same significance as in the upper part of the Figure 2.
The biphasic simulated ECGs were first processed by the following method, prior to application of the automatic computer algorithm: Amplitudes were first squared over their sampling times and then a square root was taken. This manipulation had the effect of converting the biphasic positive/negative phases of the T wave into positive/positive biphasic T waves. The automatic computer algorithm was applied to the second positive phase of the T wave.
Statistical analysis
The six different automatic computer algorithms were used to calculate the QT intervals to the nearest millisecond on each of the eight simulated ECGs contaminated with 39 different noise combinations. The mean, standard deviation and coefficient of variation (to compensate for the relative difference in QT interval magnitudes using different algorithms) were calculated and statistically compared. The single factor ANOVA was used to first identify any significant difference between the mean QT intervals calculated for the noise contaminated simulated ECGs using the different automatic algorithms. The Tukey test was then used to analyse the significant differences between each comparison.
Simulations were programmed using Mathcad 2001 mathematical and signal processing software package and run on an Advent Pentium 4, 3 GHertz processor. Statistical analysis was performed using AXUM 7 software.
Results
The simulated ECGs
Figure 3 shows the four baseline ECGs (ECGs1, 2, 3 and 4) simulated at an extracellular potassium concentration of 5.4 mmols with different combinations of resistivities and potassium channel conductances. ECGs 1, 2, 3 and 4 were simulated with respective (time dependent potassium channel conductances in mS/cm2, resistivities in Ohm cm) of (584,100,000), (262,10,000), (584,10,000) and (262,100,000) respectively. As previously described, the simulation of myocardial volume effects was achieved by assuming that each one of the four ECGs could theoretically each represent one of three possible orthogonal ECGs within a hypothetical volume of myocardium. Vector addition of various combinations of 3 ECGs would therefore produce a resultant vector ECG incorporating the characteristics of the 3 constituent orthogonal waveforms. Four resultant ECGs were produced from the following combinations Res1 = ECG2+ECG3+ECG4, Res2 = ECG1+ECG2+ECG3, Res3 = ECG1+ECG3+(-ECG2), Res4 = ECG1+ECG2+(-ECG4). Res3 and Res4 simulated positive/negative biphasic ECGs. Figures 4, 5, 6 and 7 show the baseline ECGs Res1, Res2, Res3 and Res4 respectively without superadded noise. Vertical markers above the electric baseline show the ends of the T waves as calculated by each of the algorithms without any filtering. The T waves return to the isoelectric line at times 495, 461, 461 and 495 milliseconds from the onset of each QRS complex for ECGs Res1, Res2, Res3 and Res4 respectively. The real ends of the T waves are marked with vertical markers below the baselines. Table 1 shows the results of the calculated QT intervals by each of the algorithms without filtering, on each of the resultant ECGs (Res1, 2, 3 and 4) as shown in figures 4, 5, 6 and 7. Table 2 shows the results of the calculated QT intervals by each of the algorithms without filtering, on each of the resultant ECGs Res1, Res2, Res3 and Res4 downward shifted by 0.1 millivolts. It can be seen that threshold methods T1 and T2 are very sensitive to such a simulated shift in baseline, the slope algorithms S1 and S2 being less sensitive. The threshold method T3, which calculates the time at which the graph of the first differential of the T wave with respect to time crosses the zero baseline, is unaffected by shifting the baseline. Similarly the Novel algorithm is insensitive to manoeuvres such as baseline shift.
Figure 3 The four simulated baseline ECGs. ECG1, ECG2, ECG3 and ECG4 have the following (Time dependent potassium channel conductances in mS/cm2, Resistivities in Ohm cm): (584,100,000), (262,10,000), (584,10,000) and (262,100,000). The y axis is in millivolts and the x axis is in tenth of milliseconds.
Figure 4 Resultant Vector ECG Res1. The vertical markers y1, y2, y3, y4, y5 and y6 above the electrical baseline show the calculated QT intervals by algorithms T1, T2, T3, S1, S2 and Novel respectively. The vertical marker y7 below the electrical baseline indicates the real end of the QT interval. The y axis is in millivolts and the x axis in milliseconds.
Figure 5 Resultant Vector ECG Res2. The vertical markers y1, y2, y3, y4, y5 and y6 above the electrical baseline show the calculated QT intervals by algorithms T1, T2, T3, S1, S2 and Novel respectively. The vertical marker y7 below the electrical baseline indicates the real end of the QT interval. The y axis is in millivolts and the x axis in milliseconds.
Figure 6 Resultant Vector ECG Res3. The vertical markers y1, y2, y3, y4, y5 and y6 above the electrical baseline show the calculated QT intervals by algorithms T1, T2, T3, S1, S2 and Novel respectively. The vertical marker y7 below the electrical baseline indicates the real end of the QT interval. The y axis is in millivolts and the x axis in milliseconds.
Figure 7 Resultant Vector ECG Res4. The vertical markers y1, y2, y3, y4, y5 and y6 above the electrical baseline show the calculated QT intervals by algorithms T1, T2, T3, S1, S2 and Novel respectively. The vertical marker y7 below the electrical baseline indicates the real end of the QT interval. The y axis is in millivolts and the x axis in milliseconds.
Table 1 Automatic algorithm calculation of QT intervals for unfiltered, non-amplified, non-noisy Resultant Vector ECGs. T1, T2, T3, S1, S2 and Novel are the algorithms as described in the text. Res1, Res2, Res3 and Res4 are the resultant vector ECG described in the text unfiltered, non-amplified and without added noise. The numerical values are in milliseconds.
Res1 Res2 Res3 Res4
T1 445 404 353 345
T2 430 387 348 343
T3 496 462 371 386
S1 433 379 353 346
S2 435 382 355 347
Novel 495 461 461 495
Table 2 Automatic algorithm calculation of QT intervals for unfiltered, non-amplified, non-noisy Resultant Vector ECGs 0.1 millivolt downward shifted. T1, T2, T3, S1, S2 and Novel are the algorithms as described in the text. Res1, Res2, Res3 and Res4 are the resultant vector ECG described in the text unfiltered, non-amplified and without added noise with a 0.1 millivolt downward shift of the baseline. The numerical values are in milliseconds.
Res1 Res2 Res3 Res4
T1 420 376 342 334
T2 415 365 339 332
T3 496 462 371 386
S1 420 366 343 335
S2 422 368 344 334
Novel 495 461 461 495
As previously discussed increases in conduction effects were simulated by a doubling of the amplitude at each time instant on the resultants ECGs Res1, Res2, Res3 and Res4 to give ECGs Res5, Res6, Res7 and Res8 respectively. The respective times at which the modelled T waves return to the isoelectric lines for these amplified ECGs is unchanged from the times given for their corresponding non-amplified resultant ECGs. Table 3 shows the results of the calculated QT intervals by each of the algorithms without filtering, on each of the amplified resultant ECGs Res5, Res6, Res7 and Res8. The calculated QT intervals by each of the algorithms, for each of these non-noisy, amplified, resultant ECGs, is exactly the same as the QT intervals calculated on the corresponding non-amplified resultant ECGs. All algorithms appear insensitive to uniform millivolt amplification of the T wave signal as would occur from pure conduction effects.
Table 3 Automatic algorithm calculation of QT intervals for unfiltered, amplified, non-noisy Resultant Vector ECGs. T1, T2, T3, S1, S2 and Novel are the algorithms as described in the text. Res1, Res2, Res3 and Res4 are the resultant vector ECG described in the text unfiltered, amplified by a factor of 2 and without added noise. The numerical values are in milliseconds.
Res5 Res6 Res7 Res8
T1 445 404 353 345
T2 430 387 348 343
T3 496 462 371 386
S1 433 379 353 346
S2 435 382 355 347
Novel 495 461 461 495
Figure 8 shows resultant ECG Res1 with different combinations of superadded noise as described in the Legend. Figure 9 shows the magnified T wave from the filtered signal shown in Figure 8.
Figure 8 Resultant Vector ECG Res1 with superadded noise. The superadded noise consists of 30 dB SNR white noise, 30 dB SNR mains noise and respiratory rate of 30 per minute with 0.15 amplitude modulation and 3pi/2 phase. The y axis is in millivolts and the x axis is in milliseconds.
Figure 9 Denoised T wave of Resultant Vector ECG Res1. The magnified denoised T wave from Figure 10 is shown. The y axis is in millivolts and the x axis is in milliseconds.
The means, (standard deviation) and %coefficients of variation for the measured QT intervals, on each of the baseline ECGs with superadded noise, using algorithms T1, T2, T3, SI, S2 and the Novel method are shown in Table 4. See Additional files 1, 2, 3, 4, 5, 6, 7 and 8 showing the results of QT intervals calculated by each of the algorithms on the noisy simulated ECGs; Res1, Res2, Res3, Res4, Res5, Res6, Res7 and Res8 respectively. The coefficients of variation were calculated as the percentage of the standard deviation divided by the mean value resulting from that particular algorithm method. The coefficients of variation for T1, T2, T3, S1, S2 and Novel were 0.36, 0.23, 1.9, 0.93, 0.92 and 0.62% respectively. All algorithms showed a high reproducibility and there was a significantly superior reproducibility shown by the algorithms T1, T2 and Novel compared to T3, S1 and S2. The two threshold methods T1 and T2 having a lower coefficient of variation than Novel.
Table 4 Calculation of mean QT intervals and coefficients of variation on all noisy Resultant Vector ECGs by all Automatic Algorithms. T1, T2, T3, S1, S2 and Novel are the algorithms as described in main body of text. Res1, Res2, Res3, Res4, Res5, Res6, Res7 and Res8 are the resultant vector ECGs as described in the main body of text. The mean values and their standard deviations are in milliseconds and the coefficients of variation are in percentages.
T1 T2 T3 S1 S2 Novel
Res1
Mean 448.72 431.74 466.39 426.04 446.69 493.62
SD 3.43 1.64 14.92 3.88 6.19 0.895
%CV 0.76 0.38 3.20 0.91 1.39 0.18
Res2
Mean 406.67 388.46 429.92 365.63 401.16 462.28
SD 2.03 1.06 17.34 7.64 5.38 0.88
%CV 0.50 0.27 4.05 2.09 1.34 0.19
Res3
Mean 352.59 348.85 370.44 350.61 364.09 462.87
SD 0.74 0.70 0.81 2.34 3.61 5.68
%CV 0.21 0.20 0.22 0.67 0.99 1.22
Res4
Mean 346.10 343.44 385.33 346.83 348.42 490.10
SD 0.71 0.74 1.33 0.64 1.17 0.63
%CV 0.21 0.22 0.34 0.19 0.34 0.13
Res5
Mean 447.03 430.49 462.51 427.14 443.69 495.41
SD 1.94 0.78 17.09 3.75 2.11 1.63
%CV 0.43 0.18 3.69 0.88 0.48 0.33
Res6
Mean 406.05 388 433.85 367.32 399.15 469.54
SD 1.89 1.01 14.64 7.04 5.93 8.41
%CV 0.47 0.26 3.38 1.92 1.49 1.79
Res7
Mean 352.15 348.74 370.77 351.88 361.12 463.72
SD 0.48 0.44 0.92 2.07 3.29 4.58
%CV 0.14 0.13 0.25 0.59 0.91 0.99
Res8
Mean 345.64 343.03 385.21 346.59 348.2 493.15
SD 0.58 0.70 1.36 0.60 1.48 0.70
%CV 0.17 0.20 0.35 0.17 0.42 0.14
The four ECGs with real QT intervals of 495 milliseconds (Res1, Res4, Res5 and Res8) showed mean(standard deviations) QT intervals of 396.87(1.66), 387.175(0.965), 424.86(8.675), 386.65(2.218), 396.75(2.738) and 493(0.964) milliseconds when calculated by the respective algorithms T1, T2, T3, S1, S2 and Novel. The four ECGs with real QT intervals of 461 milliseconds (Res2, Res3, Res6 and Res7) showed mean(standard deviations) QT intervals of 379.37(1.285), 368.512(0.803), 401.25(8.43), 358.86(4.773), 381.53(4.553) and 464.6(4.888) milliseconds when calculated by the respective algorithms T1, T2, T3, S1, S2 and Novel. ANOVA showed that there was a significant difference between the accuracy of the QT algorithms at a >0.95 confidence level. The Tukey test showed that at the 0.95% confidence level the Novel method was significantly more accurate in measuring real QT interval than the other algorithms. The other algorithms demonstrated the following descendent order of accuracy when measuring ECGs with real QT intervals of 495 milliseconds: T3, T1, S2, T2 and S1. Similarly, the algorithms demonstrated the following descendent order of accuracy when measuring ECGs with real QT intervals of 461 milliseconds: T3, S2, T1, T2 and S1.
The effects of respiratory noise alone did not significantly affect the reproducibility of QT interval measurements made by any of the algorithms. Analysis of the contribution from either mains noise alone or white noise alone showed them to both be responsible for QT interval measurement instability.
Discussion
This publication describes the first objective accurate physiological simulation reference standard employed to assess the accuracy of automatic computer algorithms to measure the QT interval. The reference method is based upon the synthesis of multiple ECG waveforms (derived from varying physiological parameters of resistivity and potassium channel conductance) by numerical solution of the one dimensional diffusion equation which describes current and voltage propagation through cardiac tissue [10]. The instantaneous cellular membrane currents and voltages being calculated by the thorough well founded simulations of the mammalian cardiac cell ventricular action potential derived from known ionic channel kinetics by Luo and Rudy [11]. The numerical method used to calculate the instantaneous membrane voltages was the Crank-Nicholson method which is known to be highly accurate and have good stability [16]. The four baseline ECGs (Figure 3) generated under conditions of modified potassium channel concentrations and resistivity produced T waves with a typical appearance. These four baseline ECGs were used to simulate orthogonal vector potentials at different angles to the direction of myocardial fibre orientation and through different layers of myocardial cell composition. These models are lumped models and therefore have the disadvantages of over generalisation, particularly if the investigator were interested in the electrocardiogram generated specifically from either the endocardial, M cells or epicardium. However this homogeneous approach is appropriate for the scope of the current paper and allows the description of different complexities in the myocardial repolarisation potentials generated from changes in specific and limited parameters which would arise from volume effects. Therefore it was hypothesised that each of the baseline ECGs could represent orthogonal vectors of myocardial potentials which when added in different combinations would give different resultant ECGs containing the characteristics of the three XYZ orthogonal components. The T waves generated from various hypothetical orthogonal ECGs combinations produced T wave signals of varying complexity as shown in figures 4, 5, 6 and 7. The positive/negative biphasic T waves shown in figures 6 and 7 are typical of the biphasic ECG repolarisation changes seen in-vivo. This vector resultant ECG will always contain the information of the longest QT interval from a 12 lead ECG recorded on the same myocardial volume. Indeed the information within a12 lead ECG taken on a given volume of myocardium is collapsed within the resultant ECGs from that same myocardium and is why 12 lead ECGs can be accurately reconstructed from three orthogonal vectors [21]. This is also why QT interval dispersion can be measured from orthogonal XYZ ECG vectors [22]. The scope of this study was to compare the accuracy and reproducibility of commonly used automatic QT measurement algorithms applied to eight different complexities of the scalar ECG T waves, using an accurate and entirely objective reference standard. From the above arguments, it can be appreciated that the analysis of the T waves from all 12 ECG leads was not integral to the aims of this research.
T wave inversion morphology has not been simulated, not because it poses any technical difficulty (it only requires making the baseline ECG arrays negative), but because it merely requires the squaring then square rooting of the arrays as part of pre-processing before subjecting the now upright T wave for analysis by the different algorithms. It would therefore not add any new complexity to the T waves subjected to analysis in this research.
The majority of algorithms in commercial ECG devices use a combination of leads of signals in an effort to increase signal to noise ratio as noise is a major reason for the inaccuracy and poor reproducibility of ECG interval measurement and is always present in varying degrees within ECGs taken in-vivo. Thus the use of noise simulation in order to assess the accuracy and reproducibility of ECG measurements obtained by automatic computer algorithms is commonly used [12,13,23]. This study has used established methods to produce superadded and signal amplitude modulated noise to contaminate the eight resultant ECG signals [12,13]. The types of noise and signal to noise ratios were chosen to be compatible with a resting ECG, the noise intensity levels for various physiological combinations of noise were chosen to cover a wide anticipated range without being excessive for resting conditions. Because only resting populations of ECGs were considered the study did not simulate noise artefact which would occur during ambulatory ECG recordings such as abrupt movement artefact nor was poor electrode contact noise simulated. However it is anticipated that the same Novel method could be used to accurately measure QT intervals on ambulatory ECGs with additional signal pre-processing techniques.
The commonly used algorithms which were compared in this research were based on those used in previous studies [14,20] and are those algorithms present in the majority of commercial ECG recording devices. Furthermore it is known that the slope method S2 is the most commonly used automatic algorithm in clinical ECG recorders. Previous methods used to assess different QT interval measuring algorithms have been either based on the reproducibility of the results obtained or upon the correlation with expert manual measurement which is analogous to measuring length with an elastic ruler. The methods used in assessing the algorithms cited in this report are objective, accurate and reproducible.
Table 1 shows the measured QT intervals by the various algorithms in the uncontaminated resultant ECG complexes. The Novel algorithm demonstrates a highly accurate QT measurement in both the monophasic and biphasic T wave complexes of Res3 and Res4. The T3 method is accurate for the monophasic Res1 and Res2 complexes but less accurate in the biphasic complexes. This occurs because the time of the negative phase plateau would be the time at which the first differential of the T wave crosses the isoelectric zero baseline, which is a time before the real end of the T wave. Similarly the negative phase of the T waves in Res3 and Res4 is responsible for the apparent shortening of the QT interval when calculated by T1, T2, S1 and S2, compared to measurements made by these algorithms on Res1 and Res2. Algorithms T1, T2, S1 and S2 generally underestimated the QT value and the Threshold method T1 shows the most accurate estimation out of these four methods.
Table 2 shows the measured QT intervals in the ECGs (Res1, 2, 3 and 4) with a depressed baseline. The Novel algorithm QT estimation is again highly accurate and unaffected by baseline depression. There continues to be the shortening effects of the calculated QT interval by all the algorithms for Res3 and Res4 versus Res1 and Res2 due to T wave biphasicity for all the algorithms. T3 remains unaffected and accurate for QT estimation in Res1 and Res2 because it is a threshold method which is dependent on the timing of the first differential with respect to time, which would not be affected by baseline depression.
The slope methods S1 and S2 are mildly affected by a depression in the ECG baseline because the point of maximum slope is also depressed and the unchanged projected slope gradient derived by both these methods will shorten the time of zero baseline intersection. The threshold methods T1 and T2 are most sensitive to a shift in baseline because small vertical movements in the low amplitude low frequency slope at the time of T wave offset will translate into large movements in time at which the T wave offset crosses the threshold. Table 3 demonstrates that amplification of the T wave secondary to conduction effects gives no significant changes in the results from the threshold methods. This is to be expected as the linear amplification of the T wave above the baseline has not altered the relative timings of the proportions of the T with respect to the non amplified signal. The slope methods are a function of amplitude and time and although the tangent at the point of maximum negative slope of the T wave increase with increasing T wave amplification, it increases directly in proportion to the increased amplitude of the T wave at the time of maximum down-slope i.e. the time from maximum down-slope to the time at which the tangents intersect the zero baseline is unchanged for the amplified and non amplified T waves.
The T3, S1 and S2 methods demonstrated the highest variability because they were all functions of a differential operator with respect to time on a smoothed signal with residual noise. It is known that any differentiation of a noisy signal lowers the signal to noise ratio thereby increasing the range of measured slope gradient, increasing the value of the maximum gradient and therefore increasing the potential for measurement error. The methods T1, T2 and Novel showed the least variability.
With the exception of the Novel method, the other five automatic algorithms demonstrated very poor accuracy. The method T3 showed the best accuracy out of the five algorithms because the method is in effect using the differential of the T wave signal to determine the minimum of the T wave, which in the absence of noise or any biphasic waveform would be significantly accurate
The Novel algorithm displayed a superior accuracy compared to the other measurement algorithms because the method uses the template of the T wave to in effect measure its own offset. It uses an inverted image of itself as a template to determine the first time at which the upright T wave and its inverted image return together to a common isoelectric baseline. This method therefore is the common measure of the physiological end for both the inverted and upright T waves, rather than using a surrogate marker for the end of the T wave as seen in the other algorithm methods. The Novel method is not dependent upon a preset amplitude threshold and is therefore not vulnerable to shifts in the ECG baseline like the threshold methods T1 and T2. The method does not depend upon differentiation and therefore is not vulnerable to errors arising from the biphasic waveform or the excessive noise produced by differentiation. The accuracy and reproducibility of the Novel method is susceptible to noise and although pre filtering does substantially reduce the effects of noise, recording of the ECG under optimally quiet conditions would further enhance its accuracy. As a result of the CSE study it has been emphasised that small variability rather than high accuracy is a desirable property of waveform recognition methods. It therefore follows that a combination of both these attributes is an even more desirable feature.
Although it is not in the scope of this present paper to closely examine and compare the expert manual clinical method of measuring the real simulated QT interval, it is apparent that a reader of this paper may not agree with the entirely objective, simulated T wave end point and may wish to shorten the T wave end point in order for it to be more reconciled with a conventional clinical QT interval estimation. This problem arises because the clinical measurer of the QT interval may not be visibly aware of the very low frequency low amplitude content of the terminal part of the T wave. The Novel algorithm can be used to satisfy this clinical concern by using an 85% value or negative adjusting constant of the Novel algorithm QT value. This therefore describes a QT interval fixed relatively to the QT interval as calculated by the Novel algorithm. This can be appreciated in Figure 2, which shows the 85% Novel algorithm QT in both the raw clinical ECG signal and in the post filtered ECG T wave. This topic will be the subject of future research as it is not in the scope of the present paper.
Conclusion
A new validated reference standard has been formulated based on sound physiological principles which enabled the objective assessment of the accuracy and reproducibility of commonly used automatic computer algorithms which measure the scalar ECG QT interval. A new Novel computer algorithm has also been introduced.
The commonly used computer algorithms were all shown to be inaccurate in measuring the real QT interval. The Novel algorithm demonstrated accuracy in measuring the QT interval with high reproducibility. Two threshold methods were shown to give highly reproducible but inaccurate QT interval measurements which were prone to large variable errors with ECG baseline fluctuation.
Competing interests
This paper was written with the pure scientific intention of assessing the accuracy and reproducibility of commonly used automatic computer algorithms which measure the QT interval using a new objective, reproducible and accurate reference standard. A new more accurate algorithm has been introduced which may lead to interest and further developmental research being performed. The Novel algorithm does not have a registered patent. In the long term the published results of this research may reflect favourably on PSI Heartsignals Ltd which may in turn lead to financial benefit but that could be said of any research.
Authors' contributions
The author of this paper was entirely responsible for the initial concept, the writing of the computer software and final preparation of the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Supplementary Material
Additional File 1
Automatic algorithm calculation of QT intervals for Resultant Vector ECG Res1 with superadded noise T1, T2, T3, S1, S2 and Novel are the algorithms as described in the main body of text. N1 to N39 are 39 different combinations of noise as described in the main body of the text superadded to ECG Res1. The numerical values are in milliseconds. The real QT interval is 495 milliseconds.
Click here for file
Additional File 2
Automatic algorithm calculation of QT intervals for Resultant Vector ECG Res2 with superadded noise T1, T2, T3, S1, S2 and Novel are the algorithms as described in the main body of text. N1 to N39 are 39 different combinations of noise as described in the main body of the text superadded to ECG Res2. The numerical values are in milliseconds. The real QT interval is 461 milliseconds.
Click here for file
Additional File 3
Automatic algorithm calculation of QT intervals for Resultant Vector ECG Res3 with superadded noise T1, T2, T3, S1, S2 and Novel are the algorithms as described in the main body of text. N1 to N39 are 39 different combinations of noise as described in the main body of the text superadded to ECG Res3. The numerical values are in milliseconds. The real QT interval is 461 milliseconds.
Click here for file
Additional File 4
Automatic algorithm calculation of QT intervals for Resultant Vector ECG Res4 with superadded noise T1, T2, T3, S1, S2 and Novel are the algorithms as described in the main body of text. N1 to N39 are 39 different combinations of noise as described in the main body of the text superadded to ECG Res4. The numerical values are in milliseconds. The real QT interval is 495 milliseconds.
Click here for file
Additional File 5
Automatic algorithm calculation of QT intervals for Resultant Vector ECG Res5 with superadded noise T1, T2, T3, S1, S2 and Novel are the algorithms as described in the main body of text. N1 to N39 are 39 different combinations of noise as described in the main body of the text superadded to ECG Res5. The numerical values are in milliseconds. The real QT interval is 495 milliseconds.
Click here for file
Additional File 6
Automatic algorithm calculation of QT intervals for Resultant Vector ECG Res6 with superadded noise T1, T2, T3, S1, S2 and Novel are the algorithms as described in the main body of text. N1 to N39 are 39 different combinations of noise as described in the main body of the text superadded to ECG Res6. The numerical values are in milliseconds. The real QT interval is 461 milliseconds.
Click here for file
Additional File 7
Automatic algorithm calculation of QT intervals for Resultant Vector ECG Res7 with superadded noise T1, T2, T3, S1, S2 and Novel are the algorithms as described in the main body of text. N1 to N39 are 39 different combinations of noise as described in the main body of the text superadded to ECG Res7. The numerical values are in milliseconds. The real QT interval is 461 milliseconds.
Click here for file
Additional File 8
Automatic algorithm calculation of QT intervals for Resultant Vector ECG Res8 with superadded noise T1, T2, T3, S1, S2 and Novel are the algorithms as described in the main body of text. N1 to N39 are 39 different combinations of noise as described in the main body of the text superadded to ECG Res8. The numerical values are in milliseconds. The real QT interval is 495 milliseconds.
Click here for file
==== Refs
Moss AJ Zareba W Benhorin J Malik M Couderc J-P Kennedy H Locati Heilbron E Maison-Blanche P ISHNE guidelines for electrocardiographic evaluation of drug-related QT prolongation and other alterations in ventricular repolarisation: task force summary Ann Noninvasive Electrocardiol 2001 6 333 41 11686915
Morganroth J Pyper H The use of electrocardiograms in Clinical Drug Development: Part1 Clinical Research Focus 2001 21 17 25
Myerburg RJ Castellanos A Peter Libby Cardiac arrest and sudden death Braunwald's Heart Disease: A Textbook of Cardiovascular Disease 2001 6 Saunders (W.B.) Co Ltd 899 900
Haverkamp W Breithardt G The potential for QT prolongation and proarrhythmia by non-antiarrhythmic drugs: clinical and regulatory implications European Heart Journal 2000 21 1216 1231 10924311 10.1053/euhj.2000.2249
Pratt CM Ruberg S Morganroth J McNutt B Woodward J Harris S Ruskin J Moye L Dose response relation between terfenidine and the QTc interval on the scalar electrocardiogram: Distinguishing a drug effect from spontaneous variability American Heart Journal 1996 131 472 480 8604626 10.1016/S0002-8703(96)90525-6
Ahnve S Errors in visual determination of the corrected QTc interval during acute myocardial infarction J Am Coll Cardiol 1985 5 699 702 3973268
Salvelieva I Yi G Guo X Hnatkova K Malik M Agreement and reproducibility of automatic versus manual measurement of QT interval and QT dispersion Ame J Cardiol 1998 81 471 477 10.1016/S0002-9149(97)00927-2
Kautzner J Yi G Camm AJ Malik M Short and long-term reproducibility of QT, QTc and QT dispersion measurements in healthy subjects PACE Pacing and Clinical Electrophysiology 1994 17 928 940
Willems JL Arnaud P Van Bemmel JH Bourdillon PJ Degani R Denis B Graham I Harms FM Macfarlane PW Mazzocca G Meyer J Zywietz C A reference data base for multilead electrocardiographic computer measurement programs J Am Coll Cardiol 1987 10 1313 1321 3680801
Virag N Vesin JM Kappenberger L A computer model of cardiac electrical activity for simulation of arrhythmias PACE Pacing and Clinical Electrophysiology 1998 21 2366 2371
Luo CH Rudy Y A Model of the ventricular cardiac action potential. Depolarisation, repolarisation and their interaction Circulation Research 1991 68 1501 1526 1709839
Tikkanen PE Sellin LC Kinnunen HO Huikuri HV Using simulated noise to define optimal QT intervals for computer analysis of ambulatory ECG Medical Engineering and Physics 1999 21 15 25 10220133 10.1016/S1350-4533(99)00018-1
Friesen GM Jannet TC Jadallah MA Yates SL Quint SR Nagle HT A comparison of the noise sensitivity of nine QRS detection algorithms IEEE Transactions on Biomedical Engineering 1990 37 85 97 2303275 10.1109/10.43620
Xue Q Reddy S Algorithms for computerised QT analysis Journal of Electrocardiology 1998 30 181 186 9535497 10.1016/S0022-0736(98)80072-1
Kazuttaka G Rudy Y Ionic cfurrent Basis of electrocardiographic waveforms. A model study Circ Res 2002 90 889 896 11988490 10.1161/01.RES.0000016960.61087.86
Gerald F Wheatley Parabolic and hyperbolic partial differential equations Applied Numerical Analysis 1999 6 Addison Wesley: Massachusetts 610 612
Henriquez CS Papazoglou AA Using computer models to understand the roles of tissue structure and membrane dynamics in arrhythmogenesis Proceedings of the IEEE 1996 84 334 353 10.1109/5.486738
Mirvis DM Goldberger AL Peter Libby Electrocardiography Braunwald's Heart Disease: A Textbook of Cardiovascular Disease 2001 6 Saunders (W.B.) Co Ltd 83
Plonsey R Barr RC Bioelectricity: A Quantitative Approach 1988 New York, Plenum Press 149 163
McLaughlin NB Cambell RWF Murray A Comparison of automatic QT measurement techniques in the normal 12 lead electrocardiogram British Heart Journal 1995 74 84 89 7662463
Dower G The ECGD: A derivation of the ECG from VCG Leads Journal of Electrocardiology 1984 17 189 192 6736842
Burri H Zimmermann M Bloch A Orthogonal leads for the measurement of QT dispersion: A comparison with conventional leads Int J Cardiol 2000 75 245 248 11077141 10.1016/S0167-5273(00)00339-9
Zywietz C Willems JL Arnaud P Van Bemmel JH Degani R Macfarlane PW Stability of computer ECG amplitude measurements in the presence of noise Comput Biomed Res 1990 23 10 31 2306932 10.1016/0010-4809(90)90003-U
|
16185361
|
PMC1262700
|
CC BY
|
2021-01-04 16:30:07
|
no
|
BMC Cardiovasc Disord. 2005 Sep 26; 5:29
|
utf-8
|
BMC Cardiovasc Disord
| 2,005 |
10.1186/1471-2261-5-29
|
oa_comm
|
==== Front
BMC Cardiovasc DisordBMC Cardiovascular Disorders1471-2261BioMed Central London 1471-2261-5-301618804010.1186/1471-2261-5-30Research ArticleJTT-130, a microsomal triglyceride transfer protein (MTP) inhibitor lowers plasma triglycerides and LDL cholesterol concentrations without increasing hepatic triglycerides in guinea pigs Aggarwal Dimple [email protected] Kristy L [email protected] Tosca L [email protected] Sudeep [email protected] Marcela [email protected] Maria Luz [email protected] Department of Nutritional Sciences, University of Connecticut, Storrs, CT, USA2 Department of Nutritional Sciences, University of Sinaloa, Culiacan, Mexico2005 27 9 2005 5 30 30 22 6 2005 27 9 2005 Copyright © 2005 Aggarwal et al; licensee BioMed Central Ltd.2005Aggarwal et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Microsomal transfer protein inhibitors (MTPi) have the potential to be used as a drug to lower plasma lipids, mainly plasma triglycerides (TG). However, studies with animal models have indicated that MTPi treatment results in the accumulation of hepatic TG. The purpose of this study was to evaluate whether JTT-130, a unique MTPi, targeted to the intestine, would effectively reduce plasma lipids without inducing a fatty liver.
Methods
Male guinea pigs (n = 10 per group) were used for this experiment. Initially all guinea pigs were fed a hypercholesterolemic diet containing 0.08 g/100 g dietary cholesterol for 3 wk. After this period, animals were randomly assigned to diets containing 0 (control), 0.0005 or 0.0015 g/100 g of MTPi for 4 wk. A diet containing 0.05 g/100 g of atorvastatin, an HMG-CoA reductase inhibitor was used as the positive control. At the end of the 7th week, guinea pigs were sacrificed to assess drug effects on plasma and hepatic lipids, composition of LDL and VLDL, hepatic cholesterol and lipoprotein metabolism.
Results
Plasma LDL cholesterol and TG were 25 and 30% lower in guinea pigs treated with MTPi compared to controls (P < 0.05). Atorvastatin had the most pronounced hypolipidemic effects with a 35% reduction in LDL cholesterol and 40% reduction in TG. JTT-130 did not induce hepatic lipid accumulation compared to controls. Cholesteryl ester transfer protein (CETP) activity was reduced in a dose dependent manner by increasing doses of MTPi and guinea pigs treated with atorvastatin had the lowest CETP activity (P < 0.01). In addition the number of molecules of cholesteryl ester in LDL and LDL diameter were lower in guinea pigs treated with atorvastatin. In contrast, hepatic enzymes involved in maintaining cholesterol homeostasis were not affected by drug treatment.
Conclusion
These results suggest that JTT-130 could have potential clinical applications due to its plasma lipid lowering effects with no alterations in hepatic lipid concentrations.
==== Body
Background
Microsomal triglyceride transfer protein (MTP) is a resident protein in the lumen of endoplasmic reticulum and is primarily responsible for transfer of triglycerides (TG) and other lipids from their site of synthesis in the endoplasmic reticulum into the lumen during the assembly of very low density lipoprotein (VLDL) [1]. VLDL produced by the liver are the major source of LDL in plasma and elevated levels of LDL are associated with the development of atherosclerosis and cardiovascular disease (CVD). Increased total cholesterol and LDL cholesterol (LDL-C) are both considered primary risk factors for atherosclerosis [2,3]. To reduce CHD risk factors improvements in diet and exercise are primary recommendations however when plasma cholesterol concentrations reach a certain limit drug intervention is necessary. Statins, which are targeted to 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase and are used extensively, are effective in lowering LDL-C, and somewhat effective in reducing plasma TG [4,5]. A number of studies done in the past have indicated that reduction in LDL-C values by using statins can significantly reduce the risk of CHD however a large population of patients still experience a clinical event [2,4,5]. Therefore, pharmaceutical companies are continuing to research other drug options to control hypercholesterolemia with the goal of developing a therapy for treating patients with dyslipidemias. Microsomal triglyceride transfer protein inhibitor (MTPi) is one such option. It is believed that blocking MTP will not only reduce plasma total and LDL cholesterol (LDL-C) but also plasma VLDL and TG by affecting the packaging and secretion of VLDL and chylomicrons. Certain animal and human studies [6,7] have shown that the inhibition of MTP blocks the hepatic secretion of VLDL and the intestinal secretion of chylomicrons. Consequently, this mechanism provides a highly efficacious pharmacological target for the lowering of LDL-C and reduction of postprandial lipemia. These effects could afford unprecedented benefit in the treatment of atherosclerosis and consequent cardiovascular disease. The promise of this therapeutic target has attracted widespread interest in the pharmaceutical industry.
This research study had a primary goal to evaluate whether (JTT-130), an MTPi reduces plasma cholesterol and triglyceride concentrations in male Hartley guinea pigs. Since JTT-130 is mainly targeted to the intestine, another main objective of this study was to evaluate whether this MPTi resulted in less hepatic lipid accumulation compared to other inhibitors [6,7]. Guinea pigs were used as the animal model for this study because of their similarities to humans in terms of hepatic cholesterol and lipoprotein metabolism. Previous studies done in our laboratory report that guinea pig serve as a good model for evaluating cholesterol lowering drugs [8-10].
Methods
Materials
Reagents were obtained from the following sources. JTT-130, the MTPi tested was provided by Akros Pharma Inc (Princeton, NJ). Enzymatic cholesterol and TG kits, cholesterol oxidase, cholesterol esterase and peroxidase were purchased from Roche-Diagnostics (Indianapolis, IN). Phospholipid and free cholesterol enzymatic kits were obtained from Wako Pure Chemical (Osaka, Japan). Quick-seal ultracentrifuge tubes were from Beckman (Palo Alto, CA). DL-hydroxy- [3-14C] methyl glutaryl coenzyme A (1.81 GBq/mmol), DL- [5-3H] mevalonic acid (370 GBq/mmol), cholesteryl- [1,2,6,7-3H] oleate (370 GBq/mmol), Aquasol, Liquiflor (toluene concentrate) and [14C] cholesterol were purchased from DuPont NEN (Boston, MA). Oleoyl- [1-14C] coenzyme A (1.8 GBq/mmol) and DL-3-hydroxy-3-methyl glutaryl coenzyme A were obtained from Amersham (Clearbrook, IL). Cholesteryl oleate, glucose-6-phosphate, glucose-6-phosphate dehydrogenase, nicotinamide adenine dinucleotide phosphate (NADP), sodium fluoride, Triton, bovine serum albumin and sucrose were obtained from Sigma Chemical (St. Louis, MO). Aluminum and glass silica gel plates were purchased from EM Science (Gibbstown, NJ).
Diets
Diets were prepared and pelleted by Research Diets (New Brunswick, NJ). Isocaloric diets were designed to meet all the nutritional requirements for guinea pigs. The four diets had identical composition except for the type and dose of tested drug as indicated in Table 1. The amount of cholesterol in the diets was adjusted to be 0.08 g/100 g, an amount equivalent to 600 mg/day in the human diet [11].
Table 1 Composition of Control, low dose of the inhibitor (LDI), high dose of the inhibitor (HDI) and atorvastatin diets
Components Control % LDI % HDI % Atorvastatin %
Soybean protein 22.5 22.5 22.5 22.5
Methionine 0.5 0.5 0.5 0.5
Sucrose 25 25 25 25
Corn Starch 15 15 15 15
Fat mix1 15.1 15.1 15.1 15.1
Cellulose 10 10 10 10
Guar gum 2.5 2.5 2.5 2.5
Mineral Mix2 8.2 8.2 8.2 8.2
Vitamin Mix2 1.1 1.1 1.1 1.1
Cholesterol 0.08 0.08 0.08 0.08
JTT-130 0 0.0005 0.0015 0
Statin 0 0 0 0.05
1 Fat mix for the diet contains olive oil-palm kernel oil-safflower oil (1:2:1.8), high in lauric and myristic acids.
2 Mineral and vitamin mix adjusted to meet NRC requirements for guinea pigs. Detailed composition of the vitamin and mineral mix has been reported elsewhere (Fernandez et al. 1992b).
Animals
Forty male guinea pigs (Harlan Sprague-Dawley, Hills), weighing 250–300 g, were randomly assigned to either a control, low dose of MTPi (LDI), high dose of MTPi (HDI) or an atorvastatin (AT) treatment (n = 10/group) for 4 weeks. Initially, all guinea pigs were fed the control diet for 3 weeks to raise plasma cholesterol concentrations. Two animals were housed per metal cage in a light cycle room (light from 0700–1900 h) and had free access to diets and water. Non-fasted guinea pigs were sacrificed by heart puncture after isoflurane anesthesia. Blood and livers were harvested for analysis and were stored at -80°C for further analysis. All animal experiments were conducted in accordance with U.S. Public Health Service/U.S. Department of Agriculture guidelines. Experimental protocols were approved by the University of Connecticut Institutional Care and Use Committee.
Lipoprotein isolation
Plasma samples were collected from blood obtained by heart puncture from guinea pigs under anesthesia. A preservation cocktail of aprotonin, phenyl methyl sulfonyl fluoride and sodium azide was added to plasma samples to minimize changes in lipoprotein composition during isolation. Plasma was aliquoted for LCAT and CETP determinations, plasma lipid analysis and lipoprotein isolation.
Lipoproteins were isolated by sequential ultracentrifugation [12] in a LE-80K ultracentrifuge (Beckman Instruments, Palo Alto, CA). VLDL was isolated at d = 1.006 kg/L at 125,000 g at 15°C for 19 h in a Ti-50 rotor. LDL was isolated at d = 1.019-1.09 kg/L in quick-seal tubes at 15°C for 3 h at 200,000 g in a vertical Ti-65 rotor [13]. LDL samples were dialyzed in 0.9 g/L sodium chloride-0.1 g/L ethylene diamine tetra acetic acid (EDTA), pH 7.2, for 12 h and stored at 4°C for further analysis.
Plasma and hepatic lipids
Plasma samples were analyzed for cholesterol and TG by enzymatic methods [14]. Hepatic total and free cholesterol and TG were determined according to the method by Carr et al. [15] following extraction of hepatic lipids with chloroform-methanol 2:1. Cholesteryl ester concentrations were calculated by subtracting free from total cholesterol.
Lipoprotein characterization
VLDL and LDL composition was calculated by determining free and esterified cholesterol [14], protein by a modified Lowry method [16], and TG and phospholipids by enzymatic kits. VLDL apo B was selectively precipitated with isopropanol [17]. The number of constituent molecules of LDL was calculated on the basis of one apo B per particle with a molecular mass of 412000 kD[18]. The molecular weights were 885.4, 386.6, 645 and 734 for TG, free and esterified cholesterol, and phospholipids, respectively [19]. LDL diameters were calculated according to Van Heek et al [20]. HDL cholesterol was also determined according to Warnick et al, with a modification, which consisted of using 2 mol/L MgCl2 for precipitation of apo-B containing lipoproteins [13].
Lecithin Cholesterol Acyltransferase (LCAT) and Cholesterol Ester Transfer Protein (CETP) determinations in plasma
LCAT and CETP activities were determined according to Ogawa & Fielding [21]. Physiological CETP activity was determined without inhibiting LCAT activity by measuring the mass transfer of cholesterol ester between HDL and apo B containing lipoproteins. Samples were incubated at 37°C for 6 h in a shaking water bath and total and free plasma cholesterol and HDL cholesterol were measured. LCAT activity was determined by mass analysis of the decrease in plasma free cholesterol between 0 and 6 h at 37°C. Assays were carried out concurrently with measurements of CETP. Both of these methods have been well-standardized for guinea pig plasma [22].
Hepatic microsome isolation
Hepatic microsomes were isolated as described previously [8]. Briefly a microsomal fraction was isolated by two 25-min centrifugations at 10,000 g (JA-20 rotor, J2-21) followed by ultracentrifugation at 100,000 g in a Ti-50 rotor at 4°C for 1 hour. Microsomes were resuspended in the homogenization buffer and centrifuged for an additional hour at 100,000 g. After centrifugation, microsomal pellets were homogenized and stored at -70°C.
Hepatic HMG-CoA reductase assay
The activity of microsomal HMG-CoA reductase (E.C. 1.1.1.34) was measured in hepatic microsomes as described by Shapiro et al. [23]. HMG-CoA reductase activity was expressed as pmol of [14C] mevalonate produced per min per mg microsomal protein. Recoveries of [3H] mevalonate ranged from 60–90%.
Hepatic Acyl CoA Cholesteryl Acyltransferase (ACAT) activity
Hepatic ACAT (E.C. 2.3.1.26) activity was measured by the incorporation of [14C] oleoyl CoA in cholesteryl ester in hepatic microsomes by preincubating 0.8–1 mg of microsomal protein per assay with 84 g/L albumin and buffer for microsomal isolation [24]. Recoveries of [3H] cholesteryl oleate were between 70–90%.
Hepatic cholesterol 7α-hydroxylase activity
Cholesterol 7α-hydroxylase (E.C. 1.14.13.7) activity was measured according to the method modified by Jelinik et al [25]. [14C] cholesterol was used as a substrate and delivered as cholesterol-phosphatidylcholine liposomes (1:8 by weight) prepared by sonication. An NADPH-regenerating system (glucose-6-phosphate dehydrogenase, NADP, and glucose-6-phosphate) was included in the assay as a source of NADPH.
Statistical analysis
One-way analysis of variance (ANOVA) (SSPS for Windows version 12) was used to evaluate significant differences among groups in regards to plasma and hepatic lipids, LDL composition, hepatic enzyme activities and LCAT and CETP activities. The LSD post hoc test was used to evaluate the differences among groups. Data are presented as the mean ± SD. Differences were considered significant at P < 0.05.
Results
Plasma lipids and lipoproteins
No difference in weight gain overtime was observed in guinea pigs fed the different test diets (Fig 1), indicating that animals consumed comparable amounts of their respective test diets. After feeding the test diets for a period of four weeks, blood was isolated and plasma was analyzed for cholesterol and TG concentrations. The two doses of MTPi evaluated, low dose (LDI) and high dose (HDI) decreased plasma total cholesterol values significantly by 19.2% (P < 0.01) as compared to the control animals (Table 2). There was no significant difference between the two doses of MTPi tested. Atorvastatin, which was used as a positive control, led to a significantly robust decrease in plasma total cholesterol values of 46%, which was significantly different (P < 0.01) from the two MTPi doses used. Plasma TG values were also significantly lower in LDI (50.8%) and HDI (45.3%) when compared to their control counterparts whereas atorvastatin (AT) treatment resulted in the maximum decrease of plasma TG (Table 2).
Figure 1 Weight gain of guinea pigs treated with control, low dose, high dose of JTT-130 and atorvastatin.
Table 2 Plasma total cholesterol (TC), triglycerides (TG), VLDL-C, LDL-C and HDL-C of guinea pigs fed a control diet, low dose MTPi (LDI), high dose MTPi (HDI) or atorvastatin
Diets TC TG VLDL-C LDL-C HDL-C
(mg/dL)
Control (10) 146.9 ± 42.2a 135.8 ± 118.9a 8.1 ± 5.5ab 123.9 ± 43.5a 13.6 ± 4.5a
LDI (10) 116.9 ± 29.7b 66.7 ± 29.6b 3.7 ± 3.0a 93.3 ± 26.2b 18.5 ± 8.0a
HDI (9) 116.1 ± 23.3b 74.3 ± 31.7b 9.0 ± 8.6b 90.5 ± 28.8b 15.6 ± 7.1a
Atorvastatin (9) 76.5 ± 29.8c 49.4 ± 32.2c 2.9 ± 2.6a 59.4 ± 29.3c 14.1 ± 5.6a
1Data are presented as mean ± SD for the number of guinea pigs indicated in parenthesis. Numbers in a column with different superscripts are considered significantly different (P < 0.01) as determined by one way ANOVA and LSD as a post-hoc test
The changes in plasma TC values were mostly due to decreases in the cholesterol carried by LDL. LDL-C was also significantly decreased (24.7% & 26.9%) by the MTPi diets tested. There were no major differences between these two doses for plasma lipid parameters except for VLDL-C, which were significantly higher when compared to the low dose of the drug. No significant differences were observed for HDL-C values with MTPi or with AT (Table 2).
LDL size and composition
No significant effect of MTPi on the number of CE molecules or on the size of LDL particle was found with any of the two doses tested as compared to their control counterparts (Table 3). AT treatment reduced the number of esterified cholesterol molecules (45%) as well as decreased the size of LDL particle (30%) (Table 3).
Table 3 Number of molecules of cholesteryl ester (CE), LDL diameter and LCAT and CETP activities of guinea pigs fed a control diet, low dose MTPi (LDI), high dose MTPi (HDI) or atorvastatin.
Diets CE molecules LDL diameter LCAT CETP
nanometers (pmol/min.mg)
Control (n = 10) 1993 ± 422a 16.47 ± 3.63a 19.6 ± 11.0a 36.1 ± 12.5a
LDI (n= 10) 2072 ± 536a 16.67 ± 3.05a 14.4 ± 6.4a 31.0 ± 21.1a
HDI (n = 9) 2026 ± 132.5a 16.58 ± 7.85a 14.6 ± 8.5a 19.4 ± 13.0ab
Atorvastatin (n = 9) 1080 ± 1092b 11.52 ± 2.69b 13.4 ± 10.9a 12.8 ± 4.2b
1Data are presented as mean ± SD for the number of guinea pigs indicated in parenthesis. Numbers in a column with different superscripts are considered significantly different (P < 0.01) as determined by one way ANOVA and LSD as post hoc test.
LCAT and CETP activities
Table 3 also summarizes the activities of these two proteins, which play a major role in the intravascular processing of plasma cholesterol. There were no significant differences in LCAT activity when comparing MTPi or statin groups to the control group. However, HDI decreased the activity of CETP, which was comparable to the atorvastatin treated group (P < 0.01).
Hepatic lipids and enzymes
No significant changes were found in hepatic total cholesterol, free cholesterol, cholesteyl ester or TG values in any of the four treatments (Table 4). Results suggest that MTPi did not lead to lipid accumulation in the liver, as there was no significant difference between the two doses of MTPi tested with control or with atorvastatin group. Similarly no significant changes were observed in any of the three regulatory hepatic enzymes involved, namely CYP7, ACAT and HMG-CoA Reductase (Table 5).
Table 4 Hepatic total cholesterol (TC), free cholesterol (FC) cholesteryl ester (CE) and TG of guinea pigs fed a control diet, low dose MTPi (LDI), high dose MTPi (HDI) or atorvastatin1.
Diets TC FC CE TG
(mg/g)
Control (n = 10) 1.20 ± 0.53 1.00 ± 0.44 0.20 ± 0.23 16.2 ± 2.8
LDI (n = 10) 1.47 ± 0.40 1.24 ± 0.35 0.23 ± 0.11 13.9 ± 9.4
HDI (n = 9) 1.42 ± 0.66 0.98 ± 0.39 0.44 ± 0.58 13.5 ± 7.8
Atorvastatin (n = 9) 1.71 ± 0.76 1.31 ± 0.56 0.40 ± 0.39 18.8 ± 10.8
1Data are presented as mean ± SD for n = the number of guinea pigs indicated in parenthesis.
Table 5 Hepatic HMG-CoA reductase, ACAT and cholesterol 7α-hydroxylase activities (CYP7) guinea pigs fed a control diet, low dose MTPi (LDI), high dose MTPi (HDI) or atorvastatin1
Diets HMG-CoA Reductase ACAT CYP7
(pmol/min.mg)
Control 1.8 ± 0.8 0.8 ± 0.4 1.7 ± 1.8
LDI 2.7 ± 0.8 2.2 ± 0.9 1.2 ± 0.5
HDI 2.5 ± 1.4 3.5 ± 2.4 0.6 ± 0.7
Atorvastatin 3.4 ± 2.9 4.4 ± 3.1 0.7 ± 0.4
1Data are presented as mean ± SD for n = 3–7 guinea pigs per group.
Discussion
In this study we were able to demonstrate that JTT-130, the MTPi tested has the potential to decrease the prime risk factors of cardiovascular disease, namely plasma TG and LDL-C concentrations in guinea pigs. The novelty of the drug tested is that there was no significant accumulation of lipid in the liver as seen in some other studies done with other MTPi [26,27].
Drug treatment and hepatic lipids
Previous studies evaluating MTPi have shown increase in the lipid content of the liver [7,26,27]. Chandler et al [7] treated Hep-G2 cells with CP-346086, another MTPi, for a period of two weeks. They reported that in addition to decreasing plasma TC, LDL-C, VLDL-C and TG values, this treatment also increased hepatic and intestinal TG when the MTPi was administered with food and when it was dosed away from meals, only hepatic TG were influenced. In contrast, the major finding of this study is that the MTPi tested did not lead to any fat accumulation in the liver as confirmed by no significant changes found in the hepatic lipid content as compared to their control counterparts. The main reason for these differences between MTPi could be that the main target of JTT-130 was the intestine. Because of this, we speculate that due to MTP inhibition, less TG were transferred to the chylomicron particle being packaged in the intestine. As a result a lower concentration of TG was taken up by the hepatocytes through the chylomicron remnant. Thus the VLDL particles secreted from the liver had lower concentrations of TG molecules due to the major inhibitory effect of the MTPi in the intestine. Because there were no significant changes in hepatic cholesterol concentrations, we did not find any significant differences in hepatic enzyme activities. Similar to the study by Conde et al. [9] in atorvastatin treated guinea pigs with 0.015% atorvastatin, there were no significant differences in hepatic cholesterol concentrations when compared with a control group. However, significant differences in hepatic esterified cholesterol were observed when guinea pigs were treated with a higher dose of the statin (0.05%) [9].
Drug treatment and plasma lipids and lipoproteins
Abetalipoproteinemia, a genetic disorder characterized by low plasma cholesterol and TG levels, is caused by a functional deficiency of MTP. Absence of lipid transfer activity in the microsomes of abetalipoproteinemia patients established its pivotal function in lipoprotein assembly[1]. This finding led to the suggestion that MTP inhibition could be used as a possible lipid lowering therapy. Further evidence was obtained from a cell culture study in which researchers [28] proved that MTP is limiting in the production of apo B containing lipoproteins. Another study [6] further confirmed this finding using heterozygous MTP knockout mice which had 20% less plasma total cholesterol levels compared to wild type mice fed high fat diets; however, they did not find any significant differences in plasma TG concentrations. In our study, we have demonstrated that animals treated with MTPi had not only lower plasma TC and LDL-C but also significant reductions in plasma TG. It is possible that the VLDL particle secreted by the liver was more readily catabolized and therefore there was less conversion to LDL, which contributed to the hypocholesterolemic effects of the MTPi. Conde et al. [9] demonstrated that there was a significant reduction in plasma TG in guinea pigs treated with atorvastatin when compared to controls. This was partly explained by lower secretion of VLDL particles and by increases in the LDL receptor [9], which could have contributed to the faster removal of VLDL particles. A similar mechanism may have taken place with the MTPi. By blocking MTP, JTT-130 reduced the secretion of VLDL particles, and therefore, the formation of LDL in the plasma.
There was a dose response in CETP activity with JTT-130, and in addition, guinea pigs treated with atorvastatin exhibited decreased activity of this transfer protein. The main function of CETP is to contribute to the reverse cholesterol pathway by transferring cholesteryl ester from HDL to the apo B containing lipoproteins [29]. However, this action prolongs the residence time of CE in LDL and increases the possibility of its deposition in the arterial wall. Thus lower CETP activity has been associated with decreased atherogenesis in animal studies [30]. Therefore the lowering of CETP activity by drug treatment can be considered beneficial.
Results from this study indicate that JTT-130 has the potential to reduce the primary risk factors for coronary heart disease. While these results are quite promising, more studies are needed to clarify the possibility of adverse effects including steatorhea, fat malabsorption and fat-soluble vitamin absorption. Although the lipid lowering effects were not as pronounced as those observed with atorvastain, the doses of MTPi used in the current study were lower than the doses of atorvastatin. It is possible that using JTT-130 in combination with statins could reduce the wide array of adverse effects associated with reductase inhibitors [31]. This study also demonstrates that MTP inhibitor which is mainly targeted towards the intestine may open a new avenue for treatment of hyperlipidemic patients who are at high risk for cardiovascular diseases.
List of abbreviations used
ACAT: acyl CoA cholesteryl acyl transferase; AT: atorvastatin; CETP: cholesterol ester transfer protein; CHD: coronary heart disease; CYP7: cholesterol 7α-hydroxylase; HDi: high dose of the inhibitor; HDL-C: HDL cholesterol; HMG-CoA; 3-hydroxy-3methyl glutaryl Conezyme A; LCAT: lecithin cholesterol acyltransferase; LDi: low dose of the inhibitor; LDL-C: LDL cholesterol; MTP: microsomal transfer protein; MTPi: microsomal transfer protein inhibitor; NADP: nicotinamide adenenine dinucleotide phosphate; TG: triglycerides; VLDL: very low density lipoprotein.
Competing interests
Authors received funding from Akros Pharma Inc. (Princeton, NJ) to carry out the studies presented in this manuscript.
Authors' contributions
DA did the assays, wrote the manuscript and participated in the interpretation of data; KLW: assisted in the assays for plasma lipids, CETP and LCAT; TLZ: assisted in the determination of ACAT activity and participated in data interpretation, SS: assisted in taking care of guinea pigs, isolation of microsomes and data interpretation; MVJ assisted in the determination of CYP7 and in data interpretation and MLF designed the experiment, evaluated the results, interpreted the data and participated in manuscript preparation.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
These studies were supported by Akros Pharma Inc, Princeton, NJ
==== Refs
Hussain MM Iqbal J Anwar K Rava P Dai K Microsomal triglyceride transfer protein: a multifunctional protein Front Biosci 2003 500 506
National Cholesterol Education Program Expert Panel. Second report of the National Cholesterol Education Program (NCEP) Expert panel on Detection Evaluation and Treatment of High Blood Cholesterol in Adults; Adult Treatment Panel II Circulation 1994 89 1329 1445
Association AH Heart and Stroke Statistical Update American Heart Association: 2001 2001
Shepherd J Cobbe SM Ford I Isles CG Lorimer AR MacFarlane PW McKillop JH Packard CJ Prevention of coronary heart disease with pravastatin in men with hypercholesterolemia N Eng J Med 1995 333 1301 1307 10.1056/NEJM199511163332001
Sacks FM Pfeffer MA Moye LA Rouleau JL Rutherford JD Cole TG Brown L Warnica JW Arnold JMO Wun CC Davis BR Braunwald E The effect of pravastatin on coronary events after myocardial infarction in patients with average cholesterol levels N Eng J Med 1996 335 1001 1009 10.1056/NEJM199610033351401
Raabe M Veniant MM Sullivan MA Zlot CH Bjorkegren J Nielson LB Wong JS Hamilton RL Young SG Analysis of the role of microsomal triglyceride transfer protein in the liver of tissue-specific knockout mice J Clin Invest 1999 103 1287 1298 10225972
Chandler CE Wilder DE Pettini JL Savoy YE Petras SF Chang G Vincent J Harwood HJ CP-346086: an MTP inhibitor that lowers plasma cholesterol and triglycerides in experimental animals and in humans J Lipid Res 2003 44 1887 1901 12837854 10.1194/jlr.M300094-JLR200
West KL Fernandez ML Guinea pigs as models to study the hypercholesterolemic effects of drugs Cardiovasc Drug Rev 2004 22 55 70 14978518
Conde K Pineda G Newton RS Fernandez ML Hypocholesterolemic effects of 3-hydroxy-3-methylglutaryl coenzyme a (HMG-CoA) reductase inhibitors in the guinea pig Biochem Pharmacol 1999 58 1209 1219 10484079 10.1016/S0006-2952(99)00203-8
Fernandez ML Guinea pigs as models for cholesterol and lipoprotein metabolism J Nutr 2001 131 10 20 11208932
Lin ECK Fernandez ML McNamara DJ Dietary fat type and cholesterol quantity interact to affect cholesterol metabolism in guinea pigs J Nutr 1992 122 2019 2029 1527642
Redgrave TG Roberts DC West CE Separation of plasma lipoproteins by density-gradient ultracentrifugation Anal Biochem J 1975 65 42 49 10.1016/0003-2697(75)90488-1
Fernandez ML Wilson TA Conde K Vegara-Jimenez M Nicolosi RJ Hamsters and guinea pigs differ in their plasma lipoprotein cholesterol distribution when fed diets varying in animal protein, soluble fiber or cholesterol content J Nutr 1999 129 1323 1333 10395594
Allain CC Poon LC Chan CS Richard W Fu PC Enzymatic determination of total serum cholesterol Clin Chem 1974 20 47 475
Carr TP Anderssen CJ Rudel LL Enzymatic determination of triglycerides, free cholesterol and cholesterol in tissue lipid extracts Clin Biochem 1993 26 39 42 8448837 10.1016/0009-9120(93)90015-X
Markwell MK Haas S Bieber LL Tolbert NE A modification of the Lowry procedure to simplify protein determination in membrane and lipoprotein samples Anal Biochem J 1978 87 206 210 10.1016/0003-2697(78)90586-9
Homquist L Carlson K Carlson LA Comparison between the use of isopropanol and tetramethylurea for the solublization and quantitation of human serum very low density apolipoproteins Anal Biochem J 1987 149 423 426
Chapman JM Mills GL Ledford JH The distribution and partial characterization of the serum apolipoproteins in the guinea pig Biochem J 1975 149 423 36 170915
Conde K Vergara-Jimenesz M Krause BR Newton RS Fernandez ML Hypocholesterolemic actions of atorvastatin are associated with alterations on hepatic cholesterol metabolism and lipoprotein composition in the guinea pig J Lipid Res 1996 37 739 746 8732774
Van Heek M Zilversmit D Mechanisms of hypertriglyceridemia in the coconut oil/cholesterol-fed rabbits increased secretion and decreased catabolism of VLDL Arteriosclerosis Thromb 1991 11 918 927
Ogawa YFCJ Assay of cholesterol ester transfer activity and purification of a cholesterol ester transfer protein Meth Enzymol 1985 111 274 285 3897772
Fernandez ML Conde K Vergara-Jimenesz M Behr T Abdel-Fatta G Regulation of VLDL-LDL Apo B metabolism in guinea pigs by dietary soluble fiber Am J Clin Nutr 1997 65 814 822 9062534
Shapiro DL Imblum RL Rodwell VW Thin-layer chromatographic assay for the HMG-CoA reductase and mevalonic acid Anal Biochem J 1969 31 383 390 10.1016/0003-2697(69)90279-6
Smith JL de Jersey J pillay SP Hardie IR Hepatic acyl-CoA: cholesterol acyltransferase. Development of a standard assay and determination in patients with cholesterol gallstones Clin Chim Acta 1986 158 271 282 3769201 10.1016/0009-8981(86)90291-3
Jelinek DF Andersson S Slaughter CA Russel DW Cloning and regulation of cholesterol 7a-hydroxylase, the rate limiting enzyme in bile acid biosynthesis J Biol Chem 1990 265 8190 8197 2335522
Liao W Hui TY Young SG Davis RA Blocking microsomal triglyceride transfer protein interferes with apoB secretion without causing retention or stress in the ER J Lipid Res 2003 44 978 985 12588952 10.1194/jlr.M300020-JLR200
Shiomi M Ito T MTP inhibitor decreases plasma cholesterol levels in LDL receptor-deficient WHHL rabbits by lowering the VLDL secretion Eur J Pharmacol 2001 431 127 131 11716851 10.1016/S0014-2999(01)01419-4
Jamil H Gordon DA Eustice DC Brooks CM Dickson JK JrChen Y Ricci B Chu CH Harrity TW Ciosek CP Biller SA JrGregg RE Wetterau JR An inhibitor of the microsomal triglyceride transfer protein inhibits apoB secretion from HepG2 cells Proc Natl Acad Sci USA 1996 93 11991 11995 8876250 10.1073/pnas.93.21.11991
Barter P CETP and atherosclerosis Arterioscler Thromb Vasc Biol 2000 20 2029 2031 10978244
Le Goff W Guerin M Chapman MJ Pharmacological modulation of cholesteryl ester transfer protein, a new therapeutic target in atherogenic dyslipidemia Pharmacol Ther 2004 101 17 38 14729390 10.1016/j.pharmthera.2003.10.001
Ansell BJ Rationale for combination therapy with statin drugs in the treatment of dyslipidemia Curr Atheroscler Rep 2005 7 29 33 15683599
|
16188040
|
PMC1262701
|
CC BY
|
2021-01-04 16:30:07
|
no
|
BMC Cardiovasc Disord. 2005 Sep 27; 5:30
|
utf-8
|
BMC Cardiovasc Disord
| 2,005 |
10.1186/1471-2261-5-30
|
oa_comm
|
==== Front
BMC Cardiovasc DisordBMC Cardiovascular Disorders1471-2261BioMed Central London 1471-2261-5-301618804010.1186/1471-2261-5-30Research ArticleJTT-130, a microsomal triglyceride transfer protein (MTP) inhibitor lowers plasma triglycerides and LDL cholesterol concentrations without increasing hepatic triglycerides in guinea pigs Aggarwal Dimple [email protected] Kristy L [email protected] Tosca L [email protected] Sudeep [email protected] Marcela [email protected] Maria Luz [email protected] Department of Nutritional Sciences, University of Connecticut, Storrs, CT, USA2 Department of Nutritional Sciences, University of Sinaloa, Culiacan, Mexico2005 27 9 2005 5 30 30 22 6 2005 27 9 2005 Copyright © 2005 Aggarwal et al; licensee BioMed Central Ltd.2005Aggarwal et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Microsomal transfer protein inhibitors (MTPi) have the potential to be used as a drug to lower plasma lipids, mainly plasma triglycerides (TG). However, studies with animal models have indicated that MTPi treatment results in the accumulation of hepatic TG. The purpose of this study was to evaluate whether JTT-130, a unique MTPi, targeted to the intestine, would effectively reduce plasma lipids without inducing a fatty liver.
Methods
Male guinea pigs (n = 10 per group) were used for this experiment. Initially all guinea pigs were fed a hypercholesterolemic diet containing 0.08 g/100 g dietary cholesterol for 3 wk. After this period, animals were randomly assigned to diets containing 0 (control), 0.0005 or 0.0015 g/100 g of MTPi for 4 wk. A diet containing 0.05 g/100 g of atorvastatin, an HMG-CoA reductase inhibitor was used as the positive control. At the end of the 7th week, guinea pigs were sacrificed to assess drug effects on plasma and hepatic lipids, composition of LDL and VLDL, hepatic cholesterol and lipoprotein metabolism.
Results
Plasma LDL cholesterol and TG were 25 and 30% lower in guinea pigs treated with MTPi compared to controls (P < 0.05). Atorvastatin had the most pronounced hypolipidemic effects with a 35% reduction in LDL cholesterol and 40% reduction in TG. JTT-130 did not induce hepatic lipid accumulation compared to controls. Cholesteryl ester transfer protein (CETP) activity was reduced in a dose dependent manner by increasing doses of MTPi and guinea pigs treated with atorvastatin had the lowest CETP activity (P < 0.01). In addition the number of molecules of cholesteryl ester in LDL and LDL diameter were lower in guinea pigs treated with atorvastatin. In contrast, hepatic enzymes involved in maintaining cholesterol homeostasis were not affected by drug treatment.
Conclusion
These results suggest that JTT-130 could have potential clinical applications due to its plasma lipid lowering effects with no alterations in hepatic lipid concentrations.
==== Body
Background
Microsomal triglyceride transfer protein (MTP) is a resident protein in the lumen of endoplasmic reticulum and is primarily responsible for transfer of triglycerides (TG) and other lipids from their site of synthesis in the endoplasmic reticulum into the lumen during the assembly of very low density lipoprotein (VLDL) [1]. VLDL produced by the liver are the major source of LDL in plasma and elevated levels of LDL are associated with the development of atherosclerosis and cardiovascular disease (CVD). Increased total cholesterol and LDL cholesterol (LDL-C) are both considered primary risk factors for atherosclerosis [2,3]. To reduce CHD risk factors improvements in diet and exercise are primary recommendations however when plasma cholesterol concentrations reach a certain limit drug intervention is necessary. Statins, which are targeted to 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase and are used extensively, are effective in lowering LDL-C, and somewhat effective in reducing plasma TG [4,5]. A number of studies done in the past have indicated that reduction in LDL-C values by using statins can significantly reduce the risk of CHD however a large population of patients still experience a clinical event [2,4,5]. Therefore, pharmaceutical companies are continuing to research other drug options to control hypercholesterolemia with the goal of developing a therapy for treating patients with dyslipidemias. Microsomal triglyceride transfer protein inhibitor (MTPi) is one such option. It is believed that blocking MTP will not only reduce plasma total and LDL cholesterol (LDL-C) but also plasma VLDL and TG by affecting the packaging and secretion of VLDL and chylomicrons. Certain animal and human studies [6,7] have shown that the inhibition of MTP blocks the hepatic secretion of VLDL and the intestinal secretion of chylomicrons. Consequently, this mechanism provides a highly efficacious pharmacological target for the lowering of LDL-C and reduction of postprandial lipemia. These effects could afford unprecedented benefit in the treatment of atherosclerosis and consequent cardiovascular disease. The promise of this therapeutic target has attracted widespread interest in the pharmaceutical industry.
This research study had a primary goal to evaluate whether (JTT-130), an MTPi reduces plasma cholesterol and triglyceride concentrations in male Hartley guinea pigs. Since JTT-130 is mainly targeted to the intestine, another main objective of this study was to evaluate whether this MPTi resulted in less hepatic lipid accumulation compared to other inhibitors [6,7]. Guinea pigs were used as the animal model for this study because of their similarities to humans in terms of hepatic cholesterol and lipoprotein metabolism. Previous studies done in our laboratory report that guinea pig serve as a good model for evaluating cholesterol lowering drugs [8-10].
Methods
Materials
Reagents were obtained from the following sources. JTT-130, the MTPi tested was provided by Akros Pharma Inc (Princeton, NJ). Enzymatic cholesterol and TG kits, cholesterol oxidase, cholesterol esterase and peroxidase were purchased from Roche-Diagnostics (Indianapolis, IN). Phospholipid and free cholesterol enzymatic kits were obtained from Wako Pure Chemical (Osaka, Japan). Quick-seal ultracentrifuge tubes were from Beckman (Palo Alto, CA). DL-hydroxy- [3-14C] methyl glutaryl coenzyme A (1.81 GBq/mmol), DL- [5-3H] mevalonic acid (370 GBq/mmol), cholesteryl- [1,2,6,7-3H] oleate (370 GBq/mmol), Aquasol, Liquiflor (toluene concentrate) and [14C] cholesterol were purchased from DuPont NEN (Boston, MA). Oleoyl- [1-14C] coenzyme A (1.8 GBq/mmol) and DL-3-hydroxy-3-methyl glutaryl coenzyme A were obtained from Amersham (Clearbrook, IL). Cholesteryl oleate, glucose-6-phosphate, glucose-6-phosphate dehydrogenase, nicotinamide adenine dinucleotide phosphate (NADP), sodium fluoride, Triton, bovine serum albumin and sucrose were obtained from Sigma Chemical (St. Louis, MO). Aluminum and glass silica gel plates were purchased from EM Science (Gibbstown, NJ).
Diets
Diets were prepared and pelleted by Research Diets (New Brunswick, NJ). Isocaloric diets were designed to meet all the nutritional requirements for guinea pigs. The four diets had identical composition except for the type and dose of tested drug as indicated in Table 1. The amount of cholesterol in the diets was adjusted to be 0.08 g/100 g, an amount equivalent to 600 mg/day in the human diet [11].
Table 1 Composition of Control, low dose of the inhibitor (LDI), high dose of the inhibitor (HDI) and atorvastatin diets
Components Control % LDI % HDI % Atorvastatin %
Soybean protein 22.5 22.5 22.5 22.5
Methionine 0.5 0.5 0.5 0.5
Sucrose 25 25 25 25
Corn Starch 15 15 15 15
Fat mix1 15.1 15.1 15.1 15.1
Cellulose 10 10 10 10
Guar gum 2.5 2.5 2.5 2.5
Mineral Mix2 8.2 8.2 8.2 8.2
Vitamin Mix2 1.1 1.1 1.1 1.1
Cholesterol 0.08 0.08 0.08 0.08
JTT-130 0 0.0005 0.0015 0
Statin 0 0 0 0.05
1 Fat mix for the diet contains olive oil-palm kernel oil-safflower oil (1:2:1.8), high in lauric and myristic acids.
2 Mineral and vitamin mix adjusted to meet NRC requirements for guinea pigs. Detailed composition of the vitamin and mineral mix has been reported elsewhere (Fernandez et al. 1992b).
Animals
Forty male guinea pigs (Harlan Sprague-Dawley, Hills), weighing 250–300 g, were randomly assigned to either a control, low dose of MTPi (LDI), high dose of MTPi (HDI) or an atorvastatin (AT) treatment (n = 10/group) for 4 weeks. Initially, all guinea pigs were fed the control diet for 3 weeks to raise plasma cholesterol concentrations. Two animals were housed per metal cage in a light cycle room (light from 0700–1900 h) and had free access to diets and water. Non-fasted guinea pigs were sacrificed by heart puncture after isoflurane anesthesia. Blood and livers were harvested for analysis and were stored at -80°C for further analysis. All animal experiments were conducted in accordance with U.S. Public Health Service/U.S. Department of Agriculture guidelines. Experimental protocols were approved by the University of Connecticut Institutional Care and Use Committee.
Lipoprotein isolation
Plasma samples were collected from blood obtained by heart puncture from guinea pigs under anesthesia. A preservation cocktail of aprotonin, phenyl methyl sulfonyl fluoride and sodium azide was added to plasma samples to minimize changes in lipoprotein composition during isolation. Plasma was aliquoted for LCAT and CETP determinations, plasma lipid analysis and lipoprotein isolation.
Lipoproteins were isolated by sequential ultracentrifugation [12] in a LE-80K ultracentrifuge (Beckman Instruments, Palo Alto, CA). VLDL was isolated at d = 1.006 kg/L at 125,000 g at 15°C for 19 h in a Ti-50 rotor. LDL was isolated at d = 1.019-1.09 kg/L in quick-seal tubes at 15°C for 3 h at 200,000 g in a vertical Ti-65 rotor [13]. LDL samples were dialyzed in 0.9 g/L sodium chloride-0.1 g/L ethylene diamine tetra acetic acid (EDTA), pH 7.2, for 12 h and stored at 4°C for further analysis.
Plasma and hepatic lipids
Plasma samples were analyzed for cholesterol and TG by enzymatic methods [14]. Hepatic total and free cholesterol and TG were determined according to the method by Carr et al. [15] following extraction of hepatic lipids with chloroform-methanol 2:1. Cholesteryl ester concentrations were calculated by subtracting free from total cholesterol.
Lipoprotein characterization
VLDL and LDL composition was calculated by determining free and esterified cholesterol [14], protein by a modified Lowry method [16], and TG and phospholipids by enzymatic kits. VLDL apo B was selectively precipitated with isopropanol [17]. The number of constituent molecules of LDL was calculated on the basis of one apo B per particle with a molecular mass of 412000 kD[18]. The molecular weights were 885.4, 386.6, 645 and 734 for TG, free and esterified cholesterol, and phospholipids, respectively [19]. LDL diameters were calculated according to Van Heek et al [20]. HDL cholesterol was also determined according to Warnick et al, with a modification, which consisted of using 2 mol/L MgCl2 for precipitation of apo-B containing lipoproteins [13].
Lecithin Cholesterol Acyltransferase (LCAT) and Cholesterol Ester Transfer Protein (CETP) determinations in plasma
LCAT and CETP activities were determined according to Ogawa & Fielding [21]. Physiological CETP activity was determined without inhibiting LCAT activity by measuring the mass transfer of cholesterol ester between HDL and apo B containing lipoproteins. Samples were incubated at 37°C for 6 h in a shaking water bath and total and free plasma cholesterol and HDL cholesterol were measured. LCAT activity was determined by mass analysis of the decrease in plasma free cholesterol between 0 and 6 h at 37°C. Assays were carried out concurrently with measurements of CETP. Both of these methods have been well-standardized for guinea pig plasma [22].
Hepatic microsome isolation
Hepatic microsomes were isolated as described previously [8]. Briefly a microsomal fraction was isolated by two 25-min centrifugations at 10,000 g (JA-20 rotor, J2-21) followed by ultracentrifugation at 100,000 g in a Ti-50 rotor at 4°C for 1 hour. Microsomes were resuspended in the homogenization buffer and centrifuged for an additional hour at 100,000 g. After centrifugation, microsomal pellets were homogenized and stored at -70°C.
Hepatic HMG-CoA reductase assay
The activity of microsomal HMG-CoA reductase (E.C. 1.1.1.34) was measured in hepatic microsomes as described by Shapiro et al. [23]. HMG-CoA reductase activity was expressed as pmol of [14C] mevalonate produced per min per mg microsomal protein. Recoveries of [3H] mevalonate ranged from 60–90%.
Hepatic Acyl CoA Cholesteryl Acyltransferase (ACAT) activity
Hepatic ACAT (E.C. 2.3.1.26) activity was measured by the incorporation of [14C] oleoyl CoA in cholesteryl ester in hepatic microsomes by preincubating 0.8–1 mg of microsomal protein per assay with 84 g/L albumin and buffer for microsomal isolation [24]. Recoveries of [3H] cholesteryl oleate were between 70–90%.
Hepatic cholesterol 7α-hydroxylase activity
Cholesterol 7α-hydroxylase (E.C. 1.14.13.7) activity was measured according to the method modified by Jelinik et al [25]. [14C] cholesterol was used as a substrate and delivered as cholesterol-phosphatidylcholine liposomes (1:8 by weight) prepared by sonication. An NADPH-regenerating system (glucose-6-phosphate dehydrogenase, NADP, and glucose-6-phosphate) was included in the assay as a source of NADPH.
Statistical analysis
One-way analysis of variance (ANOVA) (SSPS for Windows version 12) was used to evaluate significant differences among groups in regards to plasma and hepatic lipids, LDL composition, hepatic enzyme activities and LCAT and CETP activities. The LSD post hoc test was used to evaluate the differences among groups. Data are presented as the mean ± SD. Differences were considered significant at P < 0.05.
Results
Plasma lipids and lipoproteins
No difference in weight gain overtime was observed in guinea pigs fed the different test diets (Fig 1), indicating that animals consumed comparable amounts of their respective test diets. After feeding the test diets for a period of four weeks, blood was isolated and plasma was analyzed for cholesterol and TG concentrations. The two doses of MTPi evaluated, low dose (LDI) and high dose (HDI) decreased plasma total cholesterol values significantly by 19.2% (P < 0.01) as compared to the control animals (Table 2). There was no significant difference between the two doses of MTPi tested. Atorvastatin, which was used as a positive control, led to a significantly robust decrease in plasma total cholesterol values of 46%, which was significantly different (P < 0.01) from the two MTPi doses used. Plasma TG values were also significantly lower in LDI (50.8%) and HDI (45.3%) when compared to their control counterparts whereas atorvastatin (AT) treatment resulted in the maximum decrease of plasma TG (Table 2).
Figure 1 Weight gain of guinea pigs treated with control, low dose, high dose of JTT-130 and atorvastatin.
Table 2 Plasma total cholesterol (TC), triglycerides (TG), VLDL-C, LDL-C and HDL-C of guinea pigs fed a control diet, low dose MTPi (LDI), high dose MTPi (HDI) or atorvastatin
Diets TC TG VLDL-C LDL-C HDL-C
(mg/dL)
Control (10) 146.9 ± 42.2a 135.8 ± 118.9a 8.1 ± 5.5ab 123.9 ± 43.5a 13.6 ± 4.5a
LDI (10) 116.9 ± 29.7b 66.7 ± 29.6b 3.7 ± 3.0a 93.3 ± 26.2b 18.5 ± 8.0a
HDI (9) 116.1 ± 23.3b 74.3 ± 31.7b 9.0 ± 8.6b 90.5 ± 28.8b 15.6 ± 7.1a
Atorvastatin (9) 76.5 ± 29.8c 49.4 ± 32.2c 2.9 ± 2.6a 59.4 ± 29.3c 14.1 ± 5.6a
1Data are presented as mean ± SD for the number of guinea pigs indicated in parenthesis. Numbers in a column with different superscripts are considered significantly different (P < 0.01) as determined by one way ANOVA and LSD as a post-hoc test
The changes in plasma TC values were mostly due to decreases in the cholesterol carried by LDL. LDL-C was also significantly decreased (24.7% & 26.9%) by the MTPi diets tested. There were no major differences between these two doses for plasma lipid parameters except for VLDL-C, which were significantly higher when compared to the low dose of the drug. No significant differences were observed for HDL-C values with MTPi or with AT (Table 2).
LDL size and composition
No significant effect of MTPi on the number of CE molecules or on the size of LDL particle was found with any of the two doses tested as compared to their control counterparts (Table 3). AT treatment reduced the number of esterified cholesterol molecules (45%) as well as decreased the size of LDL particle (30%) (Table 3).
Table 3 Number of molecules of cholesteryl ester (CE), LDL diameter and LCAT and CETP activities of guinea pigs fed a control diet, low dose MTPi (LDI), high dose MTPi (HDI) or atorvastatin.
Diets CE molecules LDL diameter LCAT CETP
nanometers (pmol/min.mg)
Control (n = 10) 1993 ± 422a 16.47 ± 3.63a 19.6 ± 11.0a 36.1 ± 12.5a
LDI (n= 10) 2072 ± 536a 16.67 ± 3.05a 14.4 ± 6.4a 31.0 ± 21.1a
HDI (n = 9) 2026 ± 132.5a 16.58 ± 7.85a 14.6 ± 8.5a 19.4 ± 13.0ab
Atorvastatin (n = 9) 1080 ± 1092b 11.52 ± 2.69b 13.4 ± 10.9a 12.8 ± 4.2b
1Data are presented as mean ± SD for the number of guinea pigs indicated in parenthesis. Numbers in a column with different superscripts are considered significantly different (P < 0.01) as determined by one way ANOVA and LSD as post hoc test.
LCAT and CETP activities
Table 3 also summarizes the activities of these two proteins, which play a major role in the intravascular processing of plasma cholesterol. There were no significant differences in LCAT activity when comparing MTPi or statin groups to the control group. However, HDI decreased the activity of CETP, which was comparable to the atorvastatin treated group (P < 0.01).
Hepatic lipids and enzymes
No significant changes were found in hepatic total cholesterol, free cholesterol, cholesteyl ester or TG values in any of the four treatments (Table 4). Results suggest that MTPi did not lead to lipid accumulation in the liver, as there was no significant difference between the two doses of MTPi tested with control or with atorvastatin group. Similarly no significant changes were observed in any of the three regulatory hepatic enzymes involved, namely CYP7, ACAT and HMG-CoA Reductase (Table 5).
Table 4 Hepatic total cholesterol (TC), free cholesterol (FC) cholesteryl ester (CE) and TG of guinea pigs fed a control diet, low dose MTPi (LDI), high dose MTPi (HDI) or atorvastatin1.
Diets TC FC CE TG
(mg/g)
Control (n = 10) 1.20 ± 0.53 1.00 ± 0.44 0.20 ± 0.23 16.2 ± 2.8
LDI (n = 10) 1.47 ± 0.40 1.24 ± 0.35 0.23 ± 0.11 13.9 ± 9.4
HDI (n = 9) 1.42 ± 0.66 0.98 ± 0.39 0.44 ± 0.58 13.5 ± 7.8
Atorvastatin (n = 9) 1.71 ± 0.76 1.31 ± 0.56 0.40 ± 0.39 18.8 ± 10.8
1Data are presented as mean ± SD for n = the number of guinea pigs indicated in parenthesis.
Table 5 Hepatic HMG-CoA reductase, ACAT and cholesterol 7α-hydroxylase activities (CYP7) guinea pigs fed a control diet, low dose MTPi (LDI), high dose MTPi (HDI) or atorvastatin1
Diets HMG-CoA Reductase ACAT CYP7
(pmol/min.mg)
Control 1.8 ± 0.8 0.8 ± 0.4 1.7 ± 1.8
LDI 2.7 ± 0.8 2.2 ± 0.9 1.2 ± 0.5
HDI 2.5 ± 1.4 3.5 ± 2.4 0.6 ± 0.7
Atorvastatin 3.4 ± 2.9 4.4 ± 3.1 0.7 ± 0.4
1Data are presented as mean ± SD for n = 3–7 guinea pigs per group.
Discussion
In this study we were able to demonstrate that JTT-130, the MTPi tested has the potential to decrease the prime risk factors of cardiovascular disease, namely plasma TG and LDL-C concentrations in guinea pigs. The novelty of the drug tested is that there was no significant accumulation of lipid in the liver as seen in some other studies done with other MTPi [26,27].
Drug treatment and hepatic lipids
Previous studies evaluating MTPi have shown increase in the lipid content of the liver [7,26,27]. Chandler et al [7] treated Hep-G2 cells with CP-346086, another MTPi, for a period of two weeks. They reported that in addition to decreasing plasma TC, LDL-C, VLDL-C and TG values, this treatment also increased hepatic and intestinal TG when the MTPi was administered with food and when it was dosed away from meals, only hepatic TG were influenced. In contrast, the major finding of this study is that the MTPi tested did not lead to any fat accumulation in the liver as confirmed by no significant changes found in the hepatic lipid content as compared to their control counterparts. The main reason for these differences between MTPi could be that the main target of JTT-130 was the intestine. Because of this, we speculate that due to MTP inhibition, less TG were transferred to the chylomicron particle being packaged in the intestine. As a result a lower concentration of TG was taken up by the hepatocytes through the chylomicron remnant. Thus the VLDL particles secreted from the liver had lower concentrations of TG molecules due to the major inhibitory effect of the MTPi in the intestine. Because there were no significant changes in hepatic cholesterol concentrations, we did not find any significant differences in hepatic enzyme activities. Similar to the study by Conde et al. [9] in atorvastatin treated guinea pigs with 0.015% atorvastatin, there were no significant differences in hepatic cholesterol concentrations when compared with a control group. However, significant differences in hepatic esterified cholesterol were observed when guinea pigs were treated with a higher dose of the statin (0.05%) [9].
Drug treatment and plasma lipids and lipoproteins
Abetalipoproteinemia, a genetic disorder characterized by low plasma cholesterol and TG levels, is caused by a functional deficiency of MTP. Absence of lipid transfer activity in the microsomes of abetalipoproteinemia patients established its pivotal function in lipoprotein assembly[1]. This finding led to the suggestion that MTP inhibition could be used as a possible lipid lowering therapy. Further evidence was obtained from a cell culture study in which researchers [28] proved that MTP is limiting in the production of apo B containing lipoproteins. Another study [6] further confirmed this finding using heterozygous MTP knockout mice which had 20% less plasma total cholesterol levels compared to wild type mice fed high fat diets; however, they did not find any significant differences in plasma TG concentrations. In our study, we have demonstrated that animals treated with MTPi had not only lower plasma TC and LDL-C but also significant reductions in plasma TG. It is possible that the VLDL particle secreted by the liver was more readily catabolized and therefore there was less conversion to LDL, which contributed to the hypocholesterolemic effects of the MTPi. Conde et al. [9] demonstrated that there was a significant reduction in plasma TG in guinea pigs treated with atorvastatin when compared to controls. This was partly explained by lower secretion of VLDL particles and by increases in the LDL receptor [9], which could have contributed to the faster removal of VLDL particles. A similar mechanism may have taken place with the MTPi. By blocking MTP, JTT-130 reduced the secretion of VLDL particles, and therefore, the formation of LDL in the plasma.
There was a dose response in CETP activity with JTT-130, and in addition, guinea pigs treated with atorvastatin exhibited decreased activity of this transfer protein. The main function of CETP is to contribute to the reverse cholesterol pathway by transferring cholesteryl ester from HDL to the apo B containing lipoproteins [29]. However, this action prolongs the residence time of CE in LDL and increases the possibility of its deposition in the arterial wall. Thus lower CETP activity has been associated with decreased atherogenesis in animal studies [30]. Therefore the lowering of CETP activity by drug treatment can be considered beneficial.
Results from this study indicate that JTT-130 has the potential to reduce the primary risk factors for coronary heart disease. While these results are quite promising, more studies are needed to clarify the possibility of adverse effects including steatorhea, fat malabsorption and fat-soluble vitamin absorption. Although the lipid lowering effects were not as pronounced as those observed with atorvastain, the doses of MTPi used in the current study were lower than the doses of atorvastatin. It is possible that using JTT-130 in combination with statins could reduce the wide array of adverse effects associated with reductase inhibitors [31]. This study also demonstrates that MTP inhibitor which is mainly targeted towards the intestine may open a new avenue for treatment of hyperlipidemic patients who are at high risk for cardiovascular diseases.
List of abbreviations used
ACAT: acyl CoA cholesteryl acyl transferase; AT: atorvastatin; CETP: cholesterol ester transfer protein; CHD: coronary heart disease; CYP7: cholesterol 7α-hydroxylase; HDi: high dose of the inhibitor; HDL-C: HDL cholesterol; HMG-CoA; 3-hydroxy-3methyl glutaryl Conezyme A; LCAT: lecithin cholesterol acyltransferase; LDi: low dose of the inhibitor; LDL-C: LDL cholesterol; MTP: microsomal transfer protein; MTPi: microsomal transfer protein inhibitor; NADP: nicotinamide adenenine dinucleotide phosphate; TG: triglycerides; VLDL: very low density lipoprotein.
Competing interests
Authors received funding from Akros Pharma Inc. (Princeton, NJ) to carry out the studies presented in this manuscript.
Authors' contributions
DA did the assays, wrote the manuscript and participated in the interpretation of data; KLW: assisted in the assays for plasma lipids, CETP and LCAT; TLZ: assisted in the determination of ACAT activity and participated in data interpretation, SS: assisted in taking care of guinea pigs, isolation of microsomes and data interpretation; MVJ assisted in the determination of CYP7 and in data interpretation and MLF designed the experiment, evaluated the results, interpreted the data and participated in manuscript preparation.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
These studies were supported by Akros Pharma Inc, Princeton, NJ
==== Refs
Hussain MM Iqbal J Anwar K Rava P Dai K Microsomal triglyceride transfer protein: a multifunctional protein Front Biosci 2003 500 506
National Cholesterol Education Program Expert Panel. Second report of the National Cholesterol Education Program (NCEP) Expert panel on Detection Evaluation and Treatment of High Blood Cholesterol in Adults; Adult Treatment Panel II Circulation 1994 89 1329 1445
Association AH Heart and Stroke Statistical Update American Heart Association: 2001 2001
Shepherd J Cobbe SM Ford I Isles CG Lorimer AR MacFarlane PW McKillop JH Packard CJ Prevention of coronary heart disease with pravastatin in men with hypercholesterolemia N Eng J Med 1995 333 1301 1307 10.1056/NEJM199511163332001
Sacks FM Pfeffer MA Moye LA Rouleau JL Rutherford JD Cole TG Brown L Warnica JW Arnold JMO Wun CC Davis BR Braunwald E The effect of pravastatin on coronary events after myocardial infarction in patients with average cholesterol levels N Eng J Med 1996 335 1001 1009 10.1056/NEJM199610033351401
Raabe M Veniant MM Sullivan MA Zlot CH Bjorkegren J Nielson LB Wong JS Hamilton RL Young SG Analysis of the role of microsomal triglyceride transfer protein in the liver of tissue-specific knockout mice J Clin Invest 1999 103 1287 1298 10225972
Chandler CE Wilder DE Pettini JL Savoy YE Petras SF Chang G Vincent J Harwood HJ CP-346086: an MTP inhibitor that lowers plasma cholesterol and triglycerides in experimental animals and in humans J Lipid Res 2003 44 1887 1901 12837854 10.1194/jlr.M300094-JLR200
West KL Fernandez ML Guinea pigs as models to study the hypercholesterolemic effects of drugs Cardiovasc Drug Rev 2004 22 55 70 14978518
Conde K Pineda G Newton RS Fernandez ML Hypocholesterolemic effects of 3-hydroxy-3-methylglutaryl coenzyme a (HMG-CoA) reductase inhibitors in the guinea pig Biochem Pharmacol 1999 58 1209 1219 10484079 10.1016/S0006-2952(99)00203-8
Fernandez ML Guinea pigs as models for cholesterol and lipoprotein metabolism J Nutr 2001 131 10 20 11208932
Lin ECK Fernandez ML McNamara DJ Dietary fat type and cholesterol quantity interact to affect cholesterol metabolism in guinea pigs J Nutr 1992 122 2019 2029 1527642
Redgrave TG Roberts DC West CE Separation of plasma lipoproteins by density-gradient ultracentrifugation Anal Biochem J 1975 65 42 49 10.1016/0003-2697(75)90488-1
Fernandez ML Wilson TA Conde K Vegara-Jimenez M Nicolosi RJ Hamsters and guinea pigs differ in their plasma lipoprotein cholesterol distribution when fed diets varying in animal protein, soluble fiber or cholesterol content J Nutr 1999 129 1323 1333 10395594
Allain CC Poon LC Chan CS Richard W Fu PC Enzymatic determination of total serum cholesterol Clin Chem 1974 20 47 475
Carr TP Anderssen CJ Rudel LL Enzymatic determination of triglycerides, free cholesterol and cholesterol in tissue lipid extracts Clin Biochem 1993 26 39 42 8448837 10.1016/0009-9120(93)90015-X
Markwell MK Haas S Bieber LL Tolbert NE A modification of the Lowry procedure to simplify protein determination in membrane and lipoprotein samples Anal Biochem J 1978 87 206 210 10.1016/0003-2697(78)90586-9
Homquist L Carlson K Carlson LA Comparison between the use of isopropanol and tetramethylurea for the solublization and quantitation of human serum very low density apolipoproteins Anal Biochem J 1987 149 423 426
Chapman JM Mills GL Ledford JH The distribution and partial characterization of the serum apolipoproteins in the guinea pig Biochem J 1975 149 423 36 170915
Conde K Vergara-Jimenesz M Krause BR Newton RS Fernandez ML Hypocholesterolemic actions of atorvastatin are associated with alterations on hepatic cholesterol metabolism and lipoprotein composition in the guinea pig J Lipid Res 1996 37 739 746 8732774
Van Heek M Zilversmit D Mechanisms of hypertriglyceridemia in the coconut oil/cholesterol-fed rabbits increased secretion and decreased catabolism of VLDL Arteriosclerosis Thromb 1991 11 918 927
Ogawa YFCJ Assay of cholesterol ester transfer activity and purification of a cholesterol ester transfer protein Meth Enzymol 1985 111 274 285 3897772
Fernandez ML Conde K Vergara-Jimenesz M Behr T Abdel-Fatta G Regulation of VLDL-LDL Apo B metabolism in guinea pigs by dietary soluble fiber Am J Clin Nutr 1997 65 814 822 9062534
Shapiro DL Imblum RL Rodwell VW Thin-layer chromatographic assay for the HMG-CoA reductase and mevalonic acid Anal Biochem J 1969 31 383 390 10.1016/0003-2697(69)90279-6
Smith JL de Jersey J pillay SP Hardie IR Hepatic acyl-CoA: cholesterol acyltransferase. Development of a standard assay and determination in patients with cholesterol gallstones Clin Chim Acta 1986 158 271 282 3769201 10.1016/0009-8981(86)90291-3
Jelinek DF Andersson S Slaughter CA Russel DW Cloning and regulation of cholesterol 7a-hydroxylase, the rate limiting enzyme in bile acid biosynthesis J Biol Chem 1990 265 8190 8197 2335522
Liao W Hui TY Young SG Davis RA Blocking microsomal triglyceride transfer protein interferes with apoB secretion without causing retention or stress in the ER J Lipid Res 2003 44 978 985 12588952 10.1194/jlr.M300020-JLR200
Shiomi M Ito T MTP inhibitor decreases plasma cholesterol levels in LDL receptor-deficient WHHL rabbits by lowering the VLDL secretion Eur J Pharmacol 2001 431 127 131 11716851 10.1016/S0014-2999(01)01419-4
Jamil H Gordon DA Eustice DC Brooks CM Dickson JK JrChen Y Ricci B Chu CH Harrity TW Ciosek CP Biller SA JrGregg RE Wetterau JR An inhibitor of the microsomal triglyceride transfer protein inhibits apoB secretion from HepG2 cells Proc Natl Acad Sci USA 1996 93 11991 11995 8876250 10.1073/pnas.93.21.11991
Barter P CETP and atherosclerosis Arterioscler Thromb Vasc Biol 2000 20 2029 2031 10978244
Le Goff W Guerin M Chapman MJ Pharmacological modulation of cholesteryl ester transfer protein, a new therapeutic target in atherogenic dyslipidemia Pharmacol Ther 2004 101 17 38 14729390 10.1016/j.pharmthera.2003.10.001
Ansell BJ Rationale for combination therapy with statin drugs in the treatment of dyslipidemia Curr Atheroscler Rep 2005 7 29 33 15683599
|
16188041
|
PMC1262702
|
CC BY
|
2021-01-04 16:30:07
|
no
|
BMC Cardiovasc Disord. 2005 Sep 27; 5:31
|
latin-1
|
BMC Cardiovasc Disord
| 2,005 |
10.1186/1471-2261-5-31
|
oa_comm
|
==== Front
BMC Dev BiolBMC Developmental Biology1471-213XBioMed Central London 1471-213X-5-201618536210.1186/1471-213X-5-20Research ArticleThe roles of Bcl-xL in modulating apoptosis during development of Xenopus laevis Johnston Jillian [email protected] Robert [email protected] Maria [email protected] Sarah [email protected] Leon W [email protected] Genes and Development Research Group, Department of Biochemistry & Molecular Biology, University of Calgary, Calgary, Alberta, Canada, T2N 4N12 Genes and Development Research Group, Department of Cell Biology & Anatomy, University of Calgary, Calgary, Alberta, Canada, T2N 4N13 Universidad Nacional Autónoma de México, Centro de Ciencias de la, Atmósfera, Laboratorio de Citogenética Ambiental, Circuito Exterior S/N, Col. Coyoacán, Ciudad Universitaria, C.P.0451, México, Distrito Federal2005 26 9 2005 5 20 20 25 2 2005 26 9 2005 Copyright © 2005 Johnston et al; licensee BioMed Central Ltd.2005Johnston et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Apoptosis is a common and essential aspect of development. It is particularly prevalent in the central nervous system and during remodelling processes such as formation of the digits and in amphibian metamorphosis. Apoptosis, which is dependent upon a balance between pro- and anti-apoptotic factors, also enables the embryo to rid itself of cells damaged by gamma irradiation. In this study, the roles of the anti-apoptotic factor Bcl-xL in protecting cells from apoptosis were examined in Xenopus laevis embryos using transgenesis to overexpress the XR11 gene, which encodes Bcl-xL. The effects on developmental, thyroid hormone-induced and γ-radiation-induced apoptosis in embryos were examined in these transgenic animals.
Results
Apoptosis was abrogated in XR11 transgenic embryos. However, the transgene did not prevent the apoptotic response of tadpoles to thyroid hormone during metamorphosis. Post-metamorphic XR11 frogs were reared to sexual maturity, thus allowing us to produce second-generation embryos and enabling us to distinguish between the maternal and zygotic contributions of Bcl-xL to the γ-radiation apoptotic response. Wild-type embryos irradiated before the mid-blastula transition (MBT) underwent normal cell division until reaching the MBT, after which they underwent massive, catastrophic apoptosis. Over-expression of Bcl-xL derived from XR11 females, but not males, provided partial protection from apoptosis. Maternal expression of XR11 was also sufficient to abrogate apoptosis triggered by post-MBT γ-radiation. Tolerance to post-MBT γ-radiation from zygotically-derived XR11 was acquired gradually after the MBT in spite of abundant XR11 protein synthesis.
Conclusion
Our data suggest that Bcl-xL is an effective counterbalance to proapoptotic factors during embryonic development but has no apparent effect on the thyroid hormone-induced apoptosis that occurs during metamorphosis. Furthermore, post-MBT apoptosis triggered by irradiation before the MBT could only be restrained by maternal expression of Bcl-xL. Although maternal expression of XR11 was sufficient to abrogate apoptosis triggered by post-MBT γ-radiation, radiation tolerance from zygotically-derived XR11 was acquired gradually, indicating that synthesis of XR11 protein is not sufficient to prevent apoptosis. Thus, repression of radiation-induced apoptosis by overexpression of Bcl-xL during embryonic development depends upon the timing of its expression and post-translational events that enable the protein to become effective.
==== Body
Background
Cell death is an essential and integral aspect of embryonic development (for reviews, see [1-3]). The genome is programmed to eliminate certain cells or groups of cells at particular times during development by a process called "programmed cell death" (PCD) [4-6]. Well-known examples of PCD during vertebrate development include interdigital cell death and regression of the tail and gills during amphibian metamorphosis [3]. Considerable PCD also occurs in the central nervous system (CNS; for review, see [7]), where as much as 85% of some neuronal populations may be lost by PCD during development [8-11]. One form of PCD is apoptosis, which is characterized by membrane blebbing, nuclear and cytoplasmic shrinkage, chromatin condensation and DNA fragmentation [12]. The detection of DNA fragmentation in dying cells utilizing the TUNEL (TdT-mediated dUTP nick end labeling) assay provides the basis for a sensitive and specific in situ measure of apoptosis.
The frog Xenopus laevis provides an excellent model system for studying apoptosis during embryonic development. The TUNEL assay has been used on whole-mounts and tissue sections of Xenopus embryos to monitor spontaneous developmental apoptosis [13,14]. Apoptosis also provides the means for the embryo to rid itself of cells damaged by DNA-damaging agents, such as gamma radiation [15]. Gamma radiation during embryonic development is a threat to the viability of the embryo and to its potential to develop into a healthy, functional adult organism. Embryos have two alternative mechanisms to prevent developmental abnormalities caused by irradiation: elimination of damaged cells through apoptosis or repair of the damaged DNA. Xenopus embryos undergo a dramatic transition in their response to ionizing radiation coincident with the onset of gastrulation. Embryos irradiated before gastrulation continue to cleave normally and undergo an abrupt, catastrophic, comprehensive and synchronous apoptosis of DNA-damaged cells after the mid-blastula transition (MBT) using a maternally-derived mechanism that does not require transcription [15-20]. In contrast, cells in embryos irradiated after gastrulation reportedly do not undergo apoptosis [17,19,20].
The induction of apoptosis is subject to both pro- and anti-apoptotic cellular factors and is executed by cysteine proteases called caspases after their activation from the proenzyme state. The activation of caspases can be triggered by cytochrome c release from the intermembrane spaces of mitochondria through mitochondrial outer membrane permeabilization [21-24]. Release of cytochrome c, in turn, is regulated by members of the Bcl-2 family (for reviews, see [23,25,26]). Members of the Bcl-2 family serve a key regulatory role in apoptosis, acting as either pro- or anti-apoptotic factors. Thus, the relative levels of members with opposing functions can determine whether cells live or die [27] (for review, see [26]). One of the anti-apoptotic members of the Bcl-2 family is Bcl-xL, which functions to inhibit apoptosis in situations in which Bcl-2 does not suffice [28]. For example, bcl-2 knockout mice are viable as embryos [29,30] whereas bcl-xL knockout mice die as embryos that exhibit massive cell death in the nervous and hematopoietic systems [31]. Over-expression studies conducted both in vitro and in vivo in mice indicate that Bcl-xL is also sufficient to promote survival of embryonic neurons [32-34]. Similar results have been obtained with chick embryonic neurons in vitro [35]. Members of the Bcl family also modulate the apoptogenic effects of DNA-damaging agents. For example, lymphoid cells from bcl-2 null mice are hypersensitive to irradiation in vitro [29,30]. Conversely, over-expression of bcl-2 reduces the rate of apoptosis of γ-irradiated hematopoietic cells both in vitro and in vivo [36,37].
Cruz-Reyes and Tata [38]cloned two bcl-related genes in Xenopus laevis, which they designated as XR1 and XR11. XR11 [GenBank:X82461], which most closely resembles Bcl-xL, is expressed as early as stage 12 (gastrula) and continues to be expressed throughout embryonic, larval and adult life. Over-expression of XR11 in Rat-1 fibroblasts protected the cells from apoptosis, thus demonstrating that XR11 is an anti-apoptotic member of the Bcl family. Misexpression of human Bcl-2 or XR11 in Xenopus embryos has been shown to impede developmental- or chemically-induced apoptosis, respectively, during early Xenopus laevis embryonic development [17,39]. The transgenesis technique developed by Kroll and Amaya [40] provides the means to study the effects of genes on later developmental events and to examine the effects of the transgenes on subsequent generations. For example, Coen et al. [41] employed a neuron-specific promoter driving XR11 over-expression to demonstrate selective protection of Rohon-Beard cells during metamorphosis. In this report, we have used a constitutive promoter to examine the effects of generalized XR11 over-expression throughout Xenopus development and to examine the respective roles of maternal and zygotic XR11 in protection against γ-radiation-induced apoptosis during development.
Results
Over-expression of XR11 dramatically reduced spontaneous developmental apoptosis in embryos
Over-expression of XR11 in nuclear transplant embryos was achieved by utilizing a plasmid containing colinear CMVXR11 and CARGFP genes. CMV is a strong constitutive promoter in Xenopus, whereas XCAR promotes gene expression in muscle and cardiac tissue. Thus, XR11 should be expressed universally, whereas expression of the colinear GFP reporter gene in muscle and heart was used to select XR11 transgenics. To determine whether the over-expression of XR11 protected embryonic cells from apoptosis, we compared the incidence of TUNEL-positive nuclei between XR11 and control transgenic whole-mount and sectioned tailbud-stage embryos. Yeo and Gautier [39] demonstrated that the TUNEL assay is effective and specific for detecting apoptosis in Xenopus embryos. Examination of control whole-mount tailbud-stage embryos (Fig. 1A, C) showed that darkly labelled apoptotic nuclei are particularly prevalent in the heads, whereas XR11 transgenics (Fig. 1B, D) exhibited little or no detectable apoptosis. Examination of cross-sections (Fig. 2A) revealed that TUNEL-positive nuclei were most prevalent in the retinas and brains, whereas XR11 transgenic embryos lack TUNEL-positive nuclei. Counts of total TUNEL-positive nuclei in sections were made at stages 28, 33/34, 37 and 41 (Fig. 2B). In GFP controls, TUNEL-positive cells peaked at stage 33/34, whereas XR11 overexpressing embryos had nominal numbers of TUNEL-positive nuclei at all stages. For example, over-expression of XR11 at stage 33/34 reduced apoptosis by 95%. Student's t-tests indicate that the differences in the numbers of TUNEL-positive nuclei between the XR11 and GFP embryos are highly significant (p = 0.001 for stage 41 and 0.001 > p > 0.000 for stages 28, 33/34 and 37).
Figure 1 TUNEL assays on whole-mount tailbud-stage embryos at stages 28(A, B) and 33/34 (C, D). A, C: GFP transgenic embryos B, D: XR11 transgenic embryos.
Figure 2 Comparisons of TUNEL-positive nuclei between GFP and XR11 transgenic embryos during embryonic development. A. Cross-sections of embryos showing the distribution of TUNEL-positive nuclei at the level of the eye at stage 33. DAPI (blue) and TUNEL (red) images have been superimposed. Left: GFP transgenic embryo. R, retina; B, brain. Right: XR11 transgenic embryo. Dorsal is at the top. B. Mean numbers of TUNEL-positive nuclei in sections at stages 28, 33/34, 37/38 and 41. Error bars indicate standard error of the mean. Means were obtained by examining between 14 and 18 embryos for each category. A minimum of 4 and a maximum of 14 sections were examined for each embryo. Pooled data from multiple experiments.
Over-expression of XR11 did not impede apoptosis that occurs during normal or thyroid hormone-induced metamorphosis
If pro-apoptotic factors facilitate apoptosis during metamorphosis of tadpoles into frogs, we would expect over-expression of the anti-apoptotic XR11 to impede metamorphosis. However, XR11 transgenic tadpoles developed as apparently normal larvae that underwent metamorphosis, losing their transient larval structures, such as gills and tails (Fig. 3A, B). The loss of gills is evident by the slimming of the body profile in the pharyngeal region (arrow in A). The remnant of the tail is indicated by the arrow in B. Loss of these structures occurs via apoptosis during metamorphosis in response to thyroid hormone [42].
Figure 3 Metamorphosis of XR11 transgenic tadpoles and responses to exogenous thyroid hormone (T3). A, B. An XR11 transgenic froglet undergoing metamorphosis. C-J. Responses of one-week old (approximately stage 45) transgenic tadpoles to thyroid hormone (T3). XR11 transgenic tadpoles (G-J;) display the same gross responses to 5 days exposure to as do GFP transgenics (C-F). C, E, G and I are live tadpoles, whereas D, F, H and J are fixed tadpoles stained with Alcian blue to reveal the skeleton and the gill apparatus. K-R. The apoptotic response of one-week old transgenic tadpoles (approximately stage 45) to T3 was assessed in cross-sections through the velum. DAPI staining (K, M, O and Q) reveals nuclei, whereas the TUNEL assay (L, N, P and R) distinguishes apoptotic nuclei. The velum of the GFP transgenic tadpole (M, N) has begun dissociating in response to T3 treatment. S. Demonstration of XR11 over-expression in tadpoles by immunoprecipitation. This experiment was repeated once.
To address more specifically whether apoptosis can be triggered by thyroid hormone in XR11 transgenics, we used the metamorphosis induction assay described by Tata [43] and Huang et al. [44]. One-week tadpoles were exposed to triiodothyronine (T3) for five days. As shown in Figure 3C–J, tadpoles transgenic for XR11 displayed the same gross responses to thyroid hormone, including gill resorption, as shown by the GFP transgenics with no discernable delay. The gill apparatus, which is visible in the absence of T3 (arrow in H), regresses after T3 treatment in both cases. Meckel's cartilage (the lower jaw) shows a dramatic beak-like reshaping (arrows in F and J).
To determine whether apoptosis was occurring equally in the GFP and XR11 transgenic tadpoles, we focused on the ventral velum, a portion of the larval filter feeding apparatus that has a chevron shape in cross section before it dissociates through apoptosis during metamorphosis [45]. As shown in Figure 3K–R, TUNEL-positive nuclei, which are rare in the velums in the absence of T3, were prevalent in the velums of both T3-treated GFP and XR11 transgenic tadpoles. These results indicate that apoptosis that occurred in response to thyroid hormone was unaffected by over-expression of XR11. One possible explanation for the extensive apoptosis in the XR11 transgenic tadpoles is that the XR11 protein was no longer over-expressed. However, as shown in Figure 3S, immunoblotting with an antibody to XR11 confirmed that the 22 kDa XR11 protein was over-expressed in the transgenic tadpoles. Proteins were extracted from individual tadpoles at approximately stage 45, separated by electrophoresis, blotted and probed as described in Methods. These tadpoles had been treated with triiodothyronine (T3). Similar results were obtained with samples from untreated tadpoles (data not shown). The variation in the GFP signal reflected the intensity of GFP fluorescence in the live tadpoles, presumably reflecting differences in the number of integrated copies of GFP. The 22 kDa XR11 signal is considerably stronger in the XR11 transgenic sample than in the GFP transgenic sample, reflecting XR11 transgene expression.
XR11 transgenic frogs are fertile and transmit the transgene
The failure of XR11 over-expression to prevent metamorphosis enabled us to rear the resultant frogs to sexual maturity. We induced ovulation in adult females and used the sperm from adult males to obtain second-generation embryos overexpressing XR11. The results of representative crosses are shown in Table 1. Individual parents were designated by number (i.e., XR11-01, XR11-02). Crossing the XR11-01 male with a wild-type female gave a 1:1 ratio of transgene positive to transgene negative progeny, suggesting that his genome had integrated the transgene at a single site. Crossing him with the XR11-01 female gave a 3:1 ratio, a result consistent with a single transgene integration site. However, crossing him with the XR11-02 female yielded a 7:1 ratio, suggesting that this female had two transgene integration sites.
Table 1 Representative Crosses Illustrating Inheritance of the XR11 Transgene
Cross GFP positive GFP negative Ratio (χ2)
XR11-01 male × WT-01 female 197 199 1:1 (p < 0.9)
XR11-01 male × XR11-01 female 319 90 3:1 (p < 0.1)
XR11-01 male × XR11-02 female 321 46 7:1 (p < 0.9)
Over-expression of XR11 curtailed apoptosis induced by irradiation
The ability to produce second-generation transgenics gave us the opportunity to compare the effects of maternal and zygotic expression of the XR11 transgene on susceptibility of embryos to γ-radiation-induced apoptosis. Consistent with the literature [15], wild-type embryos subjected to -radiation before the MBT did not undergo apoptosis before the MBT (data not shown). The appearance of TUNEL-positive nuclei was delayed until after the MBT and was radiation dosage-dependent (Fig. 4A–F). After exposure to 10 Gy (Fig. 4B) blastopores were not evident, and large sectors containing TUNEL-positive nuclei were seen. More extensive damage was seen after exposure to 20 Gy (Fig. 4C). TUNEL-positive nuclei were pervasive. Dissociation of the embryos was so extensive that the embryos often fragmented during the TUNEL procedure.
Figure 4 Maternal and paternal expression of the XR11 transgene. A-F. Maternal expression of the XR11 transgene reduces the extent of apoptosis in embryos exposed to γ-radiation at stage 6–6.5 and examined 12 hours after irradiation by whole-mount TUNEL assays. A-C, Wild-type embryos. A, Representative control embryos. Arrow indicates the blastopore lip. No evidence of apoptosis is apparent. Embryos that had been exposed to 10 Gy are shown in B. More extensive damage is seen after exposure to 20 Gy (C). D-F, TUNEL assays of progeny of fertilization of XR11 eggs by wild-type sperm after exposure to either 10 Gy (E) or 20 Gy (F). This experiment was conducted three times using eggs from four XR11 females. G-L. Paternal expression of the XR11 transgene does not protect embryos from apoptosis induced by γ-radiation before the MBT. Representative whole-mount TUNEL assays comparing the effects of γ-radiation at stage 6–6.5 on wild-type embryos (G-I) and progeny of fertilization of wild-type eggs by XR11 sperm (J-L). Embryos were fixed for TUNEL assay 12 hours after irradiation. G, J. Controls. H, K. 10 Gy. I, L. 20 Gy. This experiment was conducted twice using sperm from two XR11 males.
In contrast to the results with wild-type embryos, embryos produced by fertilization of eggs from XR11 females with wild-type sperm showed evidence of partial protection from γ-radiation (Fig. 4E, F). Reduced damage from radiation was always observed after exposure to 10 Gy, but protection from 20 Gy was variable from experiment to experiment, suggesting that overexpression of XR11 was not always sufficient to overcome the effects of 20 Gy.
To investigate whether protection from radiation-induced damage was due to maternal or zygotic expression of the transgene, we irradiated embryos derived from fertilizing eggs of wild-type females with sperm from XR11 males and compared them to wild-type embryos. If zygotic expression of the transgene could partially overcome the effects of pre-MBT γ-radiation, we would have expected to see a subset of embryos with few or no TUNEL-positive nuclei that resembled embryos derived from eggs of XR11 females. Instead, all embryos resembled irradiated wild-type embryos (compare Fig. 4J–L to Fig. 4G–I). Extensive TUNEL-positive nuclei are seen in both groups at both doses. Furthermore, the embryos exposed to 20 Gy fragmented during the TUNEL procedure. These results are consistent with a requirement for a maternal source of radiation protection at these early stages.
According to the literature, embryos irradiated after the MBT do not undergo apoptosis [17,19,20]. Indeed, embryos that were irradiated after the MBT (stage 11.5) and examined 12 hours later did not show outward signs of radiation damage (data not shown). However, TUNEL assays showed that irradiation enhanced apoptosis in wild-type embryos, predominantly in the brain, eyes, pharynx and tail bud. Hence, we sought to determine whether the maternal effect of the XR11 transgene was sustained after the MBT and whether zygotic expression of the transgene could also confer radiation tolerance. To do this, we segregated the XR11 transgenic embryos of over-expressing XR11 females from their non-transgenic siblings based upon their expression of the GFP reporter gene under control of the cardiac actin promoter. The GFP negative (GFP-) embryos had maternal expression of the XR11 transgene but no zygotic expression, whereas the GFP positive (GFP+) embryos had both maternal and zygotic XR11 transgene expression. Results show that maternal expression of the XR11 transgene protected embryos irradiated at stage 11.5 even if they did not express zygotic XR11. As shown in Figure 5A–F, neither the GFP- nor the GFP+ embryos had extensive apoptosis after irradiation, although some GFP- embryos demonstrated some apoptosis in the hindbrain region, but not in the tail tip (see Fig. 5C). When embryos were irradiated during the mid-30 stages (late tail bud) and examined 12 hours later, the protection afforded by maternal expression of XR11 had diminished significantly in the head and pharynx, although the tail tips were still protected (Fig. 5K, L, O and 5P). However, zygotic expression of XR11 was very effective in conferring global protection from apoptosis (Fig. 5M, N, Q and 5R).
Figure 5 Representative TUNEL assays showing (A-F) the sustained radiation tolerance of early post-MBT embryos and (G-R) the effects of zygotic transgene expression in late post-MBT embryos derived from eggs of XR11 females. (A-F). Embryos were irradiated at stage 11.5 and fixed for TUNEL assay 12 hours later. A-C. GFP-negative embryos (lacking the XR11 transgene). A, Control. B, 10 Gy. C, 20 Gy. D-F. GFP-positive embryos (containing the XR11 transgene). D, Control. E, 10 Gy. F, 20 Gy. This experiment was conducted four times using eggs from six XR11 females. G-R. Embryos were irradiated in the mid-30 stages and fixed for TUNEL assay 12 hours later. G, H, K, L, O, P. GFP-negative embryos (lacking the XR11 transgene). G, H. Control. Small numbers of TUNEL-positive nuclei are evidence of spontaneous developmental apoptosis. K, L. 10 Gy. O, P. 20 Gy. I, J, M, N, Q, R. GFP-positive embryos (containing the XR11 transgene). I, J. Control. M, N. 10 Gy. The ventral pigmentation in M is not due to TUNEL-positive nuclei. Q, R. 20 Gy. This experiment was conducted once using eggs from two XR11 females.
To examine the effects of zygotic XR11 transgene expression in the absence of maternal expression, we irradiated stage 11.5 embryos derived from fertilizing eggs from wild-type females with sperm from XR11 males. Progeny were scored 12 hours later for GFP expression directed by the cardiac actin promoter to confirm the presence of the transgene. As shown in Figure 6A–F, there was extensive apoptosis in both the heads, pharynges and tail tips of irradiated neurula-stage embryos and no discernable difference in the levels of apoptosis observed between the GFP-positive and GFP-negative embryos. Embryos that received 20 Gy radiation (Figs. 6C, F) are stunted with blunt heads, indicative of extensive developmental defects. These results suggest that zygotic expression of the transgene had little effect in preventing apoptosis at this stage. However, when sibling embryos were irradiated in the mid-30 stages and examined 12 hours later, the transgene conferred global protection from apoptosis (Fig. 6G–L). Note particularly the extensive apoptosis that is evident in the otic capsules (arrows in Fig. 6H) of irradiated embryos lacking the transgene (Figs. 6H, I).
Figure 6 Representative TUNEL assays showing the effects of zygotic XR11 expression on radiation tolerance of (A-F) early post-MBT embryos and (G-L) late post-MBT embryos derived from eggs of wild-type females. A-F. Embryos were irradiated at stage 11.5 and fixed for TUNEL assay 12 hours later. A-C. GFP-negative embryos. A. Control. B. 10 Gy. C. 20 Gy. D-F. GFP-positive embryos (containing the XR11 transgene). D. Control. E, 10 Gy. F. 20 Gy. This experiment was conducted three times using sperm from five XR11 males. G-L. Embryos were irradiated in the mid-30 stages and fixed for TUNEL assay 12 hours later. G-I. GFP-negative embryos. G. Control. H. 10 Gy. I. 20 Gy. J-L. GFP-positive embryos (containing the XR11 transgene). J. Control. K. 10 Gy. L. 20 Gy. This experiment was conducted once using sperm from two XR11 males. The experiment could not be repeated due to the unavailability of additional XR11 males.
The TUNEL data presented above indicate that: (1) maternal expression of the XR11 transgene reduces the amount of apoptosis induced by γ-radiation during pre-MBT and early post-MBT stages, (2) protection from apoptosis by maternal transgene expression is sustained in the tail tip after tissues in the head become susceptible to apoptosis (irradiation in the mid-30 stages) and (3) radiation tolerance is acquired gradually after the MBT and is sufficient to provide complete protection from apoptosis to embryos irradiated in the mid-30 stages of development.
To correlate transgene expression with repression of radiation-induced apoptosis, we monitored XR11 RNA expression and protein levels (Fig. 7). Expression of XR11 at the RNA level in embryos was monitored by RT-PCR with primers designed against the XR11 sequence. RT-PCR was performed on first strand cDNA synthesized from RNA isolated from pre-MBT (stage 7) and post MBT (stage 10.5 and mid-30 stages) wild-type embryos and embryos derived from fertilizing eggs of two different XR11 females with wild-type sperm and the fertilization of wild-type eggs with XR11 sperm. As shown in Figure 7A, we observed maternal XR11 transcripts in pre-MBT embryos derived from transgenic females. On the contrary, there was no evidence for expression of the XR11 transgene in embryos derived from wild-type eggs fertilized by XR11 sperm. After the MBT, XR11 mRNA derived from both maternal and paternal transgenes was evident. Immunoblots (Fig. 7B) indicated the presence of maternally-derived XR11 protein and the accumulation of XR11 protein from either maternal or paternal transgene expression after the MBT.
Figure 7 Maternal and zygotic XR11 RNA expression and protein levels. A. Over-expression of XR11 at the RNA level. RT-PCR of the constitutively expressed elongation factor 1-alpha (EF1α) was conducted in parallel as a control. This experiment was conducted once. B. XR11 protein levels in pre- and post-MBT embryos were demonstrated by Western blot analysis. Because the XR11 protein is membrane-bound, it is difficult to separate it from yolk, which is abundant in embryos. The presence of yolk during electrophoresis results in wavy bands in the gel. This experiment was repeated once.
Discussion
The application of technology to modify Xenopus genetically has created opportunities to conduct functional analyses that are not readily achieved by other means. In this study, we have used XR11 transgenics to examine the role of Bcl-xL in suppressing apoptosis during embryonic development, thyroid-hormone-induced metamorphosis and in response to γ-radiation. Our ability to raise XR11 transgenic frogs to maturity has also given us the opportunity to distinguish between maternal and zygotic contributions of Bcl-xL to radiation protection.
This study supports the growing evidence in the literature that Bcl-xL is sufficient to promote neuronal survival in mice and chicken embryos [32-35]. Our experiments showed that the XR11 transgene could eliminate most apoptosis in the developing Xenopus nervous system, including the brain and eyes. This result is consistent with a model in which the execution of developmental apoptosis is dependent upon a balance between pro-and anti-apoptotic factors with Bcl-xL being one of the anti-apoptotic factors. In spite of the abatement of developmental apoptosis in XR11 transgenic embryos, there were no apparent gross developmental consequences; the embryos developed into phenotypically normal tadpoles. This result suggests that developmental apoptosis is not essential for development of vital organ systems, notably the central nervous system, and that embryos can adapt to an overabundance of neurons. However, detailed morphometric studies would be necessary to confirm this conclusion. Yeo and Gautier [39] previously showed that injection of human bcl-2 mRNA abrogated programmed cell death during early Xenopus development, resulting in an expansion of neural tissue and variations in the patterns of expression of a number of genes expressed during early neurogenesis. We did not examine gene expression patterns in our embryos. However, if such alterations did occur, they had no apparent functional consequences.
Contrary to expectation, metamorphosis of XR11 transgenic tadpoles was normal, as was the response to exogenous thyroid hormone. If over-expression of Bcl-xL had hindered metamorphosis, it would have indicated that thyroid hormone triggers apoptosis by shifting the balance between pro-apoptotic factors and Bcl-xL. However, this proved not to be the case, even though we found that Bcl-xL was still misexpressed at metamorphosis. One possibility is that other members of the Bcl-2 family counter apoptosis during these later stages. Another possibility for the failure of XR11 transgenesis to prevent apoptosis during metamorphosis is that expression of the transgene was insufficient to shift the balance between pro-and anti-apoptotic factors sufficiently to prevent apoptosis. It is also possible that subtle changes in the timing, rate or cellular selectivity of metamorphosis did occur that we failed to detect. For example, Coen et al. [41] demonstrated selective protection of Rohon-Beard neurons in XR11 over-expressing tadpoles during metamorphosis. We did not assess cell-specific neoronal survival during metamorphosis.
Upon reaching sexual maturity, both male and female frogs were fertile. Crosses of transgenic frogs indicated that most had small numbers of transgene integration sites. This result was consistent with a previous transgene transmission study [46]. The ability to produce second generation transgenic XR11 embryos provided us with the opportunity to examine both the maternal and zygotic contributions of Bcl-xL to the apoptotic response to γ-radiation. This provided evidence for a significant role for apoptotic factor balance in determining whether cells that had received damaging dosages of radiation during development were eliminated. Other Bcl-2 family members may carry out similar roles in that microinjection of human Bcl-2 mRNA into Xenopus embryos is effective in inhibiting irradiation-induced apoptosis [16].
Expression of the XR11 transgene had a profound effect on whether cells of irradiated embryos underwent apoptosis. Furthermore, the application of transgenics allowed us to demonstrate that this effect depended upon the source of the over-expressed protein. Over-expression of Bcl-xL derived from XR11 females provided partial protection from the normal massive apoptosis that occurs in irradiated pre-MBT wild-type embryos after reaching the MBT. However, if the XR11 gene was derived from XR11 males, no protection was seen. This result demonstrates conclusively that apoptosis triggered by irradiation before the MBT can be restrained by maternal expression of Bcl-xL and that zygotic expression of Bcl-xL is ineffective in overcoming the effects of prior irradiation.
Irradiation soon after the MBT (at stage 11.5) had much more subtle effects on embryos. In wild-type embryos, apoptosis was enhanced in the brain, eyes, pharynx and tail tip. However, maternal expression of XR11 was sufficient to abrogate apoptosis. Interestingly, zygotic expression of XR11 at this time was not effective in preventing apoptosis even though XR11 protein was abundant at the time of irradiation. The failure of zygotically-derived XR11 to provide protection from apoptosis implies that synthesis of XR11 protein is not sufficient to prevent apoptosis.
Embryos irradiated during the mid-30 stages of development revealed that the maternal effect conferring protection from radiation-induced apoptosis had diminished significantly in the head and pharynx but persisted in the tails. It will be interesting to learn whether the difference between the heads and tails is due to preferential distribution or stability of maternal XR11 protein itself or is due to auxiliary factors that influence XR11 protein function. Bcl-xL and other members of the Bcl-2 family are known to be subject to post-translational regulation by such events as phosphorylation, proteolysis, cleavage, protein-protein interactions and subcellular localization [47,48]. Intriguingly, at these later stages, zygotic XR11 protein conferred global radiation protection. Thus, the initial lack of functionality of zygotic XR11 protein that we observed after the MBT had been overcome.
Biologically, the responses to irradiation after the MBT are probably more significant than the universal, catastrophic apoptosis that occurs in embryos irradiated before the MBT. Failure to eliminate cells mutagenized by irradiation would result in retention of cells that may have functional deficiencies and have the potential to become malignant due to genetic damage. Thus, the mechanisms that control apoptosis in response to DNA damage have important adaptive significance. Radiotherapy and a number of chemicals have been shown to damage DNA and induce apoptosis in mammals; this property has been exploited in the clinic to induce apoptosis of cancer cells. Resistance to apoptosis has important implications for the efficacy of chemotherapy and radiotherapy because tumor cells that overexpress either bcl-2 or bcl-xL may be resistant to either mode of cancer therapy [37,49-52]. On the other hand, their over-expression in non-tumor cells could enhance the ability of these cells to survive the otherwise lethal effects of radiation and chemotherapy.
Our results indicate that synthesis of Bcl-xL is not sufficient to confer radiation protection; rather, it belatedly acquires functionality. Possible explanations for the failure of Bcl-xL synthesized after the MBT to protect against apoptosis could include the absence of posttranslational modifications to the protein or the lack of a cooperative factor that is necessary for the protein to function effectively. The presence of this mechanism that confers functionality on Bcl-xL protein before the MBT and its absence soon after the MBT may provide an assay system that would facilitate its identification. The discovery of a mechanism that could control whether cells undergo apoptosis after exposure to an apoptogenic stimulus could have important implications for improving the efficacy of chemotherapy and radiotherapy.
Conclusion
Our results are consistent with a model in which the execution of developmental apoptosis in embryos is dependent upon a balance between pro-and anti-apoptotic factors with Bcl-xL being one of the anti-apoptotic factors. The drastic reduction of apoptosis in XR11 transgenic embryos had no apparent functional consequences. Metamorphosis of transgenic tadpoles was apparently normal, as was the response to exogenous thyroid hormone. The ability to produce second-generation transgenic XR11 embryos enabled us to examine both the maternal and zygotic contributions of Bcl-xL to the apoptotic response to γ-radiation. Post-MBT apoptosis that was triggered by irradiation before the MBT was restrained by maternal expression of Bcl-xL. However, zygotic expression of Bcl-xL was ineffective in overcoming the effects of prior irradiation. Maternal expression of XR11 was also sufficient to abrogate apoptosis triggered by post-MBT γ-radiation. However, zygotic expression of XR11 at this time was not effective in preventing apoptosis even though XR11 protein was abundant.
Methods
Transgenesis
Adult Xenopus laevis were maintained in compliance with the University of Calgary guidelines for animal care. Our transgenesis protocol is based upon the technique described by Sparrow[53], which is a simplification of the procedure described by Kroll and Amaya [40]. The simplified technique eliminates the need to use egg extract and restriction enzyme to facilitate transgenesis. We have found no significant difference in the frequencies of either normal development, transgenesis or viability of embryos, tadpoles or frogs using these two techniques (data not shown). The simplified transgenic reaction involves mixing 2.5 × 105 nuclei in 2.5 μl of sperm storage buffer, 100 ng linearized DNA in 2.5 μl of water. The components are mixed and left for 10 minutes at room temperature (~20–23°C). At this point, 495 μl of sperm dilution buffer is added, and the suspension is mixed gently and thoroughly before loading into microinjection needles for nuclear transplantation. Nuclear transplant embryos were screened for expression of the transgene by detecting GFP fluorescence using an Olympus SZX9 fluorescent stereomicroscope.
Tadpole husbandry
Tadpoles were fed Sera Micron , which is a fine granulated tropical fish food. This proved to be an excellent diet for tadpoles, which were maintained on it until metamorphosis. After metamorphosis, froglets were fed progressively larger forms of NASCO Frog Brittle. Tadpoles were either maintained in plastic food containers in which water was changed manually as needed or in continuously circulating Z-MOD tanks (Marine Biotech, Beverly, MA). After metamorphosis, frogs were maintained in Z-MOD tanks until they became large enough to be transferred to larger-sized X-MOD tanks.
RNA purification
Whole tadpoles were homogenized in 100 μl TRIZOL (Invitrogen), and 400μl of the reagent were added to the homogenate. After incubation for 5 min at room temperature, 100μl of chloroform were added, the mixture was shaken vigorously for 15 seconds, incubated for 2–3 minutes at room temperature and centrifuged at 12,000 × g for 15 minutes at 4°C. 250 μl of isopropanol were added to the supernatant, which was then incubated for 10 minutes at room temperature and centrifuged at 12,000 × g for 10 minutes at 4°C. The supernatant was removed, and 500 μl of 75% ethanol were added to the pellet. After mixing well, the mixture was centrifuged at 7,500 × g for 5 minutes at 4°C. The supernatant was discarded, and the pellet was air-dried. The dried pellet was rehydrated in RNase-free water and stored at -80°C. Any contaminating DNA was removed with DNA-free (Ambion) according to the manufacturer's protocol. RNA concentration was determined by spectrophotometry at 260 nm.
RT-PCR
Ready-to-Go RT-PCR beads (Amersham Pharmacia) were used according to the manufacturer's instructions for one-step reactions. RT-PCR reactions for XR11 and EF1α were run in tandem. 500 ng total RNA and 10 pmoles XR11 primers or 5 pmoles EF1α primers were used in each 50 ml reaction. The first strand reaction ran for 15 minutes followed by a 5 minute incubation at 95°C The PCR step ran for 35 cycles for the XR11 product or 30 cycles for the EF1α product. Each cycle was 94°C for 1 minute, followed by 55°C for 1 minute and 72°C for 1 minute. The reaction was stopped by reducing the temperature to 4°C. If necessary, the reaction products could be stored at -20°C until analyzed. The PCR products were analyzed by applying 10 μl samples to a 1.5%agarose/TAE gel containing 0.25 μg ethidium bromide (Sigma-Aldrich) in parallel with a 100 bp ladder (Amersham Pharmacia).
EF1α-downstream primer: ACTGCCTTGATGACTCCTAG
EF1α-upstream primer: CAGATTGGTGCTGGATATGC
This primer pair will generate a 270 bp band.
XR11-downstream primer: GGCATTCTTTCCATACAGGC
XR11-upstream primer: CCTTCTACTTCAGAGCGCCC
This primer pair will generate a 400 bp band.
Immunoblotting
Antiserum to XR11 was prepared by Alpha Diagnostics International, Inc. (San Antonio TX) in rabbit against a synthetic peptide encompassing amino acids 11–30 (KFV SKK LSQ NEA CRK FSN NP). This antiserum was used for immunoblotting at a dilution of 1:1000. Secondary antibody was goat anti-rabbit conjugated to alkaline phosphatase (Jackson ImmunoResearch), which was used at a dilution of 1:1000.
Individual embryos were homogenized so as to correlate immunoblot results with GFP expression levels. Proteins in tissue homogenates were separated using 15% SDS-polyacrylamide gel electrophoresis and blotted onto PVDF. Blots were blocked either overnight at 4°C or for 1.5 hours at room temperature in 5% skim milk powder in TBS containing 0.1% Tween 20. The blots were then incubated with the primary antibody in TBS containing 0.1% Tween 20 for 1 to 2.5 hr at room temperature, washed 1 × 15 min and 3 × 5 min in TBS containing 0.1%Tween 20. The blots were then incubated with the conjugated secondary antibody in TBS containing 0.1% Tween 20 for 1 hr at room temperature, washed as described above, rinsed with TBS, then with staining buffer without stain reagents, followed by staining with NBT/BCIP until developed.
Embryo whole-mount TUNEL assays
Pigmented embryos were collected at stages 28, 33, 37, and 41 and processed for TUNEL assays as described by Hensey and Gautier [13]. After staining, the embryos were fixed while rocking overnight at room temperature in Bouin's solution (70 ml water, 25 ml 37% formaldehyde, 5 ml glacial acetic acid). Embryos were then washed (four times for 15 minutes each) in TE buffered ethanol. Embryos were then bleached for 1 hour at room temperature in 0.5 × SSC, 5% formamide, 1%H2O2 in a Petri dish on a piece of aluminum foil on a shaker under a fluorescent lamp, washed twice for 5 min each in methanol and cleared in BB/BA (2:1 benzylbenzoate:benzyl alcohol). Photographs were taken of embryos representative of the extremes of TUNEL-positive nuclei.
Cryosectioning
Samples were collected at the appropriate stages and fixed overnight at 4°C in MEMFA (0.1 M MOPS pH 7.4, 2 mM EGTA, 1 mM MgSO4, 3.7 % formaldehyde). Samples were then embedded in OCT Compound (Tissue Tek Cat #4583). The embedded samples could then either be stored at -80°C or sectioned immediately. 14 μm sections were made using a Microm cryostat with cabin temperature set at -12°C. Sections were stored at -20°C if not processed immediately. Sections were processed using the ApoTagRed in situ Apoptosis Detection Kit, catalogue #7165 according to the manufacturer's protocol (Intergen Co., Purchase NY). Sections were visualized using a Zeiss Axioplan 2 microscope with Spot camera and software.
Induced metamorphosis
3,3',5-Triiodo-L-thyronine (T3; Sigma Cat #T2752) was added one week after fertilization to a final concentration of 10 nM to tadpole water and changed daily for the duration of the experiment [44](Huang et al., 1999). Before and during treatment, tadpoles were maintained in 0.1 × MMR (for description, see Kroll and Amaya [40]) in the absence of food. Tadpoles responded uniformly to the hormone.
Cartilage staining
Cartilage was visualized in tadpoles according to the procedure described by Klymkowsky and Hanken [54]. Tadpoles were anesthetized in 0.02% benzocaine and immersed in 0.02% Alcian blue 8GX in 70% ethanol/30% glacial acetic acid. After staining, tadpoles were passed through an ethanol series (100, 100, 95, 70, 40, 15%) and then into distilled water. They were then washed in two changes of methanol and cleared in BB/BA.
Irradiation of embryos
Embryos to be irradiated were placed in ~2.5 ml 0.1 × MMR in 35 × 10 mm plastic Petri dishes (approx. 20/dish) and exposed to either 10 Gy or 20 Gy from a Cesium 137 source in a Gammacell 1000 irradiation chamber (MDS Nordion International Inc., Ottawa, ON Canada). After irradiation, embryos were transferred to 60 × 10 mm plastic Petri dishes containing 0.1 × MMR and allowed to recover for 12 hours at room temperature. Unless noted otherwise, each experiment was repeated at least once.
List of abbreviations
DAPI: 4',6-diamidino-2-phenylindole dihydrochloride
GFP: green fluorescent protein
MBT: Mid-blastula transition
NBT: Nitroblue tetrazolium
PCD: programmed cell death
SSC: sodium chloride/sodium citrate
T3: 3,3',5-triiodothyronine
TBS: Tris-buffered saline
TAE: 40 mM Tris-acetate and 1 mM EDTA
TE: 10 mM Tris-HCl and 1 mM EDTA
TUNEL: TdT-mediated dUTP digoxigenin nick end labeling
Authors' contributions
JJ conducted the transgenesis experiments, the molecular studies, the thyroid hormone experiments, the irradiation experiments and helped to draft the manuscript. RC assisted in perfecting the transgenesis technique and conducted the experiments on developmental apoptosis. MC-S initiated the radiation experiments. SM participated in the design of the study and helped draft the manuscript. LWB conceived of the study and participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Figure 2
Click here for file
Acknowledgements
Supported by grants to LWB and SMcF from the Natural Sciences and Engineering Research Council of Canada and to LWB from the Alberta Cancer Board. Plasmids containing the CMV-GFP-SV40 and CMV cassettes for construction of the plasmid containing co-linear CMVXR11 and CARGFP genes were generous gifts from Dr. Enrique Amaya. Dr. James Maller generously supplied the XR11 gene, which was originally cloned by Cruz-Reyes and Tata.
==== Refs
Baehrecke EH How death shapes life during development Nat Rev Mol Cell Biol 2002 3 779 787 12360194 10.1038/nrm931
Glücksmann A Cell death in normal vertebrate ontogeny Biol Rev Cambridge Philos Soc 1951 26 59 86
Sanders EJ Wride MA Programmed cell death in development Int Rev Cytol 1995 163 105 173 8522419
Lockshin RA Williams CM Programmed Cell Death--I. Cytology Of Degeneration In The Intersegmental Muscles Of The Pernyi Silkmoth J Insect Physiol 1965 11 123 133 14287218 10.1016/0022-1910(65)90099-5
Ellis RE Yuan JY Horvitz HR Mechanisms and functions of cell death Annu Rev Cell Biol 1991 7 663 698 1809356 10.1146/annurev.cb.07.110191.003311
Jacobson MD Weil M Raff MC Programmed cell death in animal development Cell 1997 88 347 354 9039261 10.1016/S0092-8674(00)81873-5
Kuan CY Roth KA Flavell RA Rakic P Mechanisms of programmed cell death in the developing brain Trends Neurosci 2000 23 291 297 10856938 10.1016/S0166-2236(00)01581-2
Cowan WM Fawcett JW O'Leary DD Stanfield BB Regressive events in neurogenesis Science 1984 225 1258 1265 6474175
Burek MJ Oppenheim RW Programmed cell death in the developing nervous system Brain Pathol 1996 6 427 446 8944315
Oppenheim RW Cell death during development of the nervous system Annu Rev Neurosci 1991 14 453 501 2031577 10.1146/annurev.ne.14.030191.002321
Voyvodic JT Cell death in cortical development: How much? Why? So what? Neuron 1996 16 693 696 8607987 10.1016/S0896-6273(00)80089-6
Kerr JF Wyllie AH Currie AR Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics Br J Cancer 1972 26 239 257 4561027
Hensey C Gautier J Programmed cell death during Xenopus development: a spatio-temporal analysis Dev Biol 1998 203 36 48 9806771 10.1006/dbio.1998.9028
Hutson LD Bothwell M Expression and function of Xenopus laevis p75(NTR) suggest evolution of developmental regulatory mechanisms J Neurobiol 2001 49 79 98 11598917 10.1002/neu.1067
Anderson JA Lewellyn AL Maller JL Ionizing radiation induces apoptosis and elevates cyclin A1-Cdk2 activity before but not after the midblastula transition in Xenopus Mol Biol Cell 1997 8 1195 1206 9243501
Hensey C Gautier J A developmental timer that regulates apoptosis at the onset of gastrulation Mech Dev 1997 69 183 195 9486540 10.1016/S0925-4773(97)00191-3
Sible JC Anderson JA Lewellyn AL Maller JL Zygotic transcription is required to block a maternal program of apoptosis in Xenopus embryos Dev Biol 1997 189 335 346 9299125 10.1006/dbio.1997.8683
Stack JH Newport JW Developmentally regulated activation of apoptosis early in Xenopus gastrulation results in cyclin A degradation during interphase of the cell cycle Development 1997 124 3185 3195 9272959
Finkielstein CV Lewellyn AL Maller JL The midblastula transition in Xenopus embryos activates multiple pathways to prevent apoptosis in response to DNA damage Proc Natl Acad Sci U S A 2001 98 1006 1011 11158585 10.1073/pnas.98.3.1006
Greenwood J Costanzo V Robertson K Hensey C Gautier J Responses to DNA damage in Xenopus: cell death or cell cycle arrest Novartis Found Symp 2001 237 221 30; discussion 230-4 11444046
Alnemri ES Livingston DJ Nicholson DW Salvesen G Thornberry NA Wong WW Yuan J Human ICE/CED-3 protease nomenclature Cell 1996 87 171 8861900 10.1016/S0092-8674(00)81334-3
Alnemri ES Mammalian cell death proteases: a family of highly conserved aspartate specific cysteine proteases J Cell Biochem 1997 64 33 42 9015752 10.1002/(SICI)1097-4644(199701)64:1<33::AID-JCB6>3.0.CO;2-0
Vander Heiden MG Thompson CB Bcl-2 proteins: regulators of apoptosis or of mitochondrial homeostasis? Nat Cell Biol 1999 1 E209 16 10587660 10.1038/70237
Green DR Kroemer G The pathophysiology of mitochondrial cell death Science 2004 305 626 629 15286356 10.1126/science.1099320
Adams JM Cory S Life-or-death decisions by the Bcl-2 protein family Trends Biochem Sci 2001 26 61 66 11165519 10.1016/S0968-0004(00)01740-0
Cory S Adams JM The Bcl2 family: regulators of the cellular life-or-death switch Nat Rev Cancer 2002 2 647 656 12209154 10.1038/nrc883
Farrow SN Brown R New members of the Bcl-2 family and their protein partners Curr Opin Genet Dev 1996 6 45 49 8791486 10.1016/S0959-437X(96)90009-X
Boise LH Gonzalez-Garcia M Postema CE Ding L Lindsten T Turka LA Mao X Nunez G Thompson CB bcl-x, a bcl-2-related gene that functions as a dominant regulator of apoptotic cell death Cell 1993 74 597 608 8358789 10.1016/0092-8674(93)90508-N
Nakayama K Nakayama K Negishi I Kuida K Shinkai Y Louie MC Fields LE Lucas PJ Stewart V Alt FW Disappearance of the lymphoid system in Bcl-2 homozygous mutant chimeric mice Science 1993 261 1584 1588 8372353
Veis DJ Sorenson CM Shutter JR Korsmeyer SJ Bcl-2-deficient mice demonstrate fulminant lymphoid apoptosis, polycystic kidneys, and hypopigmented hair Cell 1993 75 229 240 8402909 10.1016/0092-8674(93)80065-M
Motoyama N Wang F Roth KA Sawa H Nakayama K Nakayama K Negishi I Senju S Zhang Q Fujii S Massive cell death of immature hematopoietic cells and neurons in Bcl-x-deficient mice Science 1995 267 1506 1510 7878471
Gonzalez-Garcia M Garcia I Ding L O'Shea S Boise LH Thompson CB Nunez G bcl-x is expressed in embryonic and postnatal neural tissues and functions to prevent neuronal cell death Proc Natl Acad Sci U S A 1995 92 4304 4308 7753802
Parsadanian AS Cheng Y Keller-Peck CR Holtzman DM Snider WD Bcl-xL is an antiapoptotic regulator for postnatal CNS neurons J Neurosci 1998 18 1009 1019 9437022
Middleton G Wyatt S Ninkina N Davies AM Reciprocal developmental changes in the roles of Bcl-w and Bcl-x(L) in regulating sensory neuron survival Development 2001 128 447 457 11152643
Middleton G Nunez G Davies AM Bax promotes neuronal survival and antagonises the survival effects of neurotrophic factors Development 1996 122 695 701 8625820
Strasser A Harris AW Jacks T Cory S DNA damage can induce apoptosis in proliferating lymphoid cells via p53-independent mechanisms inhibitable by Bcl-2 Cell 1994 79 329 339 7954799 10.1016/0092-8674(94)90201-1
Domen J Gandy KL Weissman IL Systemic overexpression of BCL-2 in the hematopoietic system protects transgenic mice from the consequences of lethal irradiation Blood 1998 91 2272 2282 9516125
Cruz-Reyes J Tata JR Cloning, characterization and expression of two Xenopus bcl-2-like cell-survival genes Gene 1995 158 171 179 7607538 10.1016/0378-1119(95)00159-4
Yeo W Gautier J A role for programmed cell death during early neurogenesis in xenopus Dev Biol 2003 260 31 45 12885553 10.1016/S0012-1606(03)00222-7
Kroll KL Amaya E Transgenic Xenopus embryos from sperm nuclear transplantations reveal FGF signaling requirements during gastrulation Development 1996 122 3173 3183 8898230
Coen L du Pasquier D Le Mevel S Brown S Tata J Mazabraud A Demeneix BA Xenopus Bcl-XL selectively protects Rohon-Beard neurons from metamorphic degeneration Proc Natl Acad Sci U S A 2001 98 7869 7874 11427732 10.1073/pnas.141226798
Tata JR Gene expression during metamorphosis: an ideal model for post-embryonic development Bioessays 1993 15 239 248 8517853 10.1002/bies.950150404
Tata JR Early metamorphic competence of Xenopus larvae Dev Biol 1968 18 415 440 5704418 10.1016/0012-1606(68)90050-X
Huang H Marsh-Armstrong N Brown DD Metamorphosis is inhibited in transgenic Xenopus laevis tadpoles that overexpress type III deiodinase Proc Natl Acad Sci U S A 1999 96 962 967 9927676 10.1073/pnas.96.3.962
Bartel H Lametschwandtner A Regression of blood vessels in the ventral velum of Xenopus laevis Daudin during metamorphosis: light microscopic and transmission electron microscopic study J Anat 2000 197 ( Pt 2) 157 166 11005708 10.1046/j.1469-7580.2000.19720157.x
Marsh-Armstrong N Huang H Berry DL Brown DD Germ-line transmission of transgenes in Xenopus laevis Proc Natl Acad Sci U S A 1999 96 14389 14393 10588715 10.1073/pnas.96.25.14389
Fadeel B Zhivotovsky B Orrenius S All along the watchtower: on the regulation of apoptosis regulators FASEB J 1999 13 1647 1657 10506569
Wang L Chen L Benincosa J Fortney J Gibson LF VEGF-induced phosphorylation of Bcl-2 influences B lineage leukemic cell response to apoptotic stimuli 2005 19 344 15690071
Simonian PL Grillot DA Nunez G Bcl-2 and Bcl-XL can differentially block chemotherapy-induced cell death Blood 1997 90 1208 1216 9242554
Huang DC Cory S Strasser A Bcl-2, Bcl-XL and adenovirus protein E1B19kD are functionally equivalent in their ability to inhibit cell death Oncogene 1997 14 405 414 9053837 10.1038/sj.onc.1200848
Lee S Schmitt CA Chemotherapy response and resistance Curr Opin Genet Dev 2003 13 90 96 12573441 10.1016/S0959-437X(02)00014-X
Schmitt CA Rosenthal CT Lowe SW Genetic analysis of chemoresistance in primary murine lymphomas Nat Med 2000 6 1029 1035 10973324 10.1038/79542
Sparrow DB Latinkic B Mohun TJ A simplified method of generating transgenic Xenopus Nucleic Acids Res 2000 28 E12 10648800 10.1093/nar/28.4.e12
Klymkowsky MW Hanken J Whole-mount staining of Xenopus and other vertebrates Methods Cell Biol 1991 36 419 441 1725802
|
16185362
|
PMC1262703
|
CC BY
|
2021-01-04 16:30:31
|
no
|
BMC Dev Biol. 2005 Sep 26; 5:20
|
utf-8
|
BMC Dev Biol
| 2,005 |
10.1186/1471-213X-5-20
|
oa_comm
|
==== Front
BMC Evol BiolBMC Evolutionary Biology1471-2148BioMed Central London 1471-2148-5-521620971310.1186/1471-2148-5-52Research ArticleEvidence of recombination among early-vaccination era measles virus strains Schierup Mikkel H [email protected] Carl H [email protected] Claude P [email protected] Laurids S [email protected] Bioinformatics Research Center (BiRC), University of Aarhus, Hoegh Guldbergs Gade 10, Building 090, DK-8000 Aarhus C, Denmark2 Department of Virology, Statens Serum Institut, Copenhagen, Denmark3 Institute of Immunology, Laboratoire National de Santé. PO Box 1102, L-1011 Luxembourg4 Department of Clinical Microbiology, Rigshospitalet, Copenhagen, Denmark2005 6 10 2005 5 52 52 11 4 2005 6 10 2005 Copyright © 2005 Schierup et al; licensee BioMed Central Ltd.2005Schierup et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The advent of live-attenuated vaccines against measles virus during the 1960'ies changed the circulation dynamics of the virus. Earlier the virus was indigenous to countries worldwide, but now it is mediated by a limited number of evolutionary lineages causing sporadic outbreaks/epidemics of measles or circulating in geographically restricted endemic areas of Africa, Asia and Europe. We expect that the evolutionary dynamics of measles virus has changed from a situation where a variety of genomic variants co-circulates in an epidemic with relatively high probabilities of co-infection of the individual to a situation where a co-infection with strains from evolutionary different lineages is unlikely.
Results
We performed an analysis of the partial sequences of the hemagglutinin gene of 18 measles virus strains collected in Denmark between 1965 and 1983 where vaccination was first initiated in 1987. The results were compared with those obtained with strains collected from other parts of the world after the initiation of vaccination in the given place. Intergenomic recombination among pre-/early-vaccination strains is suggested by 1) estimations of linkage disequilibrium between informative sites, 2) the decay of linkage disequilibrium with distance between informative sites and 3) a comparison of the expected number of homoplasies to the number of apparent homoplasies in the most parsimonious tree. No significant evidence of recombination could be demonstrated among strains circulating at present.
Conclusion
We provide evidence that recombination can occur in measles virus and that it has had a detectable impact on sequence evolution of pre-vaccination samples. We were not able to detect recombination from present-day sequence surveys. We believe that the decreased rate of visible recombination may be explained by changed dynamics, since divergent strains do not meet very often in current epidemics that are often spawned by a single sequence type. Signs of pre-vaccination recombination events in the present-day sequences are not strong enough to be detectable.
==== Body
Background
Measles virus (MV) has a genome with negative polarity, consisting of non-segmented single-stranded RNA of approximately 15.9 kb. MV belongs to the Paramyxoviridae family in the order of Mononegavirales and only a single serotype is known. It is among the most infectious viruses known for humans, and no other host species has been identified. Only human populations of a considerable size are able to sustain circulation [1]. Global vaccination programs have resulted in a dramatic decline in measles cases and the documented discontinuation of indigenous circulation in a number of countries [2-4] has encouraged authorities to accomplish global control of measles. However, measles still cause a large number of deaths every year, mainly in developing countries, where endemic circulation of MV is still ongoing [5,6] as a result of poor vaccination coverage.
Intergenomic recombination has been documented in a vast number of virus species within most families of RNA and DNA viruses. Recombination can allow variants of a population to escape from the current fitness peak (escape from Muller's ratchet) and re-appear with a new phenotypic make-up and/or even re-establish in a new host-relationship [7,8]. However, recombination has never convincingly been documented in species of the Mononegavirales, and it is speculated whether this reflects an inability of these viruses to recombine. Yet, the order of Mononegavirales comprises groups of viruses with an apparently large evolutionary potential with frequent shifts of host species. Over the past decades, a considerable number of Mononegavirales members causing diseases in host species in which they had not previously been recognized have been identified. Examples are phocid distemper virus [9], Hendra virus [10], Menangele virus [11], Nipah virus [12] – not to forget the re-emerging divergent members of the Filoviridae (Marburg -and Ebola-like viruses) [see [13,14]]. The origins of these emerging viruses have not been identified and the mechanism(s) of their ability to explore new niches remains enigmatic.
Co-infection of host cells with phylogenetically distinct virus strains is required for recombination events to be detectable in sequence surveys. As a result of the global vaccination against measles a situation of endemic co-circulation of multiple strains [5,6,15] shifted to a situation with a limited number of strains being re-introduced to susceptible subpopulations in major geographical regions [2-4,16]. By sequencing part of the hemagglutinin gene [15], we recently characterized 18 MV strains collected during the pre-/early-vaccination era in Denmark. In the present study, the partial sequence of the hemagglutinin coding region of those older strains is subjected to further analysis and compared with strains sampled after vaccination (generally more recently identified) using various approaches to test for the presence of intergenomic recombination.
Results and discussion
The term pre-/early-vaccination era isolates used for isolates collected in Denmark during the period of 1965–83 is meant to reflect that these isolates are from a period when vaccination against measles was not practiced in Denmark but was gradually becoming common practice in many other countries in the World. Thus, it cannot be excluded that vaccination in other countries influenced measles virus strains circulating in Denmark at the time, but it is anticipated that these isolates still bear valuable information on the nature of strains circulating before an influence of vaccination was imposed.
Table 1 compares differences in basic population genetics of the data sets of the pre-/early- and post-vaccination eras. As might be expected, the global post-vaccination data set shows more sequence variability than the Danish pre-/early-vaccination era data set, whereas the latter shows relatively more singletons, i.e. variants present only in a single sequence. Some of these singletons might be sequencing errors in the pre-/early-vaccination sequences caused by partial degradation as also observed in other ancient DNA studies [33]. Sequencing errors may also explain the higher proportions of transversion substitutions (changes between purine and pyremidine base pairs) and the higher dn/ds ratio in the pre-/early-vaccination data set. For these reasons, singletons are not considered in any of the recombination analyses. Informative sites (i.e. non-singletons) are considered unlikely to be artefacts of the sequencing procedure, since one would not expect the same error to occur more than once, and all informative sites in the pre-/early-vaccination sample are also informative sites in the post-vaccination sample. Codon usage bias measured as the effective number of codons (ENC) is low in both data sets.
Table 1 Basic population genetics summary of the data sets of the pre-vaccination and post-vaccination eras.
Danish early-vaccination era data set Global post-vaccination era data set
Length of sequences 800 800
Number of sequences 18 40
Average number of differences 12.2 31.0
# synonymous substitutions 26 131
# non-synonymous substitutions 32 71
Average Pi (nucleotide diversity)
1. all sites 0.014 0.040
2. synonymous sites 0.030 0.109
3. non-synonymous sites 0.009 0.019
Transition/transversion bias 2.2 9.4
Codon usage bias (ENC) 53 55
Figure 1a shows a minimum evolution tree of the 40 post-vaccination era isolates. The tree is topologically almost identical to the one of Christensen et al. [15] and bootstrap values support many of the genome types identified at present even though the tree is based on only 800 base pairs of the hemagglutinin gene for direct comparison with the pre-/early-vaccination sample [17-19].
Figure 1 Phylogenetic trees reconstructed using the HKY substitution model and the minimum evolution criterion as implemented in MEGA 2.1 [25]. Bootstrap values >80% are shown. a) Tree based on the global post-vaccination era sample with classification of types marked, b) the same sequences as in a) plus the 18 early-vaccination era sequences, marked with a red circle.
Figure 1b shows a tree with the same sequences as in Figure 1a, but adding the 18 pre-/early-vaccination era isolates from Denmark. As shown previously [15], the Danish isolates cluster with genome types A, C2 and E. It is also clear that the inclusion of pre-/early-vaccination era samples makes the distinction of these three genome types less obvious since the similarity distances between post-vaccination era representatives of the genome types A, C2 and E are broken down to the sum of minor differences between the pre-/early-vaccination era isolates. This might reflect that the pre-/early-vaccination era sequences are generally older than the rest and thus at the basis of division of genome types as are also the strains used for vaccine development (e.g. Edmonston). Alternatively, frequent recombination among pre-/early-vaccination era genome types at that time would lead to a poorly resolved tree. Less recombination among different surviving strains after vaccination would then lead to the more differentiated genome types seen today. The subsequent recombination analysis addresses this possibility.
Analysis of recombination
Figure 2 shows all the segregating sites for the pre-/early-vaccination era sequences with positions marked by base pair, and whether a substitution is synonymous or non-synonymous or both (for multiple hits and complex codons). It is immediately clear that many incompatibilities (by the four-gamete test [27]) exist, but also that apparent groups of physically close sites are incompatible with other such groups. We identified by eye five such groups of informative sites and marked them with different colours. Comparing these five groups by the four gamete test demonstrates that some pairs of blocks are incompatible with the same evolutionary tree (Fig. 2b). Even though the identification of these groups is subjective, the presence of such groups suggests either that they have evolved on different evolutionary trees (i.e. recombination) or, alternatively, each of the sites for a given group has mutated at least twice, i.e. by parallel mutations, at approximately the same time (since there is strong linkage disequilibrium within all groups except block 1). Note that the incompatibilities within e.g. block 1 do not weaken this conclusion. A similar table was prepared for the post-vaccination era data set. Here we were not able to visually identify groups of sites failing to satisfy the four-gamete test (results not shown), and the blocks identified in Figure 2 are not incompatible with one another in the post-vaccination data set even though it contains all of the segregating sites. Thus, if the patterns observed in Table 1 were due to convergent evolution of groups of functionally important sites in different lineages, such convergent evolution does not play an important a role in present-day evolution of measles virus. The proportion of pairs of informative sites which is incompatible by the four gamete test is 43% in the pre-/early-vaccination data set as compared to only 23% in the post-vaccination data set even though the variability in general is lower among the pre-/early-vaccination sequences.
Figure 2 Segregating sites in the pre/early-vaccination era sequences with indication of position and state (synonymous or nonsynonymous). Different colours identify five different blocks of sites in strong LD. Table inserted indicates compatibility (+) or incompatibility (-) among blocks. Blocks 2 and 4 are only partly incompatible.
One way to distinguish whether incompatibilities are caused by parallel mutations (i.e. true homoplasies) or by recombination is to correlate linkage disequilibrium (LD) between sets of informative sites to distance. Recombination causes a decrease in linkage disequilibrium and more recombination is expected between sites that are further apart from each other. The block structure of close sites is visualised in a different way in Figure 3, which shows all the significant cases of linkage disequilibrium. Visual inspection of the Figure suggests that physically close sites appear to be more likely to be in strong LD than sites far apart. This observation is tested more formally by correlating LD with distance (Table 2 and Figure 4). Table 2 shows the correlation of two commonly used measures of LD for both data sets. A significant negative correlation is observed by both measures of linkage disequilibrium in the pre-/early-vaccination data set, whereas the post-vaccination data set does not show any significant decay of linkage disequilibrium with distance. The decrease in linkage disequilibrium with distance for the pre-/early-vaccination sequences is shown graphically in Figure 4.
Figure 3 Linkage disequilibrium triangle plot for informative sites in the pre/early-vaccination sample. Significant linkage disequilibrium is indicated my shading, grey shading is P < 0.05, black shading, P < 0.001 by Fisher's exact test.
Table 2 Summary of numerical analysis of recombination in the samples of the pre-vaccination and post-vaccination eras A negative correlation implies that LD decays with distance, the P-values are obtained by a permutation test (see Methods).
Early-vaccination sample Post-vaccination sample
R2-correlation (P-value) -0.26 (P < 0.01) -0.03 (P < 0.05)
D' correlation (P-value) -0.32 (P < 0.001) 0.01 (P > 0.05)
LDhat estimate of ρ (P-value) 15.8 (P < 0.002) 7.2 (P > 0.05)
Figure 4 Correlation between linkage disequilibrium and distance for the early-vaccination era data set using the R2 and D' measures of LD, respectively.
An analysis of expected number of homoplasies caused by parallel mutations and the number of apparent homoplasies in the most parsimonious tree was performed on the pre-/early-vaccination data set [see [32,35]]. The basic idea is to investigate whether the number of apparent homoplasies in the most parsimonious tree is likely to be due to parallel (recurrent) mutations. If this is not the case, it is a strong indication of recombination since recombination easily creates "incompatible pairs of sites", i.e. pairs of sites that do not fit into the same phylogenetic tree, and thus will appear as homoplasies if assuming no recombination and a single phylogenetic tree. There are 800 sites. Of these, 267 are third position sites, and among these we observe 20 transition mutations. If we assume equal mutation rates at all third position sites, then the number of transition mutations we expect have happened while observing 20 different ones is 20.8+/-0.9 (calculated as the sum of geometric distributions). In the most parsimonious tree found using PAUP*, there are 26 transition mutations. In other words, given the observed number of mutation events of the transition type, we expect 0.8+/-0.9 parallel mutations/homoplasies under the assumption of no recombination (and no codon usage bias). We observe, however, 6 parallel mutations in the tree. This suggests that some of the apparent homoplasies resulted from recombination rather than from recurrent mutations. We also calculated the effective sites number [35] from the observed codon usage bias. However, since the codon usage bias is low, the effect is minor, and only 1.0+/-1.0 homoplasies are expected under this model, again significantly smaller than the six apparent homoplasies. A large amount of mutation rate heterogeneity at the synonymous sites offers an alternative explanation of the excess homoplasies. While this explanation cannot be ruled out, analysis of the post-vaccination data set does not support large rate heterogeneity at silent sites in the evolution of present-day measles virus, since the number of homoplasies in this data set can be explained by recurrent mutations (results not shown).
Given these different lines of evidence for recombination, it is of interest to try to estimate the rate of recombination needed to explain the data. The most appropriate method is the finite site, composite likelihood approach implemented in LDhat [31]. The result (Table 2) is that a significantly positive recombination rate is found in the pre-/early-vaccination sequences, much reflecting the results of the similar R2 test. An estimated rate of ρ = 15.8 corresponds to that an expected 45 recombination events have happened in the ancestral history of the 18 pre-/early-vaccination sequences. The estimated recombination rate appears smaller than rates reported for HIV and other viruses [31]. The estimated recombination rate in the post-vaccination era data sets also shows a positive rate of ρ, but it not significantly different from zero by the permutation test (Table 2). The lack of evidence of recombination in present day sequences of MV strains is consistent with what was also observed by [31].
In conclusion, the analysis of pre-/early-vaccination era MV sequences shows evidence of recombination at rates important to the evolution of MV. The five different tests of recombination should not be considered independent tests and some of the results might be explained by alternative mechanisms such as convergent evolution of functionally important sites and rate heterogeneity of synonymous variation. However, all tests agree and provide evidence of recombination both through an excess of apparent homoplasies compared to the expected frequency of parallel mutations, and through a decrease in LD with distance, which is difficult to explain by any other hypothesis than recombination. Furthermore, it is difficult to imagine a mechanism other than recombination by which apparent homoplasies could occur pairwise or in triplets in distant parts of the sequence considering also that such patterns are not seen in the post-vaccination data set.
The evidence of recombination among pre-/early-vaccination era MV strains and the lack of detection of recombination among post-vaccination era MV strains are consistent with the shift in epidemiology from a situation of co-circulation of strains in populations to a bottleneck situation with incidental introduction of a single strain to a susceptible sub-population of a geographical region. Given that intergenomic recombination and co-infection of individuals are common phenomena of MV it might be assumed that the lineages surviving till today did emerge from a pool of recombining pre-vaccination era strains. A lower level or absence of recombination due to changed epidemiology since then has erased our ability to detect recombination in a global sample of present day MV despite its high level of variability. It is possible that the Danish pre-/early-vaccination era strains are representatives of the very pool in which recombination took place while present-day MV strains are representatives of temporally and/or geographically separated lineages. Analysing informative sites in other parts of the MV genome of present-day lineages render it unlikely that these lineages could have derived from a clonal population structure of a global pool of MV strains (L. S. Christensen, unpublished data).
The template for replication in members of the Paramyxovirinae is a nucleocapsid complex in which each nucleocapsid monomer (N) is predicted to be associated by hydrophobic bonds with 6 nucleotides in such a way as to resist non-ionic detergent and high salt and to protect the RNA from RNase digestion [36]. This tight association of RNA and protein has been dubbed "the rule of six" and excludes the intracellular presence of naked viral RNA molecules. It raises the question of the mechanism of RNA polymerase template recognition and has provided an explanation of the hypothesis that recombination possibly cannot occur in this group of viruses. Our data suggests that a mechanism of partial unwinding of the nucleocapsid structure exists to allow homologous intergenomic recombination or RNA polymerase template shift during replication.
Conclusion
Measles virus appear to possess the ability to recombine but the present-day epidemiology of the virus where different sequence types rarely or never meet make the impact of recombination on the distribution of sequence diversity negligible. However, in the prevaccination area, endemic MV allowed more divergent strains to meet and recombine. The present-day strains are thus descendants of recombined sequences but the signal of the early recombination is lost in present-day sequences.
Methods
Sequence data
The Danish pre-/early-vaccination sequences consist of an 800 base-pair region (nt. 659 to 1458) of the hemagglutinin coding region of 18 strains collected in Denmark, Greenland and the Faeroe Islands between 1965 and 1983 (GenBank accession numbers AJ417850-AJ417867) [15]. Post-vaccination sequences of 40 strains, representative of the 22 phylogenetic clusters identified [17-19], were trimmed to match the region of the pre-/early-vaccination era sequences. GenBank accession numbers of 33 of these sequences can be found in [15]. The remaining 7 HA sequences that complete the list of globally circulating genome types, identified at present, are those of strains Mvi/Gambia/93 (Type B3, Acc. No AF484955) [20], MVi/Alberta.CAN/20.00/1 (Type D7, Acc. No. AF410986) (G.A. Tipples et al., submitted 14-Aug-2001), MVi/Montreal.CAN/19.98 (Type D8, Acc. No. AF410985) (G.A. Tipples et al., sumitted 14-Aug-2001), MVi/Vic.AU/12.99 (Type D9, Acc No. AY127853) [21], (Type G2, Acc. No. AF243851) [22], MVi/Gresik.INO/18.02 (Type G3, Acc. No. AY184218) (P.A. Rota and S.L. Liffick, submitted 19-Nov-2002), and China94-1 (Type H2, Acc. No. AF045203) [23].
Sequence analysis
The alignment of the 18 pre-/early-vaccination and 40 post-vaccination sequences does not contain any gaps. The computer program DnaSP 3.99 [24] was used for estimation of standard parameters of population genetics. Segregating sites were classified as synonymous or non-synonymous, except for complex codons where a site may be classified as both. Phylogenetic trees from the post-vaccination data set and the two data sets combined were built using MEGA 2.1 [25] and the minimum evolution criterion. The HKY substitution model [26] was assumed. Bootstrap values were estimated from 2000 re-samples.
Recombination was examined using five different, but complementary, approaches. (i) A graphical method by which sets of sites in strong linkage disequilibrium (LD) were visually identified and marked with different colourings (Fig. 2). It was then investigated whether different sets of blocks were incompatible by the four-gamete test [27]. If blocks are incompatible, one will infer either recombination or recurrent mutation of the sites in a block at about the same time. (ii) A plot of significant pairwise linkage disequilibria was constructed using DnaSP. (iii) Decay of linkage disequilibrium with distance (measured as either the squared correlation coefficient R2 or the standardized measure D') was investigated following [28] and estimated using the R2-program [29]. Analysis was restricted to informative sites. (iv) Estimation of the scaled population recombination rate ρ by a composite maximum likelihood approach [30], using the LDhat program of [31]. This program also allows a test of the null hypothesis of no recombination (ρ = 0) by a permutation test. This analysis was also restricted to informative sites. (v) Comparison of the expected number of homoplasies to the number of apparent homoplasies in the most parsimonious phylogenetic tree, closely following the approach of [32], using PAUP* [32].
Authors' contributions
CHM, CPM and LSC collected the data. LSC formulated the hypothesis, and MHS performed all analyses. MHS and LSC wrote the paper.
Acknowledgements
We thank Oliver Pybus, Roald Forsberg, Freddy B. Christensen and an anonymous reviewer for very helpful comments to the manuscript,. Rikke Jonson, Lis Nielsen and Gunilla Trolle at the Department of Clinical Microbiology, Rigshospitalet, and Ellen Christensen at Department of Virology, Statens Serum Institut are acknowledged for expert technical assistance, Enette B. Knudsen for editing. The study was supported by Sygekassernes Helsefond (grant No. 11/282-94), the Danish Medical Research Council (grant No. 12-1667), the Novo Nordisk Foundation and the Danish Natural Sciences Research Council (grant no. 1262). Frank Jorgensen and Paul Sharp are acknowledged for valuable discussions.
==== Refs
Griffin DE Knipe DM, Howley PM Measles virus Virology 2001 4 Philadelphia: Lippincott Williams & Wilkins 1401 1441
Rota JS Heath JL Rota PA King GE Celma ML Carabaña J Fernandez-Muñoz R Brown D Jin L Bellini WJ Molecular epidemiology of measles virus: identification of pathways of transmission and implications for measles elimination JID 1996 173 32 37 8537679
Rima BK Earle JAP Yeo RP Herlihy L Baczko K ter Meulen V Carabana J Caballero M Celma ML Fernandez-Munoz R Temporal and geographical distribution of measles virus genotypes J Gen Virol 1995 76 73 1180 7844544
Dahl L Christensen LS Schöller S Westh H Plesner A-M Sequence analysis of the hemagglutinin gene of measles virus isolates in Denmark 1997–1998: no evidence of persistent circulation of measles virus in Denmark APMIS 2000 108 267 72 10843413 10.1034/j.1600-0463.2000.d01-54.x
Hanses F Truong AT Ammerlaan W Ikusika O Adu F Oyefolu AO Omilabu SA Muller CP Molecular epedemioology of Nigerian and Ghanian measles virus isolates reveals a genotype circulating widely in western and central Africa J Gen Virol 1999 80 871 7 10211955
Truong AT Kreis S Ammrelaan W Hartter HK Adu F Omilabu SA Berbers GAM Muller CP Genotyping and antigenic characteristics of hemagglutinin proteins of African measles virus isolates Virus Res 1999 62 89 95 10513290 10.1016/S0168-1702(99)00072-6
Christensen LS The population biology of suid herpesvirus 1 Acta Pathologica Microbiologica et Immunologica Scandinavica 1995 103 1 48
Burke DS Recombination in HIV: an important viral evolutionary strategy Emerging Infectious Diseases 1997 3 253 259 9284369
Osterhaus AD Vedder EJ Identification of viruses causing recent seal deaths Nature 1988 335 20 3412456 10.1038/335020a0
O'Sullivan JB Allworth AM Paterson DL Snow TM Boots R Gleeson LJ Fatal encephalitis due to novel paramyxovirus transmitted from horses Lancet 1997 349 93 95 8996421 10.1016/S0140-6736(96)06162-4
Philbey AW Kirkland PD Ross AD Davis RJ Gleeson AB Love RJ Daniels PW Gould AR Hyatt AD An apparently new virus (Family Paramyxoviridae) infectious for pigs, humans and fruit bats Emerging Infectious Diseases 1998 4 269 275 9621197
Chua KB Bellini WJ Rota PA Harcourt BH Tamin A Lam SK Ksiazek TG Rollin PE Zaki SR Shieh W Goldsmith CS Gubler DJ Roehrig JT Eaton B Gould AR Olson J Field H Daniels P Ling AE Peters CJ Anderson LJ Mahy BW Nipah virus: a recently emergent deadly paramyxovirus Science 2000 288 1432 1435 10827955 10.1126/science.288.5470.1432
Monath TP Ecology of Marburg and Ebola viruses: speculations and directions for future research J Infect Dis 1999 179 S127 138 9988176
Balter M On the trail of Ebola and Marburg viruses Science 2000 290 924 925
Christensen LS Schöller S Schierup MH Vestergaard BF Mordhorst CH Sequence analysis of pre- and early-vaccination era strains of measles virus in Denmark 1965–83 reveal diversity of ancient strains circulating globally APMIS 2002 110 113 22 12064866 10.1034/j.1600-0463.2002.100201.x
Mulders MN Truong AT Muller CP Monitoring of measles elimination using molecular epidemiology Vaccine 2001 19 2245 2249 11257341 10.1016/S0264-410X(00)00453-9
World Health Organisation Expanded programme on immunization-standardization of the nomenclature for describing the genetic characteristics of wild-type measles virus Weekly Epidemiological Record 1998 73 265 272 9745371
World Health Organisation Nomenclature for describing the genetic characteristics of wild-type measles viruses (update) Weekly Epidemiological Record 2001 76 249 251
Rota PA Bellini WJ Update on the global distribution of genotypes of wild type measles viruses J Infect Dis 2003 187 S270 6 12721925 10.1086/368042
Waku DK Nerrienet E Mfoupouendoun J Tene G Whittle H Wild TF Measles virus strains circulating in Central and West Africa: Geographical distribution of two B3 genotypes J Med Virol 2002 68 433 440 12226833 10.1002/jmv.10222
Chibo D Riddell M Catton M Lyon M Lum G Birch C Studies of measles viruses circulating in Australia between 1999 and 2001 reveal a new genotype Virus Res 2003 91 213 221 12573500 10.1016/S0168-1702(02)00273-3
Rota PA Liffick S Rosenthal S Heriyanto B Chua KB Measles genotype G2 in Indonesia and Malaysia Lancet 2000 355 1557 1558 10801203 10.1016/S0140-6736(05)74612-2
Xu W Tamin A Rota JS Zhang L Bellini WJ Rota PA New genetic group of measles virus isolated in the People's Republic of China Virus Res 1998 54 147 156 9696123 10.1016/S0168-1702(98)00020-3
Rozas J Sánchez-DelBarrio JC Messeguer X Rozas R DnaSP, DNA polymorphism analyses by the coalescent and other methods Bioinformatics 2003 14668244
Kumar S Tamura K Jakobsen IB Nei M MEGA2: molecular evolutionary genetics analysis software Bioinformatics 2001 17 1244 5 11751241 10.1093/bioinformatics/17.12.1244
Hasegawa M Kishino H Yano T Dating of the human-ape splitting by a molecular clock of mitochondrial DNA Journal of Molecular Evolution 1985 22 160 174 3934395 10.1007/BF02101694
Hudson RR Kaplan NL Statistical properties of the number of recombination events in the history of a sample of DNA sequences Genetics 1985 111 147 64 4029609
Awadalla P Eyre-Walker A Smith JM Linkage disequilibrium and recombination in hominid mitochondrial DNA Science 1999 286 2524 2525 10617471 10.1126/science.286.5449.2524
Recombination analysis software
Fearnhead P Donnelly P Estimating recombination rates from population genetic data Genetics 2001 159 1299 1318 11729171
McVean GAT Awadalla P Fearnhead P A coalescent-based method for detecting recombination from gene sequences Genetics 2002 160 1231 1241 11901136
Eyre-Walker A Smith NH Smith JM How clonal are human mitochondria? Proceedings of the Royal Society of London Series B-Biological Sciences 1999 266 477 483 10.1098/rspb.1999.0662
Swofford DL PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4 2001 Sinauer Associates, Sunderland, Massachusetts
Krings M Stone A Schmitz RW Krainitzki H Stoneking M Paabo S Neandertal DNA sequences and the origin of modern humans Cell 1997 90 19 30 9230299 10.1016/S0092-8674(00)80310-4
Smith JM Smith NH Detecting recombination from gene trees Mol Biol Evol 1998 15 590 599 9580989
Egelman EH Wu SS Amrein M Protner A Murti G The Sendai virus nucleocapsid exists in at least four different helical states J Virol 1989 63 2233 2243 2539515
|
16209713
|
PMC1262704
|
CC BY
|
2021-01-04 16:29:16
|
no
|
BMC Evol Biol. 2005 Oct 6; 5:52
|
utf-8
|
BMC Evol Biol
| 2,005 |
10.1186/1471-2148-5-52
|
oa_comm
|
==== Front
BMC Fam PractBMC Family Practice1471-2296BioMed Central London 1471-2296-6-401620214610.1186/1471-2296-6-40Research ArticleLength of patient-physician relationship and patients' satisfaction and preventive service use in the rural south: a cross-sectional telephone study Donahue Katrina E [email protected] Evan [email protected] Donald E [email protected] Department of Family Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, 27599, USA2 Cecil G. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, 27599, USA2005 4 10 2005 6 40 40 24 3 2005 4 10 2005 Copyright © 2005 Donahue et al; licensee BioMed Central Ltd.2005Donahue et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Physicians and patients highly value continuity in health care. Continuity can be measured in several ways but few studies have examined the specific association between the duration of the patient-doctor relationship and patient outcomes. This study (1) examines characteristics of rural adults who have had longer relationships with their physicians and (2) assesses if the length of relationship is associated with patients' satisfaction and likelihood of receiving recommended preventive services.
Methods
Cross-sectional telephone survey of health care access indicators of adults in selected non-metropolitan counties of eight U.S. predominantly southern states. Analyses were restricted to adults who see a particular physician for their care and weighted for demographics and county sampling probabilities.
Results
Of 3176 eligible respondents, 10.8% saw the same physician for the past 12 months, 11.8% for the previous 13–24 months, 20.7% for the past 25–60 months and 56.7% for more than 60 months. Compared to persons with one year or less continuity with the same physician, respondents with over five years continuity more often were Caucasian, insured, a high school graduate, and more often reported good to excellent health and an income above $25,000. Compared to those with more than five years of continuity, participants with either less than one year or one to two years of continuity with the same physician were more often not satisfied with their overall health care (OR 2.34; OR 1.78), participants with less than one year continuity were more often not satisfied with the concern shown them by their physician (O.R. 1.90) and having their health questions answered, and those with one to two years continuity were more often not satisfied with the quality of their care (OR 2.37). No significant associations were found between physician continuity and use rates of any of the queried preventive services.
Conclusion
Over half of this rural population has seen the same physician for more than five years. Longer continuity of care was associated with greater patient satisfaction and confidence in one's physician, but not with a greater likelihood of receiving recommended preventive services.
==== Body
Background
Both physicians and patients highly value continuity in healthcare [1,2]. The Institute of Medicine (IOM) holds that continuity, defined as an ongoing partnership between patients and physicians, is a central and important component of primary care [3]. Continuity fosters personal relationships and is believed to improve health outcomes, although the latter has been difficult to demonstrate formally.
Continuity has been measured in various ways with different associated outcomes. Concepts of continuity include having a usual place for healthcare, having a usual physician and actually having a greater proportion of one's visits with a particular physician [4-6]. Continuity reportedly increases use rates of some services, including some preventive services, and is associated with lower health care costs for certain populations [7-12]. However, continuity has shown no effect on the use of some particular preventive services, including Pap smears and mammograms [4,13]. Evidence for the relationship between continuity with hospitalization rates is conflicting [4,10,14,15]. Lack of continuity care is associated with fewer follow-up care visits and fewer medical prescriptions given for chronic illness [16]. Few studies, however, have examined outcomes for continuity as defined as having a longer standing relationship with one's physician [9,11]. The outcomes of this notion of continuity are less clearly known.
Elements of trust and satisfaction, important to the ongoing patient-physician relationship, are also important to examine in the context of continuity. Continuity over time allows physicians to get to know and better understand their patients [17] and is associated with greater trust [18]. In older patients, longer relationships are associated with the perception that their provider is more knowledgeable and thorough [11]. Considering these aspects of satisfaction may further the understanding of the benefits of a continuous health care relationship [19,20].
The purpose of this study is to examine the characteristics of rural adults who have longer versus shorter relationships with their current physicians. We also examine if the length of this relationship is associated with aspects of patient satisfaction and with the likelihood that patients have received recommended preventive services.
Methods
Study population
Cross-sectional data were obtained from a telephone survey of adults in the rural U.S. Southeast. The survey was conducted as part of an evaluation of the Southern Rural Access Program (SRAP), an initiative to improve access to health care in targeted rural areas of eight U.S. states: Alabama, Arkansas, Georgia, Louisiana, Mississippi, South Carolina, Texas and West Virginia [21]. Professional Research Consultants, Inc. of Omaha, Nebraska (PRC), a survey research firm, fielded the telephone survey with 4,879 adult respondents living in 150 non-metropolitan counties of these eight states from October 2002 to July 2003. PRC administered the survey using a computer-assisted telephone interview (CATI) system with randomly generated numbers within telephone exchanges and active number blocks within each county. Eligible adults included those age 18 and older who had lived in the immediate area for at least 12 months and spoke English or Spanish; a minimum of 10 attempts were made to contact someone at each number. After a number was contacted and confirmed to be a household, a specific adult to be interviewed was randomly selected from among the eligible adults in that household using the next birthday method of identification [22]. The overall participation rate was 51.0% (4879 participants and 4682 refusals). The study protocol was reviewed and exempted by the University of North Carolina School of Medicine's Committee on the Protection of Human Subjects Research.
Eligibility
Participants were asked, "Is there a place that you usually go to when you are sick or need advice about your health?", if they saw a particular person there for their care and how long that person had been their doctor. Of the 4879 respondents, 4367 (89.5%) reported a usual place of care, 3402 (69.7%) saw a particular person for their care and 3197 (65.5%) stated that person was a doctor.
Subjects' were asked 'How long has this person been your doctor?' and asked to respond in one of four duration categories: previous year (0–12 months), the past one to two years (13–24 months), the past three to five years (25–60 months) and more than 5 years (61 months or more). Of the 3197 eligible subjects, 3176 (99.3%) responded to this question and served as the population for this study.
Satisfaction questions
Several satisfaction questions were included in the questionnaire based on previous national and regional surveys and a published study [23-25]; responses were provided on five-point Likert scales with a neutral middle option offered. Participants were asked how satisfied they were with their overall health care and with the quality of care they usually received. Participants were also asked how satisfied they were with having their health questions answered during care visits and how welcome and comfortable they were made to feel by the office staff. In addition, participants were asked how satisfied they were with the concern shown for them by their doctor and how confident they were in the abilities of their doctor to help them.
Preventive service questions
Respondents who reported they had a 'routine medical checkup' in the past year were asked several questions pertaining to the preventive services they had received. They were asked how long it had been since they last had a mammogram, Pap smear, flexible sigmoidoscopy/colonoscopy, influenza vaccine and cholesterol level check. They were also asked if they were counseled about tobacco use (if a smoker), physical activity/exercise and nutrition/diet in the previous 12-months.
Standards of appropriate preventive services were adapted from the conservative recommendations of the US Preventive Services Task Force [26]: (1) mammogram in adult women 50–69 years old in the past year, (2) Pap smear in women 18–64 years old in the past three years, (3) flexible sigmoidoscopy or colonoscopy in people 50 and older ever (at least one), (4) influenza vaccine in people 65 and older in the past year and (5) cholesterol level check in persons 45 years and older in the past five years. Counseling variables included if a doctor or nurse: (1) advised a smoker to quit or stop using tobacco in the past year, (2) had given the participant advice about diet and nutrition in the past year, and (3) had given the participant advice about physical activity or exercise in the past year.
Analysis
Participants were sorted into four groups based on the number of years they had received care from the same physician: less than a year, one to two years, three to five years and more than five years. Chi-square tests were used to compare the demographic, satisfaction and preventive health care characteristics of each group of years of continuous care to the group with over five-years of care.
Logistic regression was performed to assess the relationship of continuity to the outcomes of satisfaction and preventive service use rates. Dummy variables were constructed for the years subjects had received care from their primary source of care, with more than five years of care as the omitted (comparison) category. Models for each outcome were adjusted for subject age, gender, race, income, health insurance status and self-reported general health status.
For all analyses STATA 8 (College Station, Texas) [27] was used. When respondent demographics were compared against the 2000 U.S. Census data, survey participation rates were found to be lower for males, persons 18–39 years of age, African-Americans, and those with household incomes below $15,000. Analyses were accordingly weighted to adjust for both the over-sampling in small counties and to correct differential response likelihood by demographic groups.
Results
Among the 3176 persons identifying a particular physician from whom they received their health care, 10.8% (N = 319) had seen the same individual for the past year, 11.8% (N = 369) for the past one to two years, 20.7% (N = 669) for the past three to five years and 56.7% (1819) had seen the same physician for more than five years. Associations between sociodemographic characteristics and duration of care are shown in Table 1. Compared to persons with one year or less continuity with the same physician, respondents with more than five years continuity were more often Caucasian (66.9% versus 61.5%, p = 0.007), more often had insurance (77.3% versus 68.5%, p = 0.001), had more education (43.6% versus 37.5% had at least some college, p = 0.006), more often had an income $25,000 or more (62.4% versus 45.6%, p < 0.001) and more often reported good to excellent health (76.0% versus 66.2%, p < 0.001). Respondents with more than five years continuity also more often had an income higher then $25,000 (62.4%) than the groups reporting only one to two years of continuity (50.1%, p < 0.001) and three to five years continuity (57.6%, p = 0.035). Otherwise, the one to two-year, three to five-year and over five-year continuity groups did not differ in their characteristics.
Table 1 Characteristics of subjects who have received care from the same physician for ≤ 1 year, >1 to 2 years, >2 to 5 years, and more than 5 years: Statistical comparisons to group with more than 5 years of continuity
Patient Characteristics %≤ 1 yr
(0–12 mo)
N = 319 % 1–2 yrs
(13–24 mo)
N = 369 % 3–5 yrs
(25–60 mo)
N = 669 %>5 yrs
(61 mos or more)
N = 1819
Age
18–39 35.8 38.4 36.9 33.0
40–64 42.3 38.3 44.4 44.3
65+ 21.9 23.2 18.7 22.7
Gender
Male 42.1 44.5 42.3 45.2
Female 57.9 55.5 57.6 54.8
Race/Ethnicity **
White, Non Hispanic 61.5 67.5 61.6 66.9
White, Hispanic 3.3 0.5 1.1 1.0
Black 33.5 29.5 33.2 29.3
Other 1.7 2.5 4.1 2.7
Insurance ***
Yes 68.5 72.9 77.7 77.3
No 31.5 27.1 22.2 22.7
Education **
Less than high school 26.2 17.8 19.4 18.0
High school diploma 29.3 32.7 30.1 30.4
Trade school 7.1 8.1 8.1 8.0
Some College or more 37.5 41.4 42.4 43.6
Income *** *** *
< $25,000 54.4 50.0 42.4 37.6
≥ $25,000 45.6 50.1 57.6 62.4
General health ***
Good to Excellent 66.2 72.7 73.1 76.0
Fair to Poor 33.8 27.3 26.9 24.2
*P < 0.05 difference for each group with >5 years continuity
**P < 0.01 difference for each group with >5 years continuity
***P < or = 0.001 difference for each group with >5 years continuity
In terms of putative outcomes of continuity, compared to those with more than five years continuity, more respondents with one year or less of continuity reported being neutral or dissatisfied with their overall health care (10.6% versus 4.5%, p < 0.001), with the quality of their health care (6.7% versus 3.3%, p = 0.022) with having their health questions answered (7.2% versus 3.8%, p = 0.021) and with the concerns shown to them by their physicians (7.0% versus 3.4%, p = 0.004) (Table 2). A higher proportion of persons with one year or less continuity were also neutral or not confident in the abilities of their physician to help them than those with more than five years of continuity (20.1% versus 14.5%, p = 0.035). There were no significant differences in rates of any of the preventive services, except those who reported seeing a physician for one year or less were more likely to report being counseled in nutrition over the past year (57.8% versus 48.1%, p = 0.008).
Table 2 Association between preventive services and lack of satisfaction with number of years of care from the same physician: Statistical comparisons to group with more than 5 years of continuity
Outcomes Total Sample
N = 3176 %≤ 1 yr
(0–12 mo)
N = 319 % 1–2 yrs
(12–24 mo)
N = 369 % 3–5 yrs
(25–60 mo)
N = 669 %>5 yrs
(61 mos or more)
N = 1819
Lack of Satisfaction
% Neutral or dissatisfied with overall health care N = 3150 10.6*** 7.7* 3.8 4.5
% Neutral or dissatisfied with quality of health care N = 3156 6.7* 7.3** 3.5 3.3
% Neutral or not confident in physician N = 3156 20.1* 15.4 16.6 14.5
% Neutral or dissatisfied with concern shown by physician N = 3159 7.0** 4.8 5.3* 3.4
% Neutral or dissatisfied with having questions answered N = 3161 7.2* 5.3 4.4 3.8
% Felt neutral or not welcome and comfortable by staff N = 3165 5.6 4.8 3.5 4.3
Preventive Services (only asked to those reporting a routine checkup in past year)
Mammogram within past year in women 50 and older N = 1234 56.2 61.6 59.7 60.4
Pap smear in women 18–64 years old in past 3 years N = 1579 89.0 89.3 90.7 88.0
Flu shot in persons 65 and older in past year N = 720 62.0 64.9 72.0 69.5
Flexible sigmoidoscopy or colonoscopy in persons 50 and older N = 1756 46.6 52.3 50.5 50.5
Cholesterol in persons 45 and older N = 2108 87.6 89.9 88.4 88.3
Smoking counseling in past year N = 634 68.6 73.0 72.9 69.8
Physical activity counseling in past year N = 2641 56.4 51.2 55.6 51.2
Nutrition counseling in past year N = 2632 57.8* 47.5 54.6* 48.1
*P < 0.05 difference for each group with >5 years continuity
**P < 0.01 difference for each group with >5 years continuity
***P <0.001 difference for each group with >5 years continuity
Controlling for age, gender, race, income, insurance and health status, participants with one year or less and one to two years of continuity with their physician remained more likely to be neutral or dissatisfied with their overall health care compared to those with more than 5 years of continuity (OR 2.34; 95% CI: 1.39–3.93 and OR 1.78; 95% CI: 1.04–3.06, respectively) (Table 3). Those with one to two years of continuity were also more likely to be neutral or dissatisfied with the quality of health care they usually received (OR 2.37; 95% CI: 1.35–4.15). Those with one year or less continuity were also more likely to be neutral or dissatisfied with the concern shown by their physician (OR 1.90; 95% CI: 1.12–3.23) and having their questions answered (OR 1.98; 95% CI: 1.10–3.57). There were no significant differences across groups with various lengths of relationships with their physicians and their confidence in their physician, feeling welcome by office staff, reported rates of counseling for smoking and physical activity and receipt of preventive services (Table 4). Persons with shorter periods of continuity were more likely, however, to report having received nutrition counseling in the past year (OR 1.47; 95% CI: 1.08–2.02).
Table 3 Adjusted relationships between lack of satisfaction and years of care from the same physician: logistic regression results*: Statistical comparisons to group with more than 5 years of continuity
Lack of Satisfaction %≤1 yr care
OR (95% CI) % 1–2 yrs care
OR (95% CI) % 3–5 yrs care
OR (95% CI)
% Neutral or dissatisfied with overall health care 2.34 1.78 0.81
(1.39–3.93) (1.04–3.06) (0.50–1.32)
p = 0.001 p = 0.04 p = 0.40
% Neutral or dissatisfied with quality of health care 1.87 2.37 1.05
(0.98–3.60) (1.35–4.15) (0.56–1.97)
p = 0.058 p = 0.003 p = 0.87
% Neutral or not confident in physician 1.36 1.07 1.11
(0.94–1.98) (0.77–1.50) (0.83–1.49)
p = 0.10 p = 0.65 p = 0.47
% Neutral or dissatisfied with concern shown by physician 1.90 1.42 1.37
(1.12–3.23) (0.74–2.76) (0.83–2.27)
p = 0.02 p = 0.29 p = 0.21
% Neutral or dissatisfied with having questions answered 1.98 1.55 1.18
(1.10–3.57) (0.81–2.94) (0.70–2.00)
p = 0.02 p = 0.18 p = 0.52
% Felt neutral or not welcome and comfortable by staff 1.28 1.14 0.77
(0.66–2.52) (0.60–2.15) (0.48–1.23)
p = 0.46 p = 0.69 p = 0.27
*Adjusted for age, gender, race, income ≥ 25,000, health insurance status, and health status. Statistical comparisons to group with more than 5 years
Table 4 Adjusted relationships between receipt of preventive service variables and years of care from the same physician: logistic regression results: Statistical comparisons to group with more than 5 years of continuity
Preventive services (only asked to those reporting at least one doctor visit in past year) % ≤1 yr care
OR (95% CI) % 1–2 yrs care
OR (95% CI) % 3–5 yrs care
OR (95% CI)
Mammogram in women 50 and older within past year (n = 1182) 0.94 1.17 0.96
(0.60–1.46) (0.77–1.80) (0.70–1.32)
p = 0.78 p = 0.44 p = 0.81
Pap smear in women 18–64 years old in past 3 years (n = 1543) 1.27 1.17 1.36
(0.71–2.24) (0.62–2.20) (0.83–2.24)
p = 0.41 p = 0.63 p = 0.22
Flu shot in persons 65 and older in past year (n = 679) 0.66 0.83 1.07
(0.36–1.20) (0.43–1.61) (0.66–1.74)
p = 0.17 p = 0.58 p = 0.77
Flexible sigmoidoscopy or colonoscopy in persons 50 and older 10 years (n = 1684) 0.87 1.15 1.05
(0.60–1.96) (0.83–1.59) (0.81–1.37)
p = 0.53 p = 0.39 p = 0.67
Cholesterol in persons 45 and older (n = 2024) 1.08 1.33 0.98
(0.59–1.89) (0.77–2.31) (0.68–1.42)
p = 0.79 p = 0.30 p = 0.93
Smoking counseling in past year (n = 611 smokers) 1.00 1.15 1.31
(0.55–1.82) (0.61–2.17) (0.74–2.33)
p = 0.99 p = 0.66 p = 0.36
Physical Activity counseling in past year (n = 2548) 1.23 1.06 1.16
(0.94–1.64) (0.78–1.43) (0.92–1.47)
p = 0.13 p = 0.70 p = 0.19
Nutrition counseling in past year (n = 2541) 1.47 1.05 1.25
(1.08–2.02) (0.76–1.45) (1.01–1.54)
p = 0.01 p = 0.76 p = 0.04
*Adjusted for age, gender, race, income ≥ 25,000, health insurance status, and health status. Statistical comparisons to group with more than 5 years
Discussion
Continuity can be measured in a number of ways. When measured as the length of the patient-physician relationship, we find continuity varies with patient demographics and with elements of satisfaction. Over half of adults of this study's 150 rural communities report seeing the same physician for more than five years. Those who saw the same physician for less than five years tended to be nonwhite, without health insurance, less educated, more often report income of less than $25,000 and more often report fair to poor health status. This group was also less likely to be satisfied with their overall health care and its quality and with components of the doctor-patient relationship.
Patients and their physicians value continuity [28-31]; in this study of rural participants, satisfaction appears associated with length of the continuity relationship, as similarly noted in other studies [11,18,32]. The elements of greater satisfaction found in persons with longer relationships with their physicians lend credence to the importance of the continuity relationship to patient outcomes. However, in this study, there appears to be a threshold effect at one or two years of continuity beyond which satisfaction does not rise significantly further.
Even a relationship length of two years is becoming difficult to maintain in the current U.S. healthcare system, especially in urban areas, with the pressures of competitive managed care plans which encourage patients and their employers to change health plans, and the growth of urgent care centers [33]. In the 1996–97 Community Tracking Study household survey, 17% of privately insured persons changed their health plan during the year prior to the survey [34]. Of those changing health plans, a little over half cited changes in their insurance as the reason for also changing their source of care. Increases in insurance premiums could contribute to further health provider switching.
Satisfaction, as an indicator of quality of care [28], has been found to affect other outcomes, including patient adherence to their physicians' recommendations [35]. Mothers are more likely to follow a physician's treatment recommendations for their child if she feels the physician is friendly and understands the complaint [36]. Patients with hypertension are more likely to adhere to treatment and have their blood pressure under control when the physician considers the patient an active participant in treatment [37].
Previous studies note the importance of having a usual source of care versus no source to the timely receipt of preventive services for younger adults [38,39]. However no differences in preventive care services were observed for older Americans in long-term relationships [9]. In our study preventive services outcomes did not differ significantly by length of continuity as well. For preventive services, it appears that having a usual source of care is important but no additional benefit comes with having a longer-term relationship with that source of care.
Respondents with one year or less continuity, interestingly, were more likely to report receiving nutrition counseling than those in the five-year or more continuity category. Possibly more patients are asking about nutrition issues in the first year or are specifically changing doctors to discuss nutrition. Another possibility is that the developing familiarity with the patient in a long term relationship may be associated with less vigilance by the physician or less counseling in areas in which the patient may have initially shown resistance [40].
Limitations
The cross-sectional design of this study limits our ability to attribute causation to the statistical relationships demonstrated or know the directions of any causal connections. We do not know if continuity results in higher trust and satisfaction or if the opposite is true. Our data are limited in that the outcomes of having received counseling and other preventive services were self-reported and may not accurately reflect care patients received. Telephone surveys also limit the population to those persons with a working phone, although weighting upwards for households with low incomes and minorities partially adjusts for this. The survey response rate of 51% was moderate but similar to other U.S. national telephone surveys [41]. Response bias is a possibility. This study addressed a rural population statistically representative of the U.S. rural South, thus its findings may not apply to urban and other regions of the U.S. or to other countries. However, we know of no reason to expect that the association between longer-term doctor-patient relationships and satisfaction differs elsewhere.
Conclusion
Over half of this rural population has seen the same physician for more than five years. Longer continuity was significantly related to aspects of the patient-physician relationship, specifically people's satisfaction with and confidence in their physicians, but not with one's likelihood of receiving recommended preventive services. Fostering long-term relationships between patients and their physicians may help promote the outcome of greater patient satisfaction with care.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
KD conceived and designed the study, performed the analyses, and drafted the manuscript. EA participated in the design, interpretation and helped draft the manuscript. DP participated in the design, analyses and critical review of the manuscript. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This work was supported by the Robert Wood Johnson Foundation Grant #036829. We acknowledge the thoughtful reviews and input given to us by the fellows and faculty of the UNC-Chapel Hill Southern Rural Access Program writers group. This work was presented at the Academy Health Annual Research Meeting, June 2004, San Diego, California.
==== Refs
McWhinney I Continuity of care J Fam Pract 1982 15 847 848 7130912
Fletcher RH O'Malley MS Earp JA Littleton TA Fletcher SW Greganti MA Davidson RA Taylor J Patients' priorities for medical care Med Care 1983 21 234 242 6827876
Donaldson MS, Yordy KD, Lohr KN, Vanselow NA Institute of Medicine. Primary Care: America's health in a new era 1996 Washington, DC , National Academy Press 3
Lambrew JM Defriese GH Carey TS Ricketts TC Biddle AK The Effects of Having a Regular Doctor on Access to Primary Care Med Care 1996 34 138 151 8632688 10.1097/00005650-199602000-00006
Mainous AG Baker R Love MM Gray DP Gill JM Continuity of Care and Trust in One's Physician: Evidence from Primary Care in the United States and the United Kingdom Family Medicine 2001 33 22 27 11199905
Freeman GK Richards SC How much personal care in four group practices? BMJ 1990 301 1028 1030 2249052
Ettner SL The relationship between continuity of care and the health behaviors of patients: does having a usual physician make a difference? Med Care 1999 37 547 555 10386567 10.1097/00005650-199906000-00004
Ettner SL The timing of preventive services for women and children: the effect of having a usual source of care Am J Public Health 1996 86 1748 1754 9003132
Weiss LJ Blustein J Faithful Patients: The Effect of Long-Term Physician-Patient Relationships on the Costs and Use of Health Care by Older Americans Am J Public Health 1996 86 1742 1747 9003131
Gill JM Mainous AG The Role of Provider Continuity in Preventing Hospitalizations Arch Fam Med 1998 7 352 357 9682689 10.1001/archfami.7.4.352
Wasson JH Sauvigne AE Mogielnicki P Frey WG Sox CH Gaudette C Rockwell A Continuity of Outpatient Medical Care in Elderly Men JAMA 1984 252 2413 2417 6481927 10.1001/jama.252.17.2413
O'Malley AS Mandelblatt J Gold K Cagney KA Kerner J Continuity of care and the use of breast and cervical cancer screening services in a multiethnic community Arch Intern Med 1997 157 1462 1470 9224225 10.1001/archinte.157.13.1462
Franks P Cameron C Bertakis KD On Being New to an Insurance Plan: Health Care Use Associated with the First Years in a Health Insurance Plan Ann Fam Med 2003 1 156 161 15043377 10.1370/afm.24
Bindman AB Grumbach K Osmond D Komaromy M Vranizan K Lurie N Billings J Stewart A Preventable Hospitalizations and Access to Health Care JAMA 1995 274 305 311 7609259 10.1001/jama.274.4.305
Weinberger M Oddone EZ Henderson WG Does Increased Access to Primary Care Reduce Hospital Readmissions N Engl J Med 1996 334 1441 1447 8618584 10.1056/NEJM199605303342206
Moy E Bartman BA Weir MR Access to Hypertenisve Care: Effects of Income, Insurance, and Source of Care Arch Intern Med 1995 155 1497 1502 7605151 10.1001/archinte.155.14.1497
Flocke SA Stange KC Zyzanski SJ The association of attributes of primary care with the delivery of clinical preventive services Med Care 1998 36 AS21 30 9708580 10.1097/00005650-199808001-00004
Kao AC Green DC Davis NA Koplan JP Cleary PD Patients' trust in their physicians: effects of choice, continuity, and payment method J Gen Intern Med 1998 13 681 686 9798815 10.1046/j.1525-1497.1998.00204.x
Freeman G Hjortdahl P What future for continuity of care in general practice? BMJ 1997 314 1870 1873 9224130
Saultz JW Defining and Measuring Interpersonal Continuity of Care Ann Fam Med 2003 1 134 143 15043374 10.1370/afm.23
Beachler M Holloman C Herman J Southern Rural Access Program: An Overview J Rural Health 2003 19 301 307 14526512
Salmon CT Nichols JS The next birthday method of respondent selection Public Opinion Q 1983 47 270 276 10.1086/268785
Community Tracking Study Household Survey, 1996-1997, and Followback Survey, 1997-1998: [United States]
Medical Expenditure Panel Survey
Penchansky R Thomas JW The Concept of Access: Definition and Relationship to Consumer Satisfaction Med Care 1981 19 127 140 7206846
Guide to Clinical Preventive Services, 3rd Edition, Periodic Updates
StataCorp. Stata Statistical Software: Release 8.0. College Station, TX: Stata Corporation 2003
Cleary PD McNeill T Patient Satisfaction as an Indicator of Quality Care Inquiry 1988 25 25 36 2966123
Lochman JE Factors related to Patients' Satisfaction with their medical care J Commun Health 1983 9 91 109 10.1007/BF01349873
Ware JE Snyder MK Dimensions of Patient Attitudes Regarding Doctors and Medical Care Services Med Care 1975 13 669 682 1152557
Nutting PA Goodwin MA Flocke SA Zyzanski SJ Stange KC Continuity of Primary Care: To Whom Does it Matter and When? Ann Fam Med 2003 1 149 155 15043376 10.1370/afm.63
Hjortdahl P Laerum E Continuity of care in general practice: effect on patient satisfaction BMJ 1992 304 1287 1290 1606434
Donaldson MS Continuity of Care: A Reconceptualization Med Care Res Rev 2001 58 255 290 11523291
Cunningham PJ Kohn L Health Plan Switching: Choice or Circumstance? Health Affairs 2000 19 158 164 10812794 10.1377/hlthaff.19.3.158
Becker MH Maiman LA Strategies for enhancing patient compliance J Commun Health 1980 6 113 135 10.1007/BF01318980
Francis V Korsch BM Morris MJ Gaps in doctor-patient communication: Patients' response to medical advice N Engl J Med 1969 280 535 540 5764453
Schulman BA Active patient orientation and outcomes in hypertensive treatment: Application of a socio-organizational perspective Med Care 1979 17 267 280 763004
Doescher MP Saver BG Fiscella K Franks P Preventive Care: Does Continuity Count? J Gen Intern Med 2004 19 632 637 15209601 10.1111/j.1525-1497.2004.21150.x
Ettner SL The Relationship between Continuity of Care and the Health Behaviors of Patients: Does Having a Usual Physician Make a Difference? Med Care 1999 37 547 555 10386567 10.1097/00005650-199906000-00004
Broom DH Familiarity breeds neglect? Unanticipated benefits of discontinuous primary care Fam Pract 2003 20 503 507 14507788 10.1093/fampra/cmg501
Centers for Disease Control and Prevention (CDC). "2004 Behavioral Risk Factor Surveillance System Summary Data Quality Report." Atlanta, Georgia: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, 2005
|
16202146
|
PMC1262705
|
CC BY
|
2021-01-04 16:29:13
|
no
|
BMC Fam Pract. 2005 Oct 4; 6:40
|
utf-8
|
BMC Fam Pract
| 2,005 |
10.1186/1471-2296-6-40
|
oa_comm
|
==== Front
BMC Fam PractBMC Family Practice1471-2296BioMed Central London 1471-2296-6-411620214810.1186/1471-2296-6-41Research ArticleThe future prospects of Lithuanian family physicians: a 10-year forecasting study Starkiene Liudvika [email protected] Kastytis [email protected] Zilvinas [email protected] Jack [email protected] Department of Preventive Medicine, Kaunas University of Medicine, Mickeviciaus str. 9, LT-44307 Kaunas, Lithuania2 Graduate Program in Health Services Administration, Xavier University, Ohio, USA2005 4 10 2005 6 41 41 21 6 2005 4 10 2005 Copyright © 2005 Starkiene et al; licensee BioMed Central Ltd.2005Starkiene et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
When health care reform was started in 1991, the physician workforce in Lithuania was dominated by specialists, and the specialty of family physician (FP) did not exist at all. During fifteen years of Lithuania's independence this specialty evolved rapidly and over 1,900 FPs were trained or retrained. Since 2003, the Lithuanian health care sector has undergone restructuring to optimize the network of health care institutions as well as the delivery of services; specific attention has been paid to the development of services provided by FPs, with more health care services shifted from the hospital level to the primary health care level. In this paper we analyze if an adequate workforce of FPs will be available in the future to take over new emerging tasks.
Methods
A computer spreadsheet simulation model was used to project the supply of FPs in 2006–2015. The supply was projected according to three scenarios, which took into account different rates of retirement, migration and drop out from training. In addition different population projections and enrolment numbers in residency programs were also considered. Three requirement scenarios were made using different approaches. In the first scenario we used the requirement estimated by a panel of experts using the Delphi technique. The second scenario was based on the number of visits to FPs in 2003 and took into account the goal to increase the number of visits. The third scenario was based on the determination that one FP should serve no more than 2,000 inhabitants. The three scenarios for the projection of supply were compared with the three requirement scenarios.
Results
The supply of family physicians will be higher in 2015 compared to 2005 according to all projection scenarios. The largest differences in the supply scenarios were caused by different migration rates, enrolment numbers to training programs and the retirement age. The second supply scenario, which took into account 1.1% annual migration rate, stable enrolment to residency programs and later retirement, appears to be the most probable. The first requirement scenario, which was based on the opinion of well-informed key experts in the field, appears to be the best reflection of FP requirements; however none of the supply scenarios considered would satisfy these requirements.
Conclusion
Despite the rapid expansion of the FP workforce during the last fifteen years, ten-year forecasts of supply and requirement indicate that the number of FPs in 2015 will not be sufficient. The annual enrolment in residency training programs should be increased by at least 20% for the next three years. Accurate year-by-year monitoring of the workforce is crucial in order to prevent future shortages and to maintain the desired family physician workforce.
==== Body
Background
Lithuania is situated in the Eastern Europe and is the biggest of three Baltic States, with an area of 65,000 square kilometres. Its population of 3.4 million is predominately Lithuanian with the ethnic minorities of which 6.7% are Polish, 6.3% are Russians, and 1.2% are Byelorussians. Lithuania was a part of Soviet Union until it declared its independence in 1990 [1].
The start of health care reform in 1991 found the supply of physician human resources in Lithuania dominated by specialists, and the specialty of family physician (FP) did not exist. The Soviet model of health care, which existed in Lithuania until independence, was based exclusively on the exaggerated focus and development of the hospital level, whereas the need to develop the primary health care was ignored [2].
The main goal of the health reform – to shift health care services to primary health care level – was severely hindered by the absence of FPs to provide these services. At the beginning of the reform effort, the specialty of FP was not popular with physicians as it was considered to be non-prestigious, and the system of reimbursement of FPs was inefficient [3]. In 1992, retraining courses for practicing district physicians and paediatricians to become family physicians were launched. However, these courses had limited success due to the lack of teachers in the field. Significant changes in the system started in 1994, when regulations for FPs' training in residencies were adopted. In parallel with regular residency programs, interruptive residency programs for retraining physicians to become FPs were started at Kaunas University of Medicine and Vilnius University. The aim of interruptive residency programs was to retrain practicing district physicians and paediatricians into FPs, using well-structured 10-month program. It was broken down into blocks of 2 weeks to ensure that physicians did not have to leave their jobs for a long period of time. During the period of 1994–2003, 1,908 FPs were trained, of which 77% were trained in the interruptive residencies or retraining courses. Enrolment in regular training programs was approximately 35 medical residents in 2000–2004; in addition, approximately 66 medical residents were annually admitted to training in interruptive residency programs [4]. However, in 2004, admission to interruptive residency was stopped, as it was decided that the sufficient number of FPs have been trained.
Even though the number of FPs has increased very rapidly during the last fifteen years reaching 48.6 per 100,000 population in the beginning of 2005, FPs as percentage of all physicians was only about 12%. In the other countries of the European Union, this percentage was much higher (with exception of Latvia where it was 15%) – the United Kingdom and Germany were around 30%, and the average of 25 countries of the European Union was 23%. Interestingly enough, Lithuania has one of the highest physician to 100,000 population ratios in the EU, while having one of the lowest FPs to 100,000 population ratios (Table 1) [5,6].
Table 1 Number of FPs and physicians per 100,000 population in 1997 and 2003
Country FPs per 100,000 population Physicians per 100,000 population
1997 2003* 1997 2003*
Lithuania 7.0 48.6 414.3 391.1
Latvia 16.1 45.2 296.0 298.5
Estonia 52.3 62.9 323.9 314.1
Germany 109.6 104.3 312.8 336.9
United Kingdom 60.4 62.8 188.6 212.6
EU-25 56.2 64.1 271.9 278.4
* Or the latest available year
The demographic characteristics of Lithuanian FPs are very similar to those of the overall physician population, except for the age structure. While 19.8% of the physicians were older than 60 years; only 2.5% of FPs were older than 60 years, and only 4.9% older than 50 years [7]. FPs are unequally distributed in ten counties of Lithuania, with the highest FP to population ratios being in the cities and the lowest ratios in rural areas. The physician workforce in Lithuania has been traditionally dominated by women (70%); in family medicine that proportion is even higher (84.9%) [5,7].
Since 2003, the Lithuanian health care sector has undergone restructuring in order to optimize the network of health care institutions as well as the delivery of services; specific attention has been paid to the development of services provided by FPs and shifting more health care services from the hospital level to primary health care level. According to current legislation, one FP should serve not more than 2,000 inhabitants [8], however in 2005 one FP served on average 2,058 inhabitants. As defined in the Restructuring Strategy of Health Care Institutions, the number of visits to FPs should increase by 18.7% by the end of 2005 compared with 2003. In order to meet the goals of the strategy, a sufficient number of FPs will need to be available [9].
In this paper we project the supply of and the requirement for FPs in Lithuania until 2015. Here we provide ten-year planning projections, which are essential if an adequate workforce of FPs is to be available to meet future needs. Correspondingly, changes in the number of medical residents enrolled in training programs should be managed to meet the requirements of this plan.
Methods
The standard approach to planning of health human resources was used in this study. It included the projection of supply, the projection of requirements and an analysis of the gap between supply and requirement [10,11].
Supply projections
A computer spreadsheet simulation model was used to project FP supply from 2006 to 2015 [12]. Three supply projection scenarios were used for planning purposes (Table 2). The supply of active FPs on January 1, 2005, the projected Lithuanian population by 2015, losses from the profession (due to death, retirement and migration), entry into profession from residency programs (adjusted for the drop out rate), and the duration of residency studies in family medicine (3 years) were used in this model.
Table 2 Supply projection scenarios and assumptions
Variables First scenario Second scenario Third scenario
Duration of residency studies 3 years 3 years 3 years
Population projections Optimistic Medium Pessimistic
Annual mortality rate 0.47% 0.47% 0.47%
Annual retirement rate At the age of 71 years At the age of 71 years At the age of 66 years
Annual migration rate Optimistic rate Medium rate Pessimistic rate
Enrolment in residency studies Increased by 20% (46) At the level of 2004 (38) Decreased by 20% (30)
Drop out rate during residency 1% 1.5% 2%
Population projections
Projected Lithuanian population by 2015 according to three possible scenarios: medium, optimistic and pessimistic was obtained from the Department of Statistics [13].
Annual mortality rate
There was no accessible data on the annual mortality rate of FPs, therefore we used a weighted average of the age-specific (25–64 years) and the gender-specific (84.9% women and 15.1% men) mortality rates of the general Lithuanian population obtained from the Department of Statistics [1]. This assumption tends to underestimate the annual mortality rate, since according to other studies the physician mortality rate is usually somewhat higher [10,14].
Annual retirement rate
Since no reliable data were available on the average annual retirement rate of Lithuanian FPs, we applied the method by Pace et al in order to calculate it [15]. We used the data from Physician License Registry [7]. The retirement age was set to be 66 years and then assumption was made that one fifteenth of the group of FPs aged more than 50 years would retire annually. In the other scenario the retirement age was set to be 71 years and then assumption was made that one fifteenth of the group of FPs aged more than 55 years would retire annually.
Annual migration rate
Source data for the annual drop out from profession due to migration was obtained from the Ministry of Health (data for the period of May 1, 2004 – April 30, 2005). The pessimistic rate was calculated using the following data: dividing the number of FPs who emigrated during the first year by the total number of FPs in 2005, and multiplying by 100 in order to get the annual migration rate. This rate was taken as the worst estimate because it was unlikely that the migration rate, which was observed 12 months after joining the European Union, would remain at the same level for the period of ten years. Medium rate was assumed to be half of the pessimistic rate, and optimistic rate – half of the medium rate.
Annual enrolment in residency programs and drop out rate
Numbers of annual medical resident enrolment in FPs' training programs in 2000–2004 as well as the number of graduates in 2000–2004 were obtained from the Ministry of Health. Enrolment numbers in the residency programs starting with 2005 were left at the level of 2004, increased or decreased (depending on the scenario) and then converted into future annual number of graduates using three different drop out rates (1%, 1.5% and 2%) established by Lovkyte [14].
Requirement projections
Requirement of FPs until 2015 was estimated using three different approaches:
1) The first approach was based on the survey conducted in 2000 by use of the Delphi survey technique. To determine the goal of FP workforce planning that should be reached by 2015, we surveyed the deans of the Faculties of Medicine, members of the National Board of Health, county chief physicians, directors of the Territorial Sickness Funds and the State Sickness Fund, and representatives of the Ministry of Health and the WHO Liaison Office. Out of a total of 34 questionnaires sent out, 23 were completed and returned in the first round. In the second round, the questionnaires were sent only to the 23 respondents of the first round, of whom 15 responded [2].
2) The second approach was based on the number of visits to FPs in 2003 adjusted by growth of these visits (by 18.7% until the end of 2005), defined in the Restructuring Strategy of Health Care Institutions [9]. Unfortunately, further goal on estimated growth is not available; therefore we assumed that there will be no further growth in number of visits to FPs in 2006–2015. Breakdown of health care services provided in 2003 by patients' age groups and gender was multiplied by projected changes in the population. Afterwards the projected number of visits was increased by 18.7% and converted into a number of FPs, using the number of visits to one FP in 2003.
3) The third scenario was based on the provision by the Ministry of Health that one FP should serve no more than 2,000 inhabitants [8].
Gap analysis
A gap analysis was performed comparing supply and requirement projections in order to identify future shortages or surpluses of FPs. We also attempted to determine the factors that had the largest impact on future FP workforce.
Results
Supply projections
Prior to projection exercise we had to make some calculations regarding the mortality, the retirement and the migration of FPs.
Population projections
Lithuania is characterized by a declining and aging population. It is expected that by 2015 the number of inhabitants will decrease according to all scenarios: from 100,000 according to the optimistic scenario to 250,000 according to the pessimistic scenario [13].
Annual mortality rate
The mortality of women in the 25–64 age group was 0.37% and the mortality of men was 1.06% in 2004. The weighted average of mortality in the 25–64 age group adjusted by gender (84.9% women and 15.1% men) was 0.47% [1].
Annual retirement rate
Family practice is a young physician specialty with relatively few physicians being at retirement age [7]. Only four physicians per year could be expected to retire using 71 years as the retirement age (since there were 54 physicians aged 56 years or older, one fifteenth of them should be at retirement age or older). Using a lower retirement age – 66 years, on average 11 physicians per year could be expected to retire (since there were 169 physicians aged 51 years or older, one fifteenth of them should be at retirement age or older).
Annual migration rate
The pessimistic annual migration rate was calculated to be 2.2%, using the following reasoning: 36 FPs left the country during 12 months, and there were 1,665 FPs in 2005. Medium rate was calculated to be 1.1% and optimistic rate 0.6%.
Annual enrolment in residency programs and drop out rate
Different annual enrolment rates (30, 38 and 46) also resulted in different gains to profession, varying from 444 to 569.
Figure 1 summarizes the projections of FPs' supply according to three scenarios. According to the first scenario, FPs-to-population ratio would be higher by 24.6% in 2015 than it was in 2005. The second scenario forecasts increase by 18.6%. The third projection also indicated increase by 4.5%, which would result in the ratio of 51.1 per 100,000 population.
Figure 1 Projections of FPs supply (FPs to 100,000 population ratio) according to the three scenarios.
All three supply projections were equally influenced by the annual mortality rate (Table 3). The biggest differences in supply scenarios were caused by different migration rates, enrolment numbers in training programs and the retirement age. If the retirement age was set at 66 years, 110 FPs could be expected to retire during 10-year period, compared with only 40, if retirement was 71 years. If migration remained stable at the current rate, 381 FPs could be expected to leave Lithuania in 2006–2015. The rise in the supply curves until 2008 was mainly caused by the higher annual number of graduates than in the consecutive years (Figure 1.). In addition to the graduates of regular residency programs (30–46 depending on scenario), it included graduates of interruptive residency programs (around 65 each year). Later fall was related to discontinued admission to these programs since 2004. The drop out rate from training programs is also reflected in the number of graduates, but due to very small numbers it had little influence on the future supply of FPs.
Table 3 Variables / assumptions and their influence on estimated losses / gains to the profession of FPs in 2006–2015
Variable Assumption Estimated losses from profession during 2006–2015 Estimated gains to profession during 2006–2015
Annual mortality rate 0.47% 88 -
Annual retirement rate At the age 71 years 40 -
At the age of 66 years 110 -
Annual migration rate 0.6% 116 -
1.1% 207 -
2.2% 381 -
Enrolment in residency programs Decreased by 20% - 444
At the level of 2004 - 503
Increased by 20% - 569
Drop out rate during residency 1% 5 -
1.5% 6 -
2% 6 -
Requirement projections
According to the first scenario, the requirement for FPs was 67.0 per 100,000 population [2]. This scenario would also have one FP serving 1,500 inhabitants.
The second scenario was based on the number of visits and their growth (Table 4). In 2003, there were 1,500 family physicians, who were visited more than 5.4 million times. Table 4 indicates the breakdown of population and visits by gender and age groups in 2003 and 2015. Lithuania is characterized by a declining and aging population. It is expected that by 2015 the number of inhabitants will decrease on average by 185 thousand; almost 100 thousand of them will be in the age group under 18 years. Logically, the number of visits should also decrease and if the goal to increase the number of visits by 18.7% until the end of 2005, as defined in the Restructuring Strategy of Health Care Institutions, was not taken into account, the requirement for FPs was 44.0 per 100,000. However, taking into account increasing number of visits, 52.2 FPs per 100,000 population would be needed to ensure the proper provision of services.
Table 4 Population and number of visits to FPs in 2003 and 2015
Gender, age group Population in 2003 (in thousands) Number of visits in 2003 (in thousands) Projected population in 2015 (in thousands) Projected number of visits in 2015 (in thousands) Projected number of visits in 2015, adjusted with 18.7% increase (in thousands)
Males:
0–18 438.5 736.3 336.5 565.0 670.7
19–44 650.5 537.6 597.1 493.5 585.8
45–64 353.1 496.1 409.5 575.5 683.0
>65 175.2 401.2 183.4 419.8 498.4
Totally, males 1617.3 2171.2 1526.5 2053.8 2437.9
Females:
0–18 418.8 720.9 320.0 550.0 653.9
19–44 660.5 734.6 602.5 670.0 795.3
45–64 431.5 863.2 473.1 946.5 1123.6
>65 334.5 933.3 355.3 991.5 1176.9
Totally, females 1845.3 3252.0 1750.9 3158.0 3749.7
Totally 3462.6 5423.2 3277.4 5212.8 6187.6
According to the third scenario, the requirement was based on the ruling by the Ministry of Health that one FP should serve not more than 2,000 inhabitants, i.e. 50.0 FPs per 100,000 population would be required in 2015.
Analysis of a gap between supply and requirement
In our last step we compared the three supply projections with three requirement projections. As shown in Table 5, the third requirement scenario would be exceeded by all three supply scenarios. The requirement indicated by the second scenario would be exceeded by the supply projected according to the first and the second scenarios. None of the supply scenarios would reach the requirement indicated by the first scenario (67.0); even the supply according to the first scenario would be lower (60.9 FPs per 100,000).
Table 5 Gap between supply and requirement projections (FPs per 100,000 population)
Scenarios Requirement scenarios for FPs per 100,000 population in 2015
First (67.0) Second (52.2) Third (50.0)
Supply scenarios of FPs per 100,000 population in 2015 First (60.9) -6.1 8.7 10.9
Second (58.0) -9.0 5.8 8.0
Third (51.1) -15.9 -1.1 1.1
NB. Minus indicates that the requirement will be higher than the supply
Discussion
Although recognized as an important part of health care system reform, the comprehensive planning of an FP workforce has not been a high priority in Lithuania over the last fifteen years. The national policy has been mostly limited to establishing training and retraining programs. This study is the first attempt to provide ten-year planning projections essential to ensure an adequate FP workforce in the future.
As mentioned earlier in this article, Lithuania started with no family physicians after restoration of independence. In 1996 an international expert group led by Corder set a target to increase the percentage of FPs to 20% of the overall physician population by 2005 [16]; however the goal was not reached and currently FPs make up only 12% [5]. Another target to train 2,400 FPs by the year 2010 was set in the primary health care development program, adopted by the Ministry of Health in 2000 [17]. If enrolment in residency programs was increased by 20% (as indicated by the first supply scenario), likely this training target could be reached; nevertheless the number of practicing FPs might be not sufficient. For example, out of 1,908 FPs who graduated through 2004, 91.2% held a license, but only 78.6% were practicing [5,7].
While family medicine in Lithuania has been dominated by women (84.9%), unlike other countries, women do not tend to work part-time or see fewer patients than male physicians, mainly due to an unfavourable payment system [7]. Maternal leave is also basically limited to one year due to financial disincentives to prolong it. The gender composition of medical school graduates has remained quite steady over the last decade, and it is unlikely that the number of women FPs will change in the future [2].
The migration of FPs should be monitored with particular concern. According to a survey of Lithuanian physicians conducted in 2004, 26.8% intended to leave the country and 3.8% have made a definite decision to do so. Younger age was a risk factor for leaving and is particularly important in case of FPs, since the vast majority of them are of young age. As the main reasons for leaving, salary and professional career differences were identified, and it is unlikely that this gap between Lithuania and old European Union members will diminish in a few years [18]. It is realistic that 1.1% of FPs could emigrate annually. The emigration rate of 2.2% is unlikely, because it assumes that the emigration rate, which was observed during the 12 months after accession to the European Union, would be sustained for the ten-year period.
Due to social uncertainties and an unfavourable retirement policy Lithuanian physicians are reluctant to retire at an earlier age; according to Lovkyte, 45.7% of all Lithuanian physicians were still practicing at the age of 66 and later, and it is unlikely that all would retire at once [14].
The number of visits to family physicians increased almost two-fold from 3 million in 2001, the first year for which data on the number of services provided is available, to 5.4 million in 2003. During the same period the number of visits to specialist physicians working in the primary health care setting decreased from 4.8 million in 2001 to 4.2 million in 2003 [19]. FPs currently have very busy practices, allocating an average of 10 minutes to each patient. Recent studies indicated that total job satisfaction of family physicians in Lithuania was relatively low. Compensation, high job demands, social status, and high patient load were among the key factors that caused their dissatisfaction and were significant predictors of psychosocial stress. Unfavourable job environment can also reduce the attractiveness of the profession and result in talented medical graduates choosing other medical or non-medical specialties [20,21].
In our opinion, the second supply scenario, which takes into account a 1.1% annual migration rate, a stable enrolment to residency programs and a later retirement, is the most probable. The first requirement scenario is the best as it was based on the opinion of well-informed key experts in the field. These experts took into account not only the historic number of visits, the short-term changes in the number of visits or minimal needs of the population, but also complex factors having an impact on this profession as well as future changes in health care system.
Several recommendations could be suggested, however some of them would be difficult to implement. As a general course of action, majority of the planning organizations favour adjustments to enrolment to training programs as the best long-term solution to any anticipated imbalances between expected supply and estimated requirement. This study is not exceptional in this sense and a recommendation is made to increase the enrolment to FPs' training program by 20% at least for three years in order to prevent future shortages of FPs. In the future, projections should be updated and further recommendations should be drawn. Another recommendation would be to increase retention rates in the profession, via implementing reformed and significantly improved financial and non-financial incentive system (the examples would include increased per capita reimbursement for FPs mixed with fee-for-service payments, better working conditions, lower patient load, improved access to continuous medical education courses, etc.). This could also contribute to lower emigration rates. Other recommendations such as assigning more duties to professional nurses, who would be trained to undertake part of family physicians' duties, go beyond the scope of this study.
Conclusion
Family medicine in Lithuania will face several challenges in coming years. There will likely be a lack of approximately 9 FPs per 100,000 (or 300 FPs for the whole population) in 2015, which should be considered as increased duties and responsibilities are assigned to them. Job satisfaction of FPs is relatively low, with compensation, high job demands, social status, and high patient load as key factors in causing dissatisfaction and psychosocial stress.
We recommend that the enrolment in residency programs be increased by 20% at least for the next three years. Special attention should be paid to monitoring of retirement and retention rates in profession. Every fifth graduate was not practicing in Lithuania, as he / she either chose a better paid job or moved to another country. A better retention program would reduce training requirements to achieve the desired workforce supply. Achieving a balance between the supply and the requirements is very complex, but important task in order to ensure the appropriate and efficient functioning of the health care system in the future. Requirement and supply projections should continue to be monitored annually, and be amended, if new trends in any of the FP characteristics emerge or projection assumptions change. Without some more comprehensive registry or means to link the existing databases, complete information on the FP workforce in Lithuania will remain difficult if not impossible to obtain.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
All authors participated in designing the study, making data analysis, writing the original text and read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Financial support for this study was provided by the program "Strategic planning of health human resources in Lithuania in 2003–2020" at the Ministry of Health of the Republic of Lithuania.
==== Refs
Department of Statistics of the Republic of Lithuania Demographic yearbook 2003 Vilnius 2004
Lovkyte L Reamy J Padaiga Z Physicians Resources in Lithuania: Change Comes Slowly Croat Med J 2003 44 207 13 12698513
Valius L Grabauskas V Primary health care in Lithuania: legal basis, process of reform and problems Health Policy in Lithuania in the 21st century Proceedings of the third National Conference on Health Policy in Lithuania 2000 Kaunas, University of Medicine Press 90 95
Ministry of Health of Republic of Lithuania Training of family physicians in 1994–2003 Vilnius 2004
Lithuanian Health Information Center Population health and activities of health care facilities in Lithuania Vilnius 2003
World Health Organization Regional Office for Europe European health for all database (HFA-DB) 2005 World Health Organization Regional Office for Europe
Ministry of Health of Republic of Lithuania Physician License Registry data for January 1, 2004 Report No: 2004-03-18-48/8 Vilnius 2004
Ministry of Health of the Republic of Lithuania On ratification of service list, basic prices, organization and payment for primary health care services Valstybes zinios 2000 50 1447
Ministry of Health of the Republic of Lithuania On ratification of restructuring plans of counties' health care institutions Valstybes zinios 2004 30 999
Australian Medical Workforce Advisory Committee The general practice workforce in Australia: supply and requirements 1999–2010 AMWAC Report 20002 Sydney 2000
Australian Medical Workforce Advisory Committee Medical workforce planning in Australia Aust Health Rev 2000 23 8 26 11256274
Dewdney J WPRO/RTC health workforce planning workbook 2000 University of New South Wales, Sydney
Department of Statistics of the Republic of Lithuania Prognoses of Lithuanian population 2005–2030 Vilnius 2004
Lovkyte L Planning projections of physician supply and requirement in Lithuania until 2015: doctoral dissertation 2004 Kaunas University of Medicine Press
Pace KT Provan JL Jewett MA The urology workforce in Ontario for the 21st century: feast or famine? Can J Surg 1999 42 181 9 10372014
Corder DW Planning of physician supply in Lithuania Report to the Ministry of Health of Republic of Lithuania Vilnius 1996
Ministry of Health of the Republic of Lithuania On adoption of primary health care development program Order of the Minister of Health No 441 Vilnius 2003
Stankunas M Lovkyte L Padaiga Z The survey of Lithuanian physicians and medical residents regarding possible migration to the European Union Medicina (Kaunas) 2004 40 68 74 14764985
Kaunas University of Medicine Strategic planning of health human resources in Lithuania in 2003–2020 Annual report to the Ministry of Health Kaunas 2004
Buciuniene I Blazeviciene A Bliudziute E Health care reform and job satisfaction of primary health care physicians in Lithuania BMC Family Practice 2005 6 10 15748299 10.1186/1471-2296-6-10
Vanagas G Bihari-Axelsson S The factors associated to psychosocial stress among general practitioners in Lithuania. Cross-sectional study BMC Health Services Research 2005 5 45 15946388 10.1186/1472-6963-5-45
|
16202148
|
PMC1262706
|
CC BY
|
2021-01-04 16:29:13
|
no
|
BMC Fam Pract. 2005 Oct 4; 6:41
|
utf-8
|
BMC Fam Pract
| 2,005 |
10.1186/1471-2296-6-41
|
oa_comm
|
==== Front
BMC GenetBMC Genetics1471-2156BioMed Central London 1471-2156-6-481619427810.1186/1471-2156-6-48Research ArticleEpigenetic predisposition to expression of TIMP1 from the human inactive X chromosome Anderson Catherine L [email protected] Carolyn J [email protected] Department of Medical Genetics, University of British Columbia, 2350 Health Sciences Mall, Vancouver, BC, CANADA V6T 1Z32005 29 9 2005 6 48 48 4 4 2005 29 9 2005 Copyright © 2005 Anderson and Brown; licensee BioMed Central Ltd.2005Anderson and Brown; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
X inactivation in mammals results in the transcriptional silencing of an X chromosome in females, and this inactive X acquires many of the epigenetic features of silent chromatin. However, not all genes on the inactive X are silenced, and we have examined the TIMP1 gene, which has variable inactivation amongst females. This has allowed us to examine the features permitting expression from the otherwise silent X by comparing inactive X chromosomes with and without TIMP1 expression.
Results
Expression was generally correlated with euchromatic chromatin features, including DNA hypomethylation, nuclease sensitivity, acetylation of histone H3 and H4 and hypomethylation of H3 at lysines 9 and 27. Demethylation of the TIMP1 gene by 5-azacytidine was able to induce expression from the inactive X chromosome in somatic cell hybrids, and this expression was also accompanied by features of active chromatin. Acetylated histone H3 continued to be observed even when expression was lost in cells that naturally expressed TIMP1; while acetylation was lost upon TIMP1 silencing in cells where expression from the inactive X had been induced by demethylation. Thus ongoing acetylation of inactive X chromosomes does not seem to be simply a 'memory' of expression.
Conclusion
We propose that acetylation of H3 is an epigenetic mark that predisposes to TIMP1 expression from the inactive X chromosome in some females.
==== Body
Background
Studies have shown considerable individual variability in the level of expression of genes (e.g. [1,2]). In general, however, humans cannot tolerate imbalances for expression of substantial numbers of genes, as demonstrated by the lethality of the majority of chromosomal aneuploidies. Aneuploidy for the sex chromosomes is better tolerated, being observed in approximately 1/500 births [3], presumably because all but one X chromosome is inactivated. X chromosome inactivation ensures the dosage equivalence of X-linked genes between females who have two X chromosomes and males who have a single X chromosome and the sex-determining Y chromosome [4]. However, more than 15% of human X-linked genes escape inactivation, being expressed from both the active and inactive X chromosome [5]. While such an escape from inactivation may maintain dosage equivalence for X-linked genes with Y homologs, the majority of human genes that escape inactivation no longer have functional Y equivalents, and thus may show relative overexpression in females (reviewed in [6]). Substantially fewer genes have been shown to escape inactivation in mice. Although this species difference in expression could reflect less extensive murine expression surveys, a reduced number of genes escaping inactivation is supported by the less drastic phenotype caused by monosomy of the X chromosome in mice (reviewed in [7]). In humans, the over or under-expression of genes that escape inactivation is a major contributor to the phenotypes associated with X chromosome aneuploidies, but may also contribute to expression differences between chromosomally normal males and females (e.g. [8]).
The study of genes that escape X inactivation can provide insight into such phenotypes, as well as contributing to our understanding of epigenetic silencing mechanisms. Inactivation occurs early in mammalian development, and the stable silencing of the X chromosome involves the acquisition of many features of heterochromatin. It is not known if escape from inactivation is a resistance to the initial silencing event, or rather reflects a high frequency of reactivation of an initially silenced gene, as data appear to support both possibilities. Many genes that escape inactivation in humans are clustered together, which may be indicative of regions that are resistant to the initial signal (e.g. [9]). However, analysis of Smcx, one of the few mouse genes expressed from the murine inactive X, has shown reactivation of the gene during early development [10]. Surprisingly, recent results have shown that Smcx has a histone modification pattern suggested to demarcate biallelically rather than monoallelically-expressed (imprinted and other X-linked) genes [11], suggesting that the gene is committed to escape inactivation prior to undergoing inactivation.
In addition to the genes that are subject to, or escape from, inactivation, there are some human genes that show heterogeneous X chromosome inactivation, being expressed from the inactive X in some females, but silenced on the inactive X in others [5,12,13]. Such genes provide an opportunity to study the same region when silent or active on an inactive X chromosome; and can thus provide insights into the features allowing expression from the inactive X chromosome. The human inactive X chromosome maintains its silent status when isolated in a mouse/human somatic cell hybrid, providing a model system to study the inactive X chromosome apart from its active counterpart. The largest survey of gene expression from the inactive X chromosome [5] analysed expression in a panel of nine inactive-X containing hybrids. That study defined heterogeneous inactivation as expression in three to six of the nine hybrids, which was observed for 60 of the 624 X-linked genes analysed. This variable expression is not restricted to the hybrid system, but has also been demonstrated in cells from females [5,12,13].
We now report the further characterization of one of these genes, the X-linked tissue inhibitor of metalloproteinases, TIMP1, located in Xp11.23. Our previous studies have demonstrated that TIMP1 is variably expressed from the inactive X in both somatic cell hybrids and human females. When TIMP1 is expressed from the inactive X, flanking genes (including ARAF1 and ELK1 which lie ~20 and 55 kb from TIMP1, respectively) are not expressed, suggesting that expression is being controlled in a gene-specific rather than regional fashion [12]. Quantitative RNase protection assays showed substantial variability in expression levels from the active X, precluding using expression levels to determine inactive X expression in females. Further studies in hybrids demonstrated that TIMP1-expressing clones were unstable and that methylation does not appear to be the principle controlling feature allowing variable expression of TIMP1 from the inactive X chromosome [14]. In this study we report the analysis of other features characteristic of an inactive X to determine which features might predispose TIMP1 to expression from the inactive X chromosome in a subset of females.
The inactive X acquires many of the general features of heterochromatin (reviewed in [15]), and we have now examined replication timing, nuclease sensitivity, and histone modifications for TIMP1. Late replication of the inactive X at the chromosome level is observed after Giemsa staining following bromodeoxyuridine incorporation [16]. Regions such as the distal and proximal short arm that contain a large proportion of genes that escape inactivation are not delayed in their replication, supporting a regional basis to escape from inactivation. Replication of individual X-linked genes has been analysed by fluorescent in situ hybridization (FISH) or amplification of BuDR incorporated DNA after flow cytometry. Although correspondence is not complete between different techniques [17], such methods have generally shown that genes that escape inactivation are early replicating (reviewed in [18]). DNase sensitivity is a general feature of active chromatin, and promoters of genes subject to inactivation have been seen to be less available for digestion by nucleases (e.g. [19]). Modifications to the histones associated with the inactive X are reflective of both general heterochromatic changes and ones specific to the facultative heterochromatin of the inactive X. Using antibodies to acetylated histones the inactive X chromosome stains very palely [20], while antibodies to histone H3 methylated at lysine (K) 9 (H3mK9) are generally associated with heterochromatin and those to methylated lysine 27 (H3mK27) specifically mark the inactive X chromosome [21,22]. Chromatin immunoprecipitation (ChIP) has revealed that genes subject to inactivation show limited acetylation and elevated histone methylation (H3mK9/27) at their promoters relative to genes escaping inactivation [23,24].
Results
Replication timing of TIMP1 in human cells
To evaluate the influence of replication timing upon expression of TIMP1, cell lines from a female with and a female without TIMP1 expression from the inactive X [12] were analysed by DNA FISH using probes for the TIMP1 and HPRT loci. As shown in the schematic of Figure 1, cells that have not replicated either X chromosome will show two single signals, and almost one half of the cells were of this type. Of the cells with at least one double signal, indicating replication of the locus, considerable asynchrony was observed for the HPRT locus, as 37% of all cells analysed showed single-double (SD) signals and 17% showed the double-double (DD) signals. The TIMP1 locus generally showed lower asynchrony of replication than HPRT (28% of cells were S/D vs 37% for HPRT, P < 0.0001). The cell line with TIMP1 expression from the inactive X (cell line 2 – HSC593) showed a small trend towards having a lower percentage of cells that replicated asynchronously than the GM07059 cell line which does not have TIMP1 expression from the inactive X (25% vs 29%); however this difference was not significant (P = 0.17). The observed shift in replication timing, while not statistically significant, may be a contributing factor in the expression of TIMP1 from the inactive X. However, it is likely that additional changes are also involved, as TIMP1 is within 20 kb of the ARAF1 gene and 55 kb from ELK1 (see Figure 2A), and thus at least one of these genes is likely to share a replication domain with TIMP1 despite remaining silent when TIMP1 is expressed [12].
Figure 1 Replication asynchrony for the TIMP1 (T – white bars) and HPRT (H – grey bars) genes assessed by FISH. The bars show the frequency of nuclei exhibiting unreplicated (single/single); asynchronously replicated (single/double) and completely replicated (double/double) signals as shown in the schematic below. The TIMP1 and HPRT probes were individually hybridized to interphase nuclei of the same preparations. Two human female lymphoblast cell lines were examined for the approximate degree of replication asynchrony, one that inactivated TIMP1 (1 – GM07059) and one that expressed TIMP1 from the Xi (2 – HSC593). For TIMP1, 313 cells were counted from 3 separate cell harvests for cell line 1 (T1) and 217 cells for cell line 2 (T2) from two separate harvests. 144 cells were examined for HPRT from 2 slides of one harvest for cell line 1 (H1) and 97 cells from a single harvest for cell line 2 (H2). Between different cell harvests of the same cell line the maximal difference in frequencies of nuclei in each of the three replication classes was 3%.
Figure 2 Methylation of TIMP1. A. The genes in the region surrounding TIMP1 are shown with the gene name written above the line diagram for the gene (with vertical lines representing exons and small arrows showing transcriptional orientation). The number listed on the line below at the 5' end of the genes indicates the CpG island density. The TIMP1 and ZNF147 genes do not have enough CpG sites to qualify as a CpG island so there is no number listed. The SYN1, and PFC genes are listed in grey because these genes (which are expressed in a tissue-specific manner) were not examined in this study. The presence of repeat elements (SINE, LINE and LTR) is indicated by the vertical lines below the genes. This figure is based on data generated by the UCSC browser hg17 of NCBI Build 35. B. Somatic cell hybrids analysed for TIMP1 activity. In addition to the stable active X chromosomes that express TIMP1 and are unmethylated at the 5' end of the gene (Xa/+/U) there are several categories of inactive X chromosomes. Most inactive X chromosomes previously analysed do not express TIMP1 and are methylated (Xi/-/M). Other inactive X chromosomes express TIMP1 and are unmethylated (Xi/+/U), while an intermediary class of hybrids showed both DNA methylation and lower expression levels (Xi/+/M). Subcloning of these latter cells showed that they were unstable, giving rise to additional methylated expressing clones as well as methylated silent clones and expressing unmethylated clones. The arrows show approximate proportion of cells of each class derived from subcloning. Subclones further characterized are listed below, as are clones derived by 5-azacytidine treatment and their subclones (see Tables 1 and 2). C. DNA methylation of clones from four of the demethylated clones listed in Table 1. DNA from each clone was digested with EcoRI alone (U), EcoRI plus HpaII (II) or EcoRI plus HhaI (I). Primers for the 5' end of TIMP1 and ARAF1 that flank HpaII or HhaI methylation-sensitive restriction enzyme sites were used to assess methylation, while amplification of MIC2, which is unmethylated on both active and inactive X chromosomes, served as a control for complete digestion with the methylation-sensitive enzymes. D. Comparison of expression levels in subclones of two 'sibling' subclones of t11-az-10 differing in methylation states. The t11-az-10-7 and t11-az-10-10 clones are striped, with their subclones shown to their right. Methylation (dark fill) was observed for t11-az-10-7 and its subclones while t11-az-10-10 and its subclones were unmethylated (unfilled). All subclones continued to express both TIMP1 and ARAF1 as assayed by RT-PCR. Despite the relative stability of the methylated TIMP1+ culture, the TIMP1 expression level was significantly lower in the methylated cultures (p < 0.01). E. Methylation analysis by bisulphite treatment. The 5' end of the TIMP1 gene was sequenced after bisulfite conversion, which changes unmethylated Cs to Us but leaves methylated Cs unchanged. Therefore, the presence of a C indicates that the CpG was methylated. The following CpG sites were analyzed: HhaI sites (stars) at -3 and +31; and three other sites not analyzed by methylation-sensitive enzymes (circles) at +11, +17, +20. The HpaII sites (triangles) at +61 and +81 were used in methylation-sensitive assays but were not reliably analysed by bisulphite sequencing. The open circles indicate unmethylated CpGs whereas the filled circles represent methylated CpGs. The shaded circles designate that both converted and unconverted bases were seen after sequencing, indicating that both methylated and unmethylated CpGs were present. Cell lines are listed, the male cells were GM7057 and the female cells were GM7059. t11-az-9-3 is a TIMP1- subclone of t11-az-9.
DNA methylation of TIMP1 in somatic cell hybrids
Analysis of alterations to the inactive X chromosome is complicated by the presence of the active X chromosome in female cells, so we have analysed features of inactive X chromosomes isolated in mouse/human somatic cell hybrids. We have previously described the characterization of DNA methylation status and expression levels in a number of these hybrids [14] and Figure 2B presents an outline of the hybrids analysed in this study. In addition to hybrids retaining the active X chromosome (t60-12 (t60) and AHA11aB1 (AHA)), in which TIMP1 was unmethylated and expressed, three classes of inactive X-containing hybrids were previously described. Those that were methylated and did not have TIMP1 expression (Xi/-/M – t11-4Aaz-5 (t11) and t48-1a-1Daz4a (t48)) or expressed TIMP1 and were unmethylated (Xi/+/U – t75-2maz34-1a (t75)) were stable, with no gain or loss of expression. In contrast, the two hybrids that expressed TIMP1 but were methylated at the TIMP1 promoter (Xi/+/M – t86-B1maz1b-3a (t86) and t81-az1D (t81)) demonstrated instability as subclones could be either methylated and expressing (Xi/+/M) or methylated and silent (Xi/-/M) as well as occasionally unmethylated and expressing (Xi/+/U). Consistently the Xi/-/M and Xi/+/U subclones tended to be stable while Xi/+/M subclones were unstable. As the methylated-expressing clones (Xi/+/M) are unstable and give rise to a mixed population of methylated silent and unmethylated expressing clones, these clones do not indicate if loss of methylation follows, or predisposes to, expression.
To further test the role of methylation in regulating expression in the TIMP1 region, an inactive X-containing hybrid that did not express TIMP1 (t11) was treated with 5-azacytidine, which is known to induce reactivation of X-linked genes in somatic cell hybrids. As there is not a selectable marker for TIMP1 reactivation, clones were initially selected for HPRT reactivation to ensure that the clones had been demethylated and potentially increase the frequency of reactivation, as co-reactivation of different X-linked genes has been observed after 5-azacytidine treatment [25]. 15 HPRT+ clones were selected in HAT media, and examined by RT-PCR for TIMP1 as well as flanking gene expression (Table 1). After three passages, six clones showed expression of TIMP1. A number of flanking genes were also observed to reactivate with 5-azacytidine treatment; with one clone expressing ARAF1, five expressing ELK1, three expressing ZNF41 and eight expressing ZNF157. The single ARAF1-expressing clone, and all five ELK1-expressing clones, were observed in TIMP1+ clones, and all three ZNF41 positive clones were ZNF157 positive. The TIMP1 and ZNF157 genes lack a CpG island (see Figure 2A), and TIMP1 and ZNF157 showed the highest (6/15 and 8/15) reactivation frequencies after 5-azacytidine treatment. However, ARAF1 showed the lowest reactivation frequency (1/15) despite having a smaller CpG island than either ELK1 or ZNF41, so reactivation frequency was not simply a reflection of CpG density.
Table 1 Expression of TIMP1 and surrounding genes following 5-azacytidine induced reactivation of HPRT
Clone * TIMP1 ARAF1 ELK1 ZNF41 ZNF157
t11-az-4 - - - - +
t11-az-5 - - - - +
t11-az-6 + - + - -
t11-az-7 + - + - -
t11-az-8 + - + - -
t11-az-9 + - + - +
t11-az-10 + + - + +
t11-az-11 - - - + +
t11-az-14 - - - - +
t11-az-16 - - - - +
t11-az-17 - - - - -
t11-az-18 - - - + +
t11-az-19 + - + - -
t11-az-20 - - - - -
t11-az-21 - - - - -
*The clones in bold were ones that were further subcloned to assess stability (see Table 2).
Despite 5-azacytidine treatment and subsequent expression in some clones, DNA methylation continued to be observed at 5' end of ARAF1 and TIMP1 in these clones, as shown in Figure 2C. Given our previous results demonstrating instability of expression in the presence of methylation, four clones were subcloned after 7 weeks of culture. Analysis of expression and DNA methylation for TIMP1 (Table 2) showed that, as expected, all subclones of a non-expressing clone (t11-az-4) remained silent. For both t11-az-8 and 9, the majority of clones lacked TIMP1 expression. Only two weakly positive clones were identified for each line, and even these subclones lost expression by 12 weeks in culture. Three subclasses of hybrids were observed for the t11-az-10 subclones, reminiscent of the situation seen in hybrids from females who spontaneously express TIMP1 from the inactive X. t11-az-10 was the only reactivated clone that expressed ARAF1, and nine of the 14 TIMP1-expressing subclones expressed ARAF1. The single subclone that was unmethylated at TIMP1 (t11-az-10-10) expressed TIMP1, and also expressed ARAF1, which was also unmethylated. t11-az-10-7 was methylated yet expressing for both TIMP1 and ARAF1. These clones (t11-az-10-10 and t11-az10-7) were further subcloned. The TIMP1 expression level of the subclones was determined by RNase protection. The t11-az-10-7 subclones retained DNA methylation, and expressed a lower level of TIMP1 than the unmethylated t11-az-10-10 subclones (Figure 2D).
Table 2 Stability of methylation and expression of TIMP1 in subclones of four 5-azacytidine induced HPRT reactivants
Subclones t11-az-4 (-/M) t11-az-8 (+/M)a t11-az-9 (+/M)a t11-az-10 (+/M)
TIMP-, methylated (-/M) 10 10 15 3
TIMP+, Methylated (+/M) 0 2a 2a 13
TIMP+, Unmethylated (+/U) 0 0 0 1
aExpression in these branches was initially weak and became silent after time in culture. ELK1 expression was initially retained in one of the t11-az-9 TIMP1+ clones but was not analyzed after TIMP1 expression was lost. The expression of ZNF41/157 was not examined in these subclones.
DNA methylation analysis was routinely based on methylation-sensitive restriction enzyme digestion followed by PCR that examined 4 sites, 3 of which were over 50 bp downstream of the transcription start site (Figure 2E). To examine methylation of additional CpG dinucleotides closer to the promoter, methylation analysis by bisulfite modification followed by sequencing of PCR products was performed. Three additional CpGs near the TIMP1 transcription start yielded the same methylation pattern as had been observed at the restriction enzyme sites. Direct sequencing of the PCR product was performed to allow the identification of heterogeneous populations. This approach is complementary to the methylation-sensitive restriction enzyme approach, which gave a positive signal (amplification) for low levels of methylated DNA. The high background corresponding to the unconverted (methylated) Cs, presumably due to their under-representation in the sequence, prohibits identification of low levels of methylated bases by direct sequencing. However, no background of the converted base (A) was present and thus this technique is sensitive to the presence of low levels of unmethylated DNA. If the peak corresponding to the converted base (A) was demonstrably higher than the unconverted (G), the site was called unmethylated (white circles in Figure 2E). Completely methylated sites (as seen for the inactive X hybrid) are indicated by filled in black circles. Sequencing also yielded a mix of unmethylated (converted) and methylated (unconverted) sites. This was observed in a female cell line, where the methylated (unconverted – G) peak height was greater than or equal to the unmethylated (converted – A) peak height. Such a mix, demarcated by the grey circles, was also observed for the t11-az-10-7 cells, although in this case the unmethylated peak was consistently lower than the methylated peak.
DNase sensitivity of TIMP1 in somatic cell hybrids
Nuclease sensitivity of the TIMP1 promoter, the promoter of the nearby ARAF1 gene, as well as the TIMP1 gene body and the anonymous DNA marker DXS8037 was assessed in a series of these clones (Figure 3). The promoters of TIMP1 and ARAF1 were sensitive to nuclease on the active, but not the inactive X chromosome, while the gene body or intergenic regions were generally resistant on both active and inactive X chromosomes. The TIMP1 promoter was also sensitive to digestion in the unmethylated expressing (Xi/+/U) TIMP1 clones, while it was insensitive in both the TIMP1-expressing, and silent methylated clones (Xi/+/M and Xi/-/M). A similar effect was observed for the demethylated clones, with both TIMP1 and ARAF1, being expressed in both the t11-az-10-10 and t11-az-10-7 clones, but only sensitive to nuclease digestion in the t11-az-10-10 clone, where they were unmethylated.
Figure 3 Nuclease sensitivity of TIMP1. Nuclei from various cell lines (as indicated on the left) were treated with increasing amounts of DNase I (0, 0.1, 0.25, 0.5, 1.0 units) as represented by the increasing breadth of the triangle. The DXS8037 primers flank a non-coding region and were used to check that equal digestion and PCR amplification occurred across all cell lines. The cell line designations indicate X activity, TIMP1 expression status, and methylation status as shown in Figure 2B.
Histone modifications
A growing number of specific modifications to the histone tails have been associated with both constitutive and facultative heterochromatin. To assess the potential role of such modifications in the escape from inactivation of TIMP1, we analysed the hybrids by ChIP with antibodies to acetylated histone H3 (antibody recognizes acetylated K9 and 14), acetylated histone H4 (antibody recognizes acetylated K 5, 8, 12, and 16) and methylated histone H3 (antibody to di-methyl K9, although the antibody will cross-react with tri-methyl K27 see: ). The promoter regions of TIMP1, ARAF1, ELK1 and the XIST gene (the latter is expressed solely from the inactive X chromosome) were analysed (shown in Figure 4). Additional analyses of the promoters of the PGK1 gene that is expressed only from the active X chromosome, and the ZFX gene that is expressed from both active and inactive X chromosomes, gave the anticipated results for active and inactive X chromosomes (data not shown). Antibody to acetylated H3 immunoprecipitated the TIMP1, ARAF1 and ELK1 promoters for the active, but not inactive X-containing hybrids, while the reverse was observed for the XIST promoter. ELK1 was not expressed in the inactive X hybrids examined, and was not immunoprecipitated. However for TIMP1 and ARAF1 association with acetylated histone H3 was observed for all expressing clones whether on the active or inactive X chromosome. In addition, immunoprecipitation of the TIMP1 promoter was observed for Xi/-/M clones (t86-6P/t86-1U and t81-4). Immunoprecipitation was completely concordant with expression for the acetylated histone H4 antibody at all of the genes examined. Methylation of H3 at K9 corresponds to silent chromatin, and thus, as anticipated, results were generally opposite to those seen for acetylation. Immunoprecipitation was observed when the gene was silent – i.e. XIST on the active X and TIMP1, ARAF1 and ELK1 on the inactive X. However, in all cases where there was expression in the presence of ongoing promoter methylation (for both ARAF1 and TIMP1) the promoter was immunoprecipitated by the methylated H3 antibody. Thus we detect a distinctive pattern of histone acetylation and methylation that does not correspond simply with the expression or DNA methylation status of the TIMP1 gene.
Figure 4 ChIP analysis of TIMP1 expressing or non-expressing hybrid clones. PCR products were amplified from DNA at various X-linked gene promoters after ChIP using antibodies to the modifications listed below the panels for the cell lines listed across the top of each panel (see Figure 2B for derivation of lines). t11-az-10-7a is a subclone of t11-az-10-7 (see Figure 2D). The DNA template for 'no Ab' lanes was prepared following the ChIP procedure without an antibody. A. PCR amplification products after ChIP with acetylated histone H3. Similar analysis to A, using antibody to acetylated histone H4 (panel B), or methylated histone H3 (C).
Discussion
The facultative heterochromatin of the inactive X chromosome is a fascinating system to study the epigenetic modifications associated with silent chromatin as both an active and inactive version of most X-linked genes exist in female cells. However, not all genes on the 'inactive' X chromosome are subject to silencing [26], and these genes that escape inactivation must somehow maintain an active state on an otherwise silent chromosome. TIMP1 is subject to inactivation in many females, and we now show that when subject to inactivation the anticipated features of inactive chromatin are assembled – promoter DNA hypermethylation, nuclease insensitivity and histone methylation and hypoacetylation. In other females, however, we have previously demonstrated that TIMP1 continues to be expressed from the inactive X chromosome, and in this study we have exploited such chromosomes to analyze the epigenetic features that are associated with expression from the inactive X. By comparing expression of TIMP1 that occurs naturally with that induced by the demethylating agent 5-azacytidine we are also able to propose which feature may predispose to expression of TIMP1 in otherwise silent chromatin. We observed that expression of TIMP1 was associated with an active chromatin structure, despite the presence of the gene on the inactive X chromosome, except in three situations.
First, in female cell lines with or without TIMP1 expression from the inactive X there was a very similar extent of replication asynchrony, suggesting that the expression of TIMP1 was ocurring from a late-replicating region of the X chromosome. This is not surprising, as the genes flanking TIMP1, which are not variable in their inactivation, are located within 55 kb, so it is likely that at least one of ARAF1 or ELK1 shares a replication origin with TIMP1. We assessed replication timing by FISH, and detected a frequency of replication asynchrony for the HPRT locus in close agreement with the frequencies previously reported for this gene [27]. It has been suggested that the 'double dot' versus 'single dot' pattern may reflect chromatid association in addition or instead of replication asynchrony [17]. Regardless of the underlying cause, the results were highly reproducible and consistently showed less asynchrony for the TIMP1 locus relative to the HPRT locus, although still in the range that is generally considered asynchronous [28]. This reduced asynchrony at TIMP1 could reflect, or contribute to, weaker epigenetic silencing being established in this region of the X chromosome. There was a slight, but not statistically significant, trend for the female showing expression of TIMP1 from the inactive X chromosome to show even less replication asynchrony for TIMP1, however it seems unlikely that such a small difference could contribute substantially to TIMP1 expression from the inactive X, and thus other factors must be involved in permitting the expression of TIMP1 from the inactive X.
The other two situations in which expression and active chromatin features were not concordant were found in the somatic cell hybrids, and these results are summarized in Figure 5. As shown in the grey box, DNA methylation and additional features of silent chromatin were detected along with expression in the Xi/+/M (t86, t81) cells. We attribute this to heterogeneity in the population of cells, as subcloning yielded both methylated/silent and unmethylated/expressing clones. In addition to DNA methylation, these clones were observed to be insensitive to DNase and have methylated H3K9 residues near the TIMP1 promoter. As the assays for these features relied on PCR, a positive signal could be obtained from a subpopulation of cells with a silent chromatin structure, while another population of cells could be positive for expression and active chromatin modifications (histone acetylation). We also suggest that heterogeneity accounts for expression in the presence of DNA methylation in several of the demethylated clones, including the t11-az-10-7 clone. However, unlike the mixed population of subclones obtained with all other Xi/+/M cells, eight of eight subclones of t11-az-10-7 were methylated and expressing, and six of these subclones examined by RPA showed a consistently reduced level of expression relative to their unmethylated counterparts. While it was surprising that subcloning did not isolate distinct subpopulations, the presence of an unmethylated subpopulation was detectable by bisulphite sequencing. Thus heterogeneity again seems the most likely explanation for DNA methylation, DNase insensitivity and histone H3K9 methylation in these expressing cells. However, without single cell assays it is not possible to rule out that there are methylated cells that express TIMP1 at reduced levels and show additional features of silent chromatin. These two classes of clones were the only exceptions to concordance between nuclease sensitivity and gene expression at the promoters of the TIMP1 and ARAF1 genes; and previous studies of the nucleosomal organization of the HPRT1 gene promoter have shown that methylation does not directly affect the differential positioning of nucleosomes on active and inactive X-linked promoters [19]. Thus we believe that the methylated/expressing/nuclease insensitive clones reflect the presence of a subpopulation of silent cells. This heterogeneity demonstrates the unstable nature of silencing for TIMP1 in these cells.
Figure 5 Summary of chromatin features observed in somatic cell hybrid clones for the TIMP1 gene. Most subclones and demethylated clones follow the patterns seen for the active and inactive X hybrids that are outlined in bold. A positive (+) designates the presence of the feature, while a negative (-) depicts the absence of the feature, while ND means that the feature has not been examined in that class of clones. The assays used are PCR-based and would detect a small population of cells. DNase sensitivity is listed as the inverse – DNase resistance – as it is the presence of undigested (resistant) DNA that will yield a 'positive' PCR signal. In each category of clones not all clones have been examined for all features. For the clones shaded in grey the results are an amalgamation of results anticipated from an active and inactive X, and we suggest these clones represent a mixed population. This suggestion is generally supported by sub-cloning experiments (see text for discussion). The ongoing H3 acetylation of the Xi/-/M hybrids (highlited in darker grey) cannot be attributed to heterogeneous cell populations, and since it is not seen for demethylated hybrids that have lost TIMP1 expression (t11-az/-/M) we suggest that this modification reflects a predisposing feature of inactive X chromosomes that express TIMP1.
The third exception is the most interesting, and is highlighted in Figure 5 with a dark grey fill. Acetylation of histones is generally seen for active genes, and acetylated histone H4 showed complete concordance with expression for all genes. However, the Xi/-/M clones (t86-6P and 1 U as well as t81-4) showed ongoing acetylation at H3 despite having lost TIMP1 expression from the inactive X chromosome. The Xi/-/M subclones were derived from two different unstable, naturally expressing inactive X chromosomes (in t86 and t81), however expression of TIMP1 in these hybrids was now stably silenced, and thus this result was not a reflection of a mixed population of cells. Furthermore, acetylation cannot simply reflect a failure of this region to reset chromatin structure, as the demethylated t11-az-9-10 hybrid which had lost expression had also lost acetylation. Methylation of H3 K9 was observed in these silent clones, so the Xi/-/M cells, despite being a homogeneous population, appear to show both methylation and acetylation of histone H3. This may reflect the lack of specificity of the antibodies used for these experiments, as the acetylated H3 antibody recognizes acetylation at residues 9 and 14, while the antibody to methylated K9 can cross-react with methylation of K27. Thus immunoprecipitation by both antibodies may reflect modification of specific lysines. It is also possible that only a particular set of nucleosomes show the acetylation mark, and as the sonicated fragments immunoprecipitated in the assay averaged ~600 bp, modifications on several different nucleosomes flanking the primers used could result in immunoprecipitation. Regardless of the specific site of modification, acetylation of histone H3 was the one feature that was consistently associated with X chromosomes that were naturally predisposed to expression from the inactive X chromosome, regardless of expression status. Thus we propose that histone H3 acetylation differs at the TIMP1 genes in females, predisposing some females to expression from the inactive X. Unfortunately this hypothesis is difficult to test, as females generally show acetylation at TIMP1, due to the presence of the active X chromosome. It would be necessary to study clonal populations of cells from a female with a polymorphism close enough to the promoter to be analysed by ChIP, and currently no such polymorphisms are known. Our previous work did not show an association between expression of TIMP1 and a downstream polymorphism in exon 5 [12].
A different predisposing epigenetic mark for expression from the inactive X has previously been reported in mice. Methylation of H3 at K4 was restricted to the promoter in undifferentiated embryonic stem cells for genes that would subsequently be expressed monoallelically (i.e. genes subject to X inactivation or imprinting), while autosomal genes or genes that escape inactivation had H3 K4 methylation in the gene body as well as the promoter [11]. Such methylation was observed in the body of the Smcx gene which is initially silenced, and then reactivates, reminiscent of the H3 acetylation mark and instability of silencing that we observe for TIMP1. While many genes escape inactivation in humans, TIMP1 provides us with a unique opportunity to examine the gene in its expressed and silent state in hybrid somatic cells. Thus, it will be interesting to study additional chromatin modifications for TIMP1, including H3 K4 methylation, as well as additional modifications recently associated with the inactive X chromosome such as H4 K20 methylation [29] or H2A ubiquitinylation [30,31]. Even more interesting would be to determine if the changes seen are observed for additional X-linked genes that are variable in their inactivation status, if suitable model systems could be developed.
Previous examinations of genes that escape inactivation have shown an active chromatin structure (reviewed in [7]), and recent results further demonstrate that epigenetic modifications seem to be heterogeneous in their distribution along the X [32]. These 'flavours' of inactive X chromatin may correspond to clusters of genes that escape inactivation. While genes that escape inactivation tend to be clustered in blocks and enriched on the short arm of the X chromosome, genes such as TIMP1 that are variable in their inactivation are more evenly distributed along the X chromosome ([5], reviewed in [6]), and thus may reflect different mechanisms leading to expression from the inactive X. The CTCF boundary factor has recently been shown to be found between the domain of escape and that of inactivation [33]. It is not known whether such boundaries flank the variably inactivated genes. The five clones that showed reactivation of ELK1 were all TIMP1+, suggesting a co-ordinate reactivation between the two genes. Only one ARAF1 reactivant was identified, and this clone was also TIMP1+. However, this co-ordinate reactivation of TIMP1 with flanking gene(s) is different from the TIMP1-specific expression seen for the inactive X naturally. The ZNF157/41 genes did not seem to show co-ordinate reactivation with the TIMP1 region, however all three ZNF41 positive clones were ZNF157 positive, which suggests they may be in a separately controlled domain.
While association with a CpG island does not differ between genes subject to or escaping from inactivation, genes heterogenous in their inactivation are more likely to lack an island [5], consistent with the TIMP1 promoter being the least CpG dense of the region. Although TIMP1 lacks a CpG island, our analysis suggests that methylation of the few CpGs present near the promoter generally correlates well with silencing. Demethylation resulted in at least transient expression of all genes examined, confirming the importance of DNA methylation in stable maintenance of X chromosome silencing. Examination of the ELK1, TIMP1 and ARAF1 promoters after demethylation showed the presence of DNA methylation, even when the clones were expressing the genes. The majority of these clones subsequently resilenced these genes, as has been observed for other genes following demethylation [34], perhaps reflecting only partial initial demethylation or the spread of silencing from adjacent silent regions that retained DNA methylation or other epigenetic marks of silencing.
It has been proposed that the evolutionarily recent addition of the X short arm may predispose genes located there to escape from silencing [35]. Interestingly, TIMP1 is very close to the evolutionary breakpoint between the region added to the X after marsupial divergence. ARAF1 is found on the marsupial X while TIMP1 and its surrounding gene SYN1 are autosomal in marsupials, suggesting that they are part of the eutherian addition to the human X [36]. Since genes closely flanking TIMP1 are normally subject to X inactivation, even when TIMP1 is expressed, evolutionary history alone cannot explain the escape from inactivation, although it may result in a different genomic context that contributes to expression from the inactive X. However, inspection of genomic features in the region surrounding TIMP1 does not show a substantial difference in the frequency of repetitive elements near TIMP1 relative to the flanking ARAF1 and ELK1 genes (Figure 2A).
Conclusion
Several factors, including reduced replication asynchrony, lower CpG density, and more recent evolutionary addition to the X, may contribute to less stringently controlled inactivation for TIMP1. However, we propose that there is a difference between females for a feature unique to TIMP1 that predisposes some females to expression of the gene. This mark appears to be at least reflected in a difference in acetylation of histone H3 on the inactive X in females predisposed to expression of TIMP1 from the inactive X chromosome. While histone acetylation appears to be a predisposing mark for expression of TIMP1, there may be a different hierarchy of epigenetic modifications permitting expression of other genes from the inactive X. Elucidation of these mechanisms is important not only as a model for epigenetic gene regulation but because genes that escape inactivation contribute to the phenotype of X chromosome anueploidies, and may also result in differential male/female expression levels and disease susceptibilities.
Methods
Cell culture
Lymphoblast cell lines were grown in RPMI 1640 media (Stem Cell Technologies) supplemented with 15% fetal calf serum (Cansera), L-glutamine (Invitrogen) and penicillin/streptomycin (Invitrogen). Cells were harvested 12–26 hours after addition of fresh media by centrifugation. The human/rodent somatic cell hybrids were grown in alpha minimal essential media (Invitrogen) supplemented with 7.5% fetal calf serum, L-glutamine, penicillin/streptomycin, and non-essential amino acids (Invitrogen) to sub-confluence before harvesting with trypsin-EDTA (0.25%). Cell lines have been previously described [14]. To generate single cell clones, the hybrid cultures were plated to a final concentration of 3 to 17 cells/60 mm plate. After 5 to 10 days in culture, well-separated colonies were isolated by trypsinization in cloning cylinders and transferred to new 60 mm plates. To induce demethylation, an inactive X-containing hybrid that had never expressed TIMP1 (t11-4Aaz-5) was treated with 5-azacytidine as previously described [37]. The cells were grown in media supplemented with HAT (Invitrogen) to select for HPRT reactivants and then single cell cloned to test for TIMP1-positive cultures. To remove any confounding effects of HPRT selection, the cells were transferred back to alpha minimal essential media after 2 weeks of selection. The expression of genes in the TIMP1 region was determined by RT-PCR as described previously [12].
Replication timing
Approximately 4 × 106 lymphoblast cells were harvested one day after subculture to ensure that they were actively growing. The cell pellet was resuspended in 8 ml of prewarmed hypotonic (0.75 M KCl) and then incubated at 37C for 10 minutes. 2 to 3 ml of fixative (3:1 methanol:glacial acetic acid) was slowly added before spinning at 200 g for 10 minutes. The cell pellet was washed three more times with 5 ml fixative and then resuspended in 10 ml fixative for storage at -20C for up to a week. The nuclei preparations were dropped onto slides and left at room temperature overnight. The slides were then incubated in 2X SSC at 37C for 30 minutes followed by 2 minute room temperature incubations in each of: 70%, 85%, and 95% ethanol, and then allowed to air dry. To denature, the slides were incubated in fresh 70% formamide/2 × SSC at 74C for 2 minutes followed by an ice cold ethanol series of rinses (70%, 85%, 95% for 2 minutes each) and air drying. Probes for both the TIMP1 locus (lambda phage TIMP-3.9X) and the HPRT locus (Hulambda4x-8, ATCC 57236) were labelled with dUTP-digoxigenin (Roche 1745816) by nick translation (Roche 0976776). Unincorporated nucleotides were removed with a PCR clean-up kit (Qiagen), and 1 ul (approximately 100 ng) of labelled probe was mixed with 10 ul of 70% formamide hybridization buffer and 40 ng of human Cot-1 DNA (Invitrogen 15279011). The probes were denatured at 74C for 10 minutes, pre-annealed at 37C for one hour and then added to the prepared slide to hybridize overnight in a humidified chamber at 37C. The slides were washed in 50% formamide/2 × SSC for 15 minutes at 43C, followed by two washes in 2 × SSC for 4 minutes at 37C and then 3 washes at room temperature for 2 minutes each in 1 × PBD (0.1 M NaH2PO4, 0.1 M Na2HPO4, 0.1% Triton X). To visualize the probe, slides were incubated with 500 ng of anti-DIG (sheep – Roche 1207741) conjugated to fluorescein, for 5 minutes at 37C, followed by three two-minute washes in 1 × PBD. To amplify the signal, FITC-anti-sheep (IgG FI-6000 (Vector Laboratories)) was incubated and washed as above. The slides were then counterstained with DAPI mixed with antifade (VectorLabs). Cells were scored for nuclei with single-single, single-double, or double-double TIMP1 or HPRT signals on a Ziess Axioplan II microscope by a single individual who was blinded as to the cell line being analysed.
Methylation analyses
For methylation-sensitive restriction enzyme analysis, genomic DNA was pre-digested with EcoRI at 37C overnight, followed by incubation with 2 ul of RNase at 37C for 15 minutes. After phenol extraction and ethanol precipitation, the DNA was quantified by spectrophotometry. Pre-digested DNA (2 ug) was then incubated overnight at 37C in a total volume of 20 ul with 20 units of one of the following: mock enzyme (uncut), HpaII, or HhaI. An aliquot of 1 ul (100 ng) was then used as a template in the PCR reactions as described previously [14]. All primers flanked a region of genomic DNA that did not contain EcoRI restriction enzyme sites and contained 1–2 HpaII or HhaI sites. For bisulfite analysis, 500 ng of genomic DNA was first denatured with 3 ul of 3 M NaOH at 37C for 10 minutes. After 15 ul of 20 mM hydroquinone and 255 ul of 3.9 M sodium bisulfite were added and mixed well, reactions were left at 50C for 16 hours to allow the unmethylated cytosines to convert to uracil. DNA was purified using the DNA Wizard Clean-Up Kit (Promega). After amplification with the TIMP-S primers (see Table 3), the 3' reverse oligonucleotide was used as the primer for sequencing of the PCR product.
Table 3 Primers for PCR analyses
Primer pair Sequence Use
TIMP1 5' (promoter) [14] 5'A: CCCTTGGGTTCTGCACTGA*
5'B: CCAAGCTGAGTAGACAGGC Methylation
ChIP
DNase sensitivity
TIMP1 CA (gene body) [12] CA1: GGGTTCCAAGCCTTAGGGGA
CA2: AGGCTGTTCCAGGGAGCCGC DNase sensitivity
TIMP1 S (bisulfite) 5S: GttttTTGGtTTtTGtAtTGATGGT
3S: CCAAaCTaAaTAaACAaaCATCTAaC** Bisulfite sequencing
ARAF1 M1:M4 (promoter) [14] M1: TGCCAAAGCCCTAAGGTCA
M4: CGCTGTCGACGATGGTCT
M3: GTGAGGAAACAAGAAGAGAG Methylation
ChIP
DNase sensitivity
XIST 3':5' (gene body) [39] 3':GAAGTCTCAAGGCTTGAGTTAGAAG
5': TTGGGTCCTCTATCCATCTAGGTAG Methylation
DNase sensitivity
XIST A5:29r (promoter) [37] A5: TTTCTTACTCTCTCGGGGCT
29r: ATCAGCAGGTATCCGATACC ChIP
ELK1 5' (promoter) [14] A: GCACAGCTCTGTAGGGAA
B: AGCTCACCTGTGTGTAGCG Methylation
ChIP
STA A:B (intergenic) A: CACCTGTGTGTCATGTATAC
B: CCAGTATTGGTCTTCCAGTT DNase sensitivity
8037 A:B (intergenic) A: GAGGCAAGACATCCATTCC
B: TGACTTTGAGCGAGCAGGT Reference region
* There is a mismatch in the TIMP 5'A primer, the underlined G should be C.
** The lower-case letters in the primers are the bases modified by the bisulfite reaction. All C nucleotides should have been converted because there were no CpG pairs with possible protective methylation.
RNase Protection quantitation of TIMP1 levels
RNase protection analysis was performed with Ambion's RPA II kit, following the manufacturer's directions. RNA probes were isolated after in vitro transcription with 32P-UTP. After solution hybridization overnight of excess antisense radiolabelled probe to 10 ug of total RNA, any unhybridized probe and sample RNA was removed by RNase digestion. The hybridized product was then separated on a native 5% polyacrylamide gel, visualized by autoradiography, and bands quantified by phosphoimager (BioRad FX). The intensity of the TIMP1 fragment was compared to the intensity of the band detected for MIC2 used to control for the amount of input RNA. All RPA results were normalized to one stock RNA to decrease variability between gels (see [14]).
DNaseI sensitivity
The protocol for nuclease sensitivity was adapted primarily from [38]. The somatic cell hybrids were grown to 75% confluence on 60 mm plates before harvesting with 0.25% trypsin-EDTA. The cell pellets were resuspended in 300 ul ice-cold DNase buffer (0.3 M sucrose, 60 mM KCl, 15 mM NaCl, 5 mM MgCl2, 0.1 mM EGTA, 0.5 mM DTT, and 15 mM Tris-HCL pH7.5). The suspensions were then split into 6 tubes of 50 ul each and another 50 ul of ice-cold DNase buffer with 0.4% Nonidet P40 was added. The tubes were mixed gently and placed on ice for four to five minutes before 100 ul of freshly diluted DNaseI (Invitrogen 18047-019) was added, such that 2.5 U, 1 U, 0.5 U, 0.25 U, 0.1 U, and no enzyme were used. The reaction was then incubated at 25C for 5 minutes, followed by 95C for 15 minutes. The DNA was isolated with a standard phenol/chloroform extraction protocol, quantified with spectrophotometry, and diluted to 60 ng/ul, 12 ng/ul, and 4 ng/ul. 1 ul was used as a PCR template with primers listed in Table 3. The diluted DNAs showed product intensity inversely proportional to concentration, indicating that the PCRs at 25 cycles of amplification were in the linear range of amplification, and the results shown in figure 3 are for the 12 ng/ul template.
Chromatin Immunopreciptation Assay (ChIP Assay)
The hybrid cells were grown in a t25 flask and were harvested after treatment with 5 drops of 0.25% trypsin-EDTA for 2 minutes at room temperature, and then washed in 1 × PBS. The cells were placed in a 1.5 ml tube with one ml of 0.37% formaldehyde in minimal essential media (Invitrogen) and incubated at 37C for ten minutes. From this point, the cells were kept on ice. The cells were washed twice with 1/100 proteinase inhibitor cocktail (Sigma) in 1 × PBS and then centrifuged at 2500 rpm for 4 minutes at 4C. The pellet was resuspended in 200 ul SDS lysis buffer (Upstate) with 1/100 proteinase inhibitor cocktail and placed on ice for 10 minutes. The suspension was drawn up with a 25 gauge needle and then sonicated. Immunoprecipitation was performed (Upstate Biochemicals catalogue number 17-295) with the following antibodies: Anti-acetyl H3 against lysine 9 and 14 (catalogue number 06-599); Anti-acetyl H4 against lysine 5, 8, 12, and 16 (catalogue number 06-866); and Anti-dimethyl-H3 against lysine 9 (catalogue number 07-212).
Abbreviations
H3 – histone H3; H4 – histone H4; K – lysine; ChIP – chromatin immunoprecipitation; X – X chromosome; Y – Y chromosome; Xi – inactive X; Xa – active X; M – methylated; U – unmethylated; + expressing; - not expressing; S – single signal; D – double signal.
Authors' contributions
CA carried out the molecular genetic studies and data analysis and drafted the manuscript. CB conceived of the study, participated in the experimental design and data analysis and wrote the final manuscript. Both authors read and approved the final manuscript.
Acknowledgements
This study was supported by a Canadian Institutes of Health Research operating grant (MOP13690).
==== Refs
Cheung VG Conlin LK Weber TM Arcaro M Jen KY Morley M Spielman RS Natural variation in human gene expression assessed in lymphoblastoid cells Nat Genet 2003 33 422 425 12567189 10.1038/ng1094
Whitney AR Diehn M Popper SJ Alizadeh AA Boldrick JC Relman DA Brown PO Individuality and variation in gene expression patterns in human blood Proc Natl Acad Sci U S A 2003 100 1896 1901 12578971 10.1073/pnas.252784499
Thomson MW McInnes RR Willard HF Genetics in Medicine 1991 5 Philadelphia, W.B. Saunders Co.
Lyon MF Gene action in the X-chromosome of the mouse (Mus musculus L.) Nature 1961 190 372 373 13764598
Carrel L Willard HF X-inactivation profile reveals extensive variability in X-linked gene expression in females Nature 2005 434 400 404 15772666 10.1038/nature03479
Brown CJ Greally JM A stain upon the silence: genes escaping X inactivation Trends Genet 2003 19 432 438 12902161 10.1016/S0168-9525(03)00177-X
Disteche CM Filippova GN Tsuchiya K Escape from X inactivation Cytogenet Genome Res 2002 99 36 43 12900543 10.1159/000071572
Craig IW Harper E Loat CS The genetic basis for sex differences in human behaviour: role of the sex chromosomes Ann Hum Genet 2004 68 269 284 15180708 10.1046/j.1529-8817.2004.00098.x
Tsuchiya KD Greally JM Yi Y Noel KP Truong JP Disteche CM Comparative sequence and X-inactivation analyses of a domain of escape in human Xp11.2 and the conserved segment in mouse Genome Res 2004 14 1275 1284 15197169 10.1101/gr.2575904
Lingenfelter PA Adler DA Poslinski D Thomas S Elliot RW Chapman VM Disteche CM Escape from X inactivation of Smcx is preceded by silencing during mouse development Nat Genet 1998 18 212 213 9500539 10.1038/ng0398-212
Rougeulle C Navarro P Avner P Promoter-restricted H3 Lys 4 di-methylation is an epigenetic mark for monoallelic expression Hum Mol Genet 2003 12 3343 3348 14570711 10.1093/hmg/ddg351
Anderson C Brown CJ Polymorphic X-chromosome inactivation of the human TIMP1 gene Am J Hum Genet 1999 65 699 708 10441576 10.1086/302556
Carrel L Willard HF Heterogeneous gene expression from the inactive X chromosome: an X-linked gene that escapes X inactivation in some human cell lines but is inactivated in others. Proc Natl Acad Sci USA 1999 96 7364 7369 10377420 10.1073/pnas.96.13.7364
Anderson CL Brown CJ Variability of X chromosome inactivation: effect on levels of TIMP1 RNA and role of DNA methylation Hum Genet 2002 110 271 278 11935340 10.1007/s00439-002-0676-8
Plath K Mlynarczyk-Evans S Nusinow DA Panning B Xist RNA and the mechanism of X chromosome inactivation Annu Rev Genet 2002 36 233 278 12429693 10.1146/annurev.genet.36.042902.092433
Willard HF Latt SA Analysis of deoxyribonucleic acid replication in human X chromosomes by fluorescence microscopy Am J Hum Genet 1976 28 213 227 1266850
Azuara V Brown KE Williams RR Webb N Dillon N Festenstein R Buckle V Merkenschlager M Fisher AG Heritable gene silencing in lymphocytes delays chromatid resolution without affecting the timing of DNA replication Nat Cell Biol 2003 5 668 674 12833066 10.1038/ncb1006
Chow JC Brown CJ Forming facultative heterochromatin: silencing of an X chromosome in mammalian females Cell Mol Life Sci 2003 60 2586 2603 14685685 10.1007/s00018-003-3121-9
Chen C Yang TP Nucleosomes are translationally positioned on the active allele and rotationally positioned on the inactive allele of the HPRT promoter Mol Cell Biol 2001 21 7682 7695 11604504 10.1128/MCB.21.22.7682-7695.2001
Jeppesen P Turner B The inactive X chromosome in female mammals is distinguished by a lack of histone H4 acetylation, a cytogenetic marker for gene expression Cell 1993 74 281 289 8343956 10.1016/0092-8674(93)90419-Q
Heard E Rougelle C Arnaud D Avner P Allis CD Spector DL Methylation of histone H3 at Lys-9 is an early mark on the X chromosome during X inactivation Cell 2001 107 727 738 11747809 10.1016/S0092-8674(01)00598-0
Plath K Fang J Mlynarczyk-Evans SK Cao R Worringer KA Wang H de la Cruz CC Otte AP Panning B Zhang Y Role of histone H3 lysine 27 methylation in X inactivation Science 2003 300 131 135 12649488 10.1126/science.1084274
Gilbert SL Sharp PA Promoter-specific hypoacetylation of X-inactivated genes Proc Natl Acad Sci U S A 1999 96 13825 13830 10570157 10.1073/pnas.96.24.13825
Rougeulle C Chaumeil J Sarma K Allis CD Reinberg D Avner P Heard E Differential histone H3 Lys-9 and Lys-27 methylation profiles on the X chromosome Mol Cell Biol 2004 24 5475 5484 15169908 10.1128/MCB.24.12.5475-5484.2004
Hors-Cayla MC Heuertz S Frezal J Coreactivation of four inactive X genes in a hamster x human hybrid and persistence of late replication of reactivated X chromosome Somat Cell Genet 1983 9 645 657 6197760 10.1007/BF01539470
Carrel L Cottle AA Goglin KC Willard HF A first-generation X-inactivation profile of the human X chromosome Proc Natl Acad Sci U S A 1999 96 14440 14444 10588724 10.1073/pnas.96.25.14440
Boggs BA Chinault AC Analysis of replication timing properties of human X-chromosomal loci by fluorescence in situ hybridization Proc Natl Acad Sci 1994 91 6083 6087 8016119
Gribnau J Luikenhuis S Hochedlinger K Monkhorst K Jaenisch R X chromosome choice occurs independently of asynchronous replication timing J Cell Biol 2005 168 365 373 15668296 10.1083/jcb.200405117
Kohlmaier A Savarese F Lachner M Martens J Jenuwein T Wutz A A chromosomal memory triggered by Xist regulates histone methylation in X inactivation PLoS Biol 2004 2 E171 15252442 10.1371/journal.pbio.0020171
de Napoles M Mermoud JE Wakao R Tang YA Endoh M Appanah R Nesterova TB Silva J Otte AP Vidal M Koseki H Brockdorff N Polycomb Group Proteins Ring1A/B Link Ubiquitylation of Histone H2A to Heritable Gene Silencing and X Inactivation Dev Cell 2004 7 663 676 15525528 10.1016/j.devcel.2004.10.005
Fang J Chen T Chadwick B Li E Zhang Y Ring1b-mediated H2A ubiquitination associates with inactive X chromosomes and is involved in Initiation of X-inactivation J Biol Chem 2004 279 52812 5 15509584 10.1074/jbc.C400493200
Chadwick BP Willard HF Multiple spatially distinct types of facultative heterochromatin on the human inactive X chromosome Proc Natl Acad Sci U S A 2004 101 17450 17455 15574503 10.1073/pnas.0408021101
Filippova GN Cheng MK Moore JM Truong JP Hu YJ Nguyen DK Tsuchiya KD Disteche CM Boundaries between chromosomal domains of X inactivation and escape bind CTCF and lack CpG methylation during early development Dev Cell 2005 8 31 42 15669143 10.1016/j.devcel.2004.10.018
Hansen RS Gartler SM 5-Azacytidine-induced reactivation of the human X chromosome-linked PGK1 gene is associated with a large region of cytosine demethylation in the 5' CpG island Proc Natl Acad Sci U S A 1990 87 4174 4178 1693431
Lahn BT Page DC Four evolutionary strata on the human X chromosome Science 1999 286 964 967 10542153 10.1126/science.286.5441.964
Wilcox SA Watson JM Spencer JA Graves JA Comparative mapping identifies the fusion point of an ancient mammalian X-autosomal rearrangement Genomics 1996 35 66 70 8661105 10.1006/geno.1996.0323
Tinker AV Brown CJ Induction of XIST expression from the human active X chromosome in mouse/human somatic cell hybrids by DNA demethylation Nucl Acids Res 1998 26 2935 2940 9611238 10.1093/nar/26.12.2935
Gregory RI Feil R Analysis of chromatin in limited numbers of cells: a PCR-SSCP based assay of allele-specific nuclease sensitivity Nucleic Acids Res 1999 27 e32 10536165 10.1093/nar/27.22.e32
Brown CJ Ballabio A Rupert JL Lafreniere RG Grompe M Tonlorenzi R Willard HF A gene from the region of the human X inactivation centre is expressed exclusively from the inactive X chromosome Nature 1991 349 38 44 1985261 10.1038/349038a0
|
16194278
|
PMC1262707
|
CC BY
|
2021-01-04 16:30:39
|
no
|
BMC Genet. 2005 Sep 29; 6:48
|
utf-8
|
BMC Genet
| 2,005 |
10.1186/1471-2156-6-48
|
oa_comm
|
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-1201615939410.1186/1471-2164-6-120Research ArticleDuplication and positive selection among hominin-specific PRAME genes Birtle Zoë [email protected] Leo [email protected] Chris [email protected] MRC Functional Genetics Unit, University of Oxford, Department of Human Anatomy and Genetics, South Parks Road, Oxford OX1 3QX UK2005 13 9 2005 6 120 120 24 3 2005 13 9 2005 Copyright © 2005 Birtle et al; licensee BioMed Central Ltd.2005Birtle et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The physiological and phenotypic differences between human and chimpanzee are largely specified by our genomic differences. We have been particularly interested in recent duplications in the human genome as examples of relatively large-scale changes to our genome. We performed an in-depth evolutionary analysis of a region of chromosome 1, which is copy number polymorphic among humans, and that contains at least 32 PRAME (Preferentially expressed antigen of melanoma) genes and pseudogenes. PRAME-like genes are expressed in the testis and in a large number of tumours, and are thought to possess roles in spermatogenesis and oogenesis.
Results
Using nucleotide substitution rate estimates for exons and introns, we show that two large segmental duplications, of six and seven human PRAME genes respectively, occurred in the last 3 million years. These duplicated genes are thus hominin-specific, having arisen in our genome since the divergence from chimpanzee. This cluster of PRAME genes appears to have arisen initially from a translocation approximately 95–85 million years ago. We identified multiple sites within human or mouse PRAME sequences which exhibit strong evidence of positive selection. These form a pronounced cluster on one face of the predicted PRAME protein structure.
Conclusion
We predict that PRAME genes evolved adaptively due to strong competition between rapidly-dividing cells during spermatogenesis and oogenesis. We suggest that as PRAME gene copy number is polymorphic among individuals, positive selection of PRAME alleles may still prevail within the human population.
==== Body
Background
Humans and chimpanzees shared a common ancestor approximately 6–7 million years ago (MYA) [1]. Distinguishing characteristics, such as those relating to cognitive abilities, language, habitual upright gait, dentition, and susceptibility to malaria, are all assumed to be associated with genetic differences between these two species [2]. However, these phenotypic differences have been associated with specific human-chimpanzee sequence differences in fewer than a handful of cases. In humans, two coding changes in FOXP2 have been proposed to contribute to language acquisition [3], disruption in the MYH16 myosin heavy chain gene is proposed to have led to a reduction in masticatory muscles [4], and the pseudogenisation of a type I hair keratin has been associated with modifications in our hair keratin phenotype [5]
It is also unclear at which developmental stages, and in which tissues, such human-specific adaptations are first manifested. For example, the abnormal spindle-like microcephaly-associated (ASPM) gene has roles in mitosis, meiosis and cytokinesis, and is broadly expressed in many tissues. Yet it is a major determinant of cerebral cortical size [6] and has evolved adaptively in recent hominin evolution (reviewed in Ponting & Jackson (2005))[7]. As the signatures of recent adaptation are identified in the human genome it will be important to associate these DNA changes with molecular, cellular and physiological innovation.
With the sequencing of the human and chimpanzee genomes comes the possibility of discerning nucleotide changes that have been acquired adaptively and thus might be associated with physiological innovation [2]. Two factors, however, often confound such studies. First, the scarcity of substitutions (~1% [8,9]) between human and chimpanzee orthologous coding sequence provides insufficient statistical power to distinguish adaptive from neutral substitutions. Second, the chimpanzee genome has been sequenced only to low coverage (4-fold statistical coverage for panTro1; see [10]). As a result, the chimpanzee genome sequence contains many gaps, sequence inaccuracies and assembly artefacts. Such problems are exacerbated in regions containing identical or almost identical tandem segments which pose particular problems for both sequencing and assembly. Juxtaposed and virtually identical sequences are frequently represented either by only single versions in genome assemblies, or are absent altogether, thereby giving rise to gaps in the assembly.
By contrast, the human genome assembly is virtually complete and is accurate to approximately one error every 105 bases [11]. The human sequence's high statistical coverage gives rise to an assembly that is a mosaic of contributions from multiple individuals and thus does not represent any single genome. This mosaicism is less important for single nucleotide polymorphisms (SNPs) than it is for larger-scale polymorphisms, such as copy number polymorphisms (CNPs). This is because most SNPs are selectively neutral whereas the evidence suggests that this appears not to be the case for CNPs [12].
Identifying sequence changes that distinguish human and chimpanzee physiology, development and behaviour is a challenge not only because of errors and polymorphisms in genome assemblies, but also because the very types of sequence differences that contribute most to these characteristics remain ill-determined. The near-identity of human and chimpanzee orthologous coding sequence led to an initial suggestion that gene expression, rather than coding sequence change, is the major contributor to our differences [13]. However, it has become clear that most of the variations between human and chimpanzee in non-coding sequence are not adaptive either [14]. Identifying adaptive substitutions, whether in coding or non-coding sequence, remains a considerable problem.
Our approach has been not to investigate single nucleotide substitutions as potential substrates of adaptation. Rather, we wish to consider larger sequence differences between human and chimpanzee genomes, namely genes which have duplicated in a lineage-specific manner in the past 6–7 MY since the last common ancestor of the two species.
To this end, we recently determined the number of synonymous (silent) mutations per synonymous site (KS) between closely-related human genes and used this to predict the lineage-specificity of duplication events. We identified a relatively large fraction (5%) of human genes that have participated in duplication events since the last common ancestor with the rodents [11]. Gene pairs that together have accumulated few substitutions in synonymous sites (KS < 0.3) were suggested to be primate-specific. The vast majority of these paralogues pairs have accumulated even fewer silent substitutions (KS < 0.015), indicating that most human duplications occurred only in the past 3–4 million years, after the divergence of Homo and Pan lineages. It is not yet known whether these recent duplications in the mosaic human genome assembly are fixed in the human population, or instead represent CNPs, although the latter explanation now appears increasingly likely [15]. The functions of these recently-duplicated genes are not uniformly distributed. Genes involved in reproduction, chemosensation and host defense and immunity are over-represented [11]. 'Cancer Testis antigen' (CTA) genes, most of which are normally expressed in the testis but are also highly active in certain cancers [16], are another prominent category among the recently-duplicated human gene set. They are represented among a small number of gene families, including one whose founding member is PRAME ('Preferentially expressed antigen of melanoma'), a human gene that is expressed highly in a large proportion of tumours [17,18]. In all cases, the physiological role of CTA genes in normal cells remains unclear, but their recent and extensive duplications are consistent with adaptive functions, such as chemosensation, immunity and reproduction [19]. Moreover, their specific expression in the testis and ovary argues for their involvement in the acquisition of innovative reproductive function during recent primate evolution.
CTA genes frequently have been duplicated on the human X chromosome [11,20] which might indicate a male selective advantage in possessing these genes. A mouse PRAME-like X-linked gene is known to be expressed specifically in spermatogonia, and may perform roles in the early stages of spermatogenesis [21]. Other members of this family are clustered together on an autosome, mouse chromosome 4. Because mammalian sex chromosomes undergo inactivation in late stages of spermatogenesis, it is possible that X-linked PRAME genes may play a part early in spermatogenesis, whereas the cluster of autosomal PRAME genes functions either in later stages, or in other tissues. Indeed, one autosomal mouse PRAME-like gene is known to be expressed in both oocytes and early cleavage-stage embryos [22].
Here we describe the extraordinary recent evolution of autosomal PRAME-like gene clusters on human and chimpanzee chromosomes 1, and mouse chromosome 4. We use a molecular clock, calibrated using synonymous or intronic nucleotide substitutions, to infer the recent origin of many of these PRAME genes. This is corroborated independently by comparison with available chimpanzee genomic sequence. Our analyses confirm that these human genes have duplicated unusually rapidly within the last 3 MY, with concomitant and substantial sequence diversification resulting from adaptive evolution. We predict that the differences between human and chimpanzee PRAME genes contributed to the functional divergence along the hominin lineage.
Results
Evolutionary survey of 7 human CT-Antigen gene families
We investigated whether rapidly-duplicating members of seven CTA families have experienced rapid sequence diversification as a result of adaptive evolution. Using ENSEMBL gene predictions, we initially used codeml [23] to predict sites in their amino acid alignments that have been subject to positive selection. Only 2 of the 7 families, namely the PRAME genes and SSX-like genes, were predicted to contain positively-selected sites (posterior probabilities > 0.9 for each of three model pairs (see Methods)). Because of its large size and because of the large number (23) of positively-selected sites found in an initial analysis (using the NCBI34 genome assembly (data not shown)), we decided to perform a more comprehensive analysis of PRAME genes and pseudogenes in human, chimpanzee and mouse genomes; the human SSX-like family contains only 7 members, for which 6 positively-selected sites were predicted (data not shown).
Recent origin for the PRAME gene cluster
We then investigated whether PRAME genes, located between RefSeq genes DHRS3 and T1A-2 on HSA1, are present in orthologous locations in other vertebrates. Indeed, the mouse genome contains PRAME-like genes in its orthologous region [24], as does the rat genome. However, the orthologous regions of both dog and chicken genome assemblies possess no PRAME homologous genes, as determined by searches of these regions using TBlastn [25]. Moreover, this region of the dog genome assembly contains no clone gaps and insufficiently large (≥ 2.5 kb) fragment gaps to accommodate any missing dog PRAME genes. As humans and rodents shared a more recent common ancestor than either humans and dogs, or humans and chickens, it thus appears likely that one or more PRAME genes were translocated into this genomic location after the divergence with the Laurasiatherian lineage containing extant carnivores (approximately 95 MYA), but before the rodent-primate split (approximately 85 MYA) [26].
Human, chimpanzee and mouse PRAME genes
Our comprehensive reprediction of human PRAME homologues from the 0.74 Mb region of HSA1 yielded a total of 22 PRAME genes and 10 pseudogenes (Figure 1). (These we number sequentially along the assembly, Homo_1, Homo_2, etc.) Each of these genes is approximately 3.0 kb long (average 3069 bases), contain three exons (labelled A, B and C) and two introns (a and b) both with consensus (GT-AG) splice sites. The translated protein is approximately 474 amino acids in length with the three exons having median lengths of 95, 193 and 186 amino acids.
Figure 1 Dot plot representation [63] of a 0.74 Mb region of human chromosome 1 (bases 1276000–1350000) annotated (below) according to the locations of PRAME genes (blue arrowheads) and pseudogenes (red arrowheads), approximately to scale. Gene or pseudogene orientation is indicated by arrowhead direction. PRAME gene or pseudogene numbers are provided beneath the arrowheads. Single short diagonals represent alignments of two PRAME genes or pseudogenes. Gaps in the assembly (bases 13015219–13065218 and 13302469–13352468) are indicated, on the axes, by thick black bars. Two recent segmental duplications (Homo_7–12 and 15–20, and Homo_19–25 and 26–32; see text) are highlighted in blue and pink, respectively. Regions identified by Sebat et al. [12] or by Iafrate et al. [36], as being copy number polymorphic are indicated by a yellow, or a black-and-yellow-striped, bar, respectively.
Initial predictions of chimpanzee PRAME genes, from placed and unplaced PTR1 sequence and from unmapped assembled sequence, yielded 17 candidate genes. Careful inspection, however, revealed these gene predictions to be of poor quality with many predicted genes spanning suspiciously large (>> 3 kb) genomic distances and exhibiting regions of poor sequence similarity. We believe this is a result of the sequence incompleteness and the low (4-fold) statistical coverage of the chimpanzee genome sequence in the assembly. We thus instead resorted to independently predicting each of the three chimpanzee PRAME exons. This resulted in 16 exon A, 24 exon B and 19 exon C predictions. Several of these predictions appear to be identical and could thus be redundant. Adjacent exons and introns were assembled to give 12 putative chimpanzee PRAME genes and pseudogenes (labelled Pan_1, Pan_2 etc.), all of whose introns were confirmed to contain consensus (GT-AG) splice sites. Of these predictions, only 3 appear to be full length, with 3 exons and 2 introns lacking gaps. 30 predictions contain a single exon only. 7 of the 12 sequences contain 3 stop codons and 12 frameshifts, and so might be pseudogenes. (At suggested nucleotide substitution rate of 3 × 10-4 and insertion/deletion error rates of approximately 2 × 10-4 (Tarjei Mikkelsen, personal communication) we expect few if any of these disruptions to arise from sequencing errors (data not shown)).
We inferred relationships between chimpanzee exons and intron, and their human orthologous sequences, using phylogenetic trees (see below). This revealed both well assembled chimpanzee sequence, with consecutive exons and introns assigned to the same human orthologous gene, and poorly assembled sequences, manifested by short contigs, separated by gaps, in a disordered arrangement.
18 PRAME-like genes and 15 pseudogenes were predicted in the orthologous region of mouse chromosome 4. (These are numbered sequentially Mus_1, Mus_2 etc., in the same orientation as that used for the human and chimpanzee numbering scheme.) Of these 5 (Mus_1, Mus_9, Mus_10, Mus_18 and Mus_30) have previously been investigated by Dade et al. [24], who describe these as having roles in oogenesis.
Local gene duplication
In order to visualise the chromosomal landscape of this region of HSA1, we compared its repeat-masked DNA sequence with itself using a dotplot representation (Figure 1). As befits tandemly-duplicated and highly similar sequence, a strong pattern of many diagonals was evident. Each short diagonal represents the DNA alignment of two PRAME genes or pseudogenes. The orientation of the diagonal indicates whether these two genes or pseudogenes are situated on the same, or else the opposite, strand. We observed two pairs of long diagonals (highlighted in colour in Figure 1) which represent two predicted events of segmental duplication (see below).
Human and mouse PRAME genes are monophyletic
We then were able to exploit these gene predictions from human, chimpanzee and mouse to infer the genes' evolutionary relationships and the sequential order of gene duplications. At this stage, we do not rule out that paralogous sequences have been subject to recent inter-locus gene conversion [27,28] which may result in greater sequence similarity and, hence, an apparently more recent date of evolutionary divergence (see Discussion). Dendrograms were constructed from two types of quasi-neutral nucleotide substitution rates: KS values, either for single coding exons, or for complete coding sequence, and KI values, defined as the numbers of nucleotide substitutions per site within intronic sequence.
A phylogenetic tree constructed from human and mouse PRAME gene KS values revealed that mouse sequences are monophyletic, as are human sequences (Figure 2). No pair of mouse and human PRAME genes thus possesses a simple 1:1 orthology relationship. This is a striking result since the vast majority (approximately 80%) of mouse genes possess a single human ortholog [29]. As predicted earlier [11], many human PRAME genes thus have arisen by duplication recently in the primate lineage. What was unexpected, however, is that all mouse, and similarly all human, PRAME sequences have arisen by duplication events that occurred since their last common ancestor, approximately 85 MYA.
Figure 2 Phylogenetic relationships of mouse and human full-length PRAME homologues, inferred using KS as a distance metric. Mouse PRAME homologues (blue lineages) are monophyletic, as are human PRAME homologues (red lineages).
Human PRAME genes have frequently and recently duplicated
Three further phylogenetic trees compared human and chimpanzee KS values from alignments of each of the three PRAME exons (Figure 3; Additional Information). Each of these trees indicates that PRAME gene sequence duplicated frequently in the terminal human branch (i.e. the lineage from the common ancestor of humans and chimpanzees to humans).
Figure 3 Phylogenetic relationships of exons A of human and chimpanzee PRAME homologues, inferred using KS as a distance metric. Phylogenetic relationships derived using alignments of exons B and C are available as Additional files 1 and 2. Homo_9 and Homo_18 are not shown, as these pseudogenes each appears to lack exon A.
Importantly, many pairs of human PRAME genes, and their constituent exons (Figure 3) and introns (Figure 4), were found to exhibit low synonymous rates (Figure 5) that are more typical of duplications in the terminal human branch, than they are of duplications that occurred prior to the common ancestor of humans and chimpanzees, approximately 6–7 MYA [1]. Later in the manuscript we return to the issue of whether these recently-duplicated human genes are present or absent from the chimpanzee genome.
Figure 4 Phylogenetic relationships of introns a of human and chimpanzee PRAME homologues, inferred using K>I as a distance metric, and a neighbour-joining tree. Percentage bootstrap support (1000 iterations) is shown on branches where the support was less than 50%. Phylogenetic relationships derived using an alignment of intron b is available as Additional file 3.
Figure 5 Scatter plot of the lowest neutral rate estimates (either KS calculated from exon, or KI for intron, alignments) for human PRAME genes and either their human paralogues (indicated in red) or their chimpanzee orthologues (indicated in black). Circles represent averages of intronic rate (KI) estimates, whereas squares represent averages of exonic rate (KS) estimates. The horizontal axis represents genomic location within a 0.74 Mb region of human chromosome 1 (see Figure 1). Two recent segmental duplications (Homo_7–12 and 15–20, and Homo_19–25 and 26–32; see text) are highlighted in blue and pink, respectively. The dark line represents the median KS value (3.58 × 10-3) for human paralogues. The grey band identifies 25–75% of this median value (second and third quartiles). The blue line represents the median KS (0.011) for human-chimpanzee coding sequence [30, 31]. The exonic KS value for Homo_12 vs Homo_15 is not shown due to incongruencies in KS-derived phylogenetic trees (see text). Homo-Pan rate estimates are missing when the most-closely related sequences, that are available, are relatively divergent KI or KS > 0.1. These missing values are likely to reflect the incompleteness of the current chimpanzee genome assembly. Homo-Homo rate estimates are missing for 4 genes (Homo_1, 3, 5 and 13) which appear not to have duplicated recently (KI or KS > 0.1).
Assignment of human paralogues and chimpanzee orthologues
By testing for congruency among the three exon (KS) and the two intron (KI) trees (Figures 3 and 4; Additional Information Files 1, 2, 3) we were able to identify the closest human paralogue to each human PRAME gene. The set of assignments was found to be unambiguous and internally consistent, with one notable exception: Homo_15 and Homo_12 are almost identical in their first two exons but are divergent in their exons C. Upon closer inspection it appears that either genome assembly error has generated a chimaeric Homo_12 gene, or else its exon C and a portion of intron b, have been subjected to inter-locus gene conversion with an, as yet unknown, PRAME homologue. Consequently, in subsequent evolutionary rate calculations, comparisons between exons C from Homo_12 and Homo_15 have been discarded. All human PRAME genes, with only 4 exceptions (Homo_1, Homo_3, Homo_5 and Homo_13), are little diverged (KS < 0.1 or KI < 0.1; Figure 5) from another human gene, and are thus part of a pair of sequence similar paralogues which have apparently been generated by a recent gene duplication.
A similar protocol was adopted to identify chimpanzee orthologues of human PRAME exons and introns. For each human exon (or intron), we assigned as its orthologue the chimpanzee exon (or intron) with the lowest KS (or KI) value from the tree, whilst checking to see that these values were approximately 0.011, the median KS value between chimpanzee and human orthologues [30,31]. This process resulted in at least one orthology assignment to the exons or introns of all but 9 (Homo_4, Homo_6, Homo_10, Homo_14, Homo_17, Homo_21, Homo_25, Homo_28, and Homo_32) of the human PRAME homologues; these missing orthologues can be assumed to be present in the chimpanzee genome but absent from its current assembly. For each human orthologue, we then examined the chimpanzee genome assembly for contiguity of its assigned chimpanzee orthologous exons and introns. For example, Pan_1_A, Pan_2_B and Pan_3_C, which are the chimpanzee orthologues of the three exons of Homo_1, appear consecutively within the chimpanzee genome sequence, complete with intervening intronic sequence, and thus were assigned as a full length chimpanzee PRAME, Pan_1. Several chimpanzee orthologue exon pairs appeared not to be contiguous in the current assembly, which again indicates that considerable additional data and attention will be required to provide an accurate assembly of this region.
Pseudogenes
Of 32 HSA1 PRAME homologues, 10 are predicted to be pseudogenes. A similar proportion of chimpanzee PRAME exons are disrupted by at least one stop codon: 19 (3 exon A, 7 exon B and 9 exon C) out of 59 chimpanzee predicted exons contain at least one such disruption. It is probable that some of these are due merely to sequencing or assembly errors due to the low (4-fold) statistical coverage of the chimpanzee genome.
We can safely infer that at least three of these pseudogenes (Homo_9, Homo_13, and Homo_18) were present in the common ancestor to both human and chimpanzee simply because in each case the disruptions coincide between orthologues. Five human sequences (Homo_3, Homo_11, Homo_16, Homo_20 and Homo_27) appear to have become pseudogenes in the hominin lineage as a result of disruptions which are absent from their chimpanzee orthologues. Homo_20 and Homo_27, which differ by only two synonymous substitutions, acquired their disrupting mutation (a stop codon) only recently, since their divergence from the Homo_7 gene, within the last 1 MY (see below).
Dating segmental duplications in the human genome
The branching order of human genes, both from exon KS-based trees (Figure 3; Additional Information) and from intron KI-based trees (Figure 4; Additional Information), indicates that two large-scale duplication events occurred recently in a human ancestral genome. The most recent event appears to have been a single tandem duplication of 7 PRAME homologues to generate a pair of segments encompassing genes Homo_19–25 and genes Homo_26–32 (Figure 1; Figure 5).
We can estimate the age of this duplication using the neutral rate estimates as a molecular clock and calibrating this by the divergence time (6–7 MY) between the human and chimpanzee lineages (see Figure 3 and Additional Information). Previous large-scale studies have shown that the median KS value between human and chimpanzee orthologues is 0.011 [30,31]. The mean KS value between the seven Homo_19–32 genes and their assigned orthologues in chimpanzee was found to be 0.00995. Divergence between these regions of HSA1 and PTR1 thus is typical of these genomes as a whole.
We expect, therefore, that pairs of human paralogues possessing KS values less than approximately 0.01 are likely to have arisen in the terminal human branch, within the past 6–7 MY, whereas human paralogues possessing KS values greater than 0.01 arose due to duplications that occurred prior to the divergence of chimpanzee and human lineages. We calculated that the seven least divergent pairs between Homo_19–25 and Homo_26–32 exhibit a mean divergence of 1.46 × 10-3 which is nearly seven-fold lower than the chimpanzee-human divergence. This indicates an age for this duplication of approximately (1.46 × 10-3 /9.95 × 10-3) × 6 ≈ 0.9 MY. A similar calculation using intronic nucleotide substitution rates KI predicts an age of 0.8 MY. As these predicted ages considerably postdate the split between chimpanzee and human lineages (6–7 MYA), the tandem duplication of Homo_19–25 and Homo_26–32 genes appears to have been a hominin-specific event.
This conclusion is reinforced by the high identity of genomic sequence between the two duplications. Genomic sequences encompassing Homo_19–25 (HSA1 bases 13132294–13272173) and Homo_26–32 (bases 13353079–13493033) PRAME genes are 99.82% identical (161 mismatched bases over ~140 kb). 0.18% divergence, again, is almost seven-fold lower than 1.23%, the average divergence between human and chimpanzee sequence [8,31,32]. This divergence is also twice the average polymorphism rate (0.08%; [11,33,34]) between human individuals and in the human genome assembly. This most recent large-scale duplication of human PRAME genes thus appears to be recent, with respect to the human-chimpanzee divergence event, but ancient, compared with the appearance of most human polymorphisms, within approximately the last 0.10 MY [35].
A more ancient, but still apparently hominin-specific, large, segmental and inverted duplication is that of Homo_7–12 and Homo_15–20 PRAME genes (Figure 1; Figure 5). This duplication's average divergence (mean KI = 0.00275; mean KS = 0.00447) is 2.2–3.6-fold smaller than that expected divergence (≈ 0.010, see above) for human-chimpanzee comparisons, which corresponds to an estimated divergence time of between 1.7 and 2.7 MYA. These estimates again considerably postdate the chimpanzee-human divergence.
Copy number polymorphisms (CNPs)
The recent segmental duplications of human PRAME genes suggest that this region of HSA1 might contain CNPs within the human population. By querying the database of genomic variants [36] we determined that HSA1p36.21, which encompasses these PRAME genes, is one of only 11 polymorphic loci found in two large-scale CNP investigations [12,36,37] (see Figure 1). This implies that not only has this region undergone two large-scale duplications in ~ 3 MY, but that there have been additional, more recent, duplications which are not fixed in the human population and have not been captured in the human genome reference sequence.
Positive selection of PRAME genes
Gene duplication in a genome provides a substrate upon which selection may act. The preservation of duplicates without disruption to their open-reading frames over millennia is itself an indication that these duplicates confer a selective benefit to the host organism. More direct evidence of positive selection comes from the elevated values of the ratio of KA, the number of nonsynonymous (amino acid changing) substitutions per nonsynonymous site, to KS. After discarding closely-related sequences (KS < 0.02), the median KA/KS ratio between pairs of human PRAME genes is 0.73, and 19 gene pairs exhibit KA/KS ratios greater than 1, with a maximum value of 1.73 between Homo_6 and Homo_10. These values are considerably higher than the average ratio between human and rodent single gene orthologues (median KA/KS~ 0.12) [29].
KA/KS values of approximately 1 might be due to positive selection of nonsynonymous nucleotide substitutions, or to reduced selective constraints due to the loss of PRAME genes' functions. In order to distinguish between these hypotheses, and to further investigate the evolution of these genes, we used codeml [38-40] to infer positive selection at single sites within multiple alignments of human or mouse PRAME genes.
Among human PRAME genes, a large number (30) of amino acid sites were identified as having been subject to positive selection. By mapping these sites to a homologous protein structure, that of porcine ribonuclease inhibitor, we observed that these sites aggregate to form a pronounced cluster on one exterior face (Figure 6). The majority of these sites would thus be available to participate in binding interactions. A similar analysis of mouse PRAME genes also demonstrated the impact of positive selection: 17 positively selected sites were identified, of which 4 coincide with such sites among human PRAME genes.
Figure 6 Structure of porcine ribonuclease inhibitor (PDB code 2BNH) with amino acid sites that are positively-selected among human and mouse PRAME proteins shown in red and blue, respectively, ((A) front view, (B) rear view).
Discussion
Our findings demonstrate an extraordinarily rapid expansion within this PRAME gene family that occurred independently in both primate and rodent lineages. Given the high conservation of gene order among chicken, dog, human and rodent genomes, we can date the origin of this cluster to between approximately 95 and 85 MYA [26]. This is because PRAME homologues are undetectable in the orthologous region of the chicken and dog genomes, but are present in syntenic portions of primate and rodent genomes. Thereafter, many episodes of gene duplication have occurred in both primate and rodent lineages.
In order to infer the most recent of these duplication events, we identified 13 pairs of human PRAME paralogues which appear to have arisen by duplication since the common ancestor with chimpanzees: their divergence is considerably less than both the expected and the observed divergence between orthologous human and chimpanzee sequence (Figure 1 and Figure 5). Using a molecular clock, and a palaeontological calibration of divergence between these two species of 6–7 MYA [1], we estimate that two large segmental duplications of PRAME genes occurred independently in the terminal human branch, within approximately the past 3 MY.
The low divergence of these human paralogues, compared with the divergence between chimpanzee and human orthologues, argues strongly that chimpanzee lacks single orthologues of many, if not all, of these human duplicated genes. Insufficient nucleotide substitutions at synonymous and intronic sites have accumulated to indicate that Homo_7–12 and Homo_15–32 genes were all present in single copies in the common ancestor of chimpanzee and human. Even when the chimpanzee genome is completed, we thus expect that chimpanzee single orthologues of these genes will not be identified.
In addition to these duplications which are apparent in the human genome assembly, it appears, from two independent studies [36,12], that the region of human chromosome 1 (HSA1) containing these PRAME genes is copy number polymorphic. These human genes are present in different numbers among the human population, thus providing further evidence that human-specific duplications are a feature of this region. CNPs are thought not to be selectively neutral [12]. Their persistence in the human population suggests, rather, that at least a subset of CNPs may be adaptive.
In support of this hypothesis, we found that a large number (30) of codons in the HSA1 PRAME family have been subject to positive selection. 26 of these adaptive codons were confirmed using the "sitewise likelihood-ratio" (SLR) method [41] (data not shown). These sites are clustered onto one surface in a homology model of protein structure, thereby demarcating a likely surface-accessible functional site. Mouse PRAME genes also contain a large number (17) of positively selected sites, which cluster within a site equivalent to that for human PRAME genes (Figure 6).
Expansion of this PRAME gene family has occurred independently in both primate and rodent lineages. In each of these lineages, PRAME genes appear to have evolved by 'birth-and-death' processes, such as occurs for immunity genes [42]: genes both persist as duplications, and are lost by pseudogene creation. Sequence similarities between paralogues, however, could have arisen also from concerted evolution, as the result of homologous recombination, in particular, gene conversion and unequal crossing over [43]. Nevertheless, the recent origin of the PRAME progenitor gene just prior to the common ancestor of primates and rodents, and its rapid duplication thereafter, and the occurrence of CNPs in the human population, each indicates that the predominant process in this expansion has been gene duplication. Moreover, the congruency of dendrograms associated with separate exons or introns (Figures 3 and 4, and Additional files 1, 2, 3), and the tandem segmental duplications we have inferred (Figure 1; Figure 5), also argue against concerted evolution as a dominant evolutionary mechanism.
PRAME genes have arisen by rapid gene duplication and pseudogene creation, and their sequences have been subject to positive selection. Nevertheless, the adaptive advantages conferred on these genes by their duplication and sequence diversification remain unclear, as are the genes' functions in normal tissues. Their expression profile often is limited to testes and to a wide variety of tumors, which suggests that PRAME proteins might perform important mitotic roles in rapidly dividing cells. This is consistent with the observation that a PRAME-like gene (oogenesin) in mouse accumulates in the nucleus only at the late one-cell and early two-cell stages of early embryos [22].
At least one of the mouse orthologues of these PRAME genes is expressed in spermatogonial cells [21]. Interestingly, it is known that a mutation in FGFR2 expressed in these cells confers a selective advantage, thereby leading to clonal expansion similar to that seen in tumours [44]. We suggest that similar nonsynonymous substitutions in PRAME genes might have conferred comparable benefits to these cells, and these have driven positive selection for both gene duplication and sequence change.
The evidence thus points mainly to darwinian selection in spermatogonia, and adaptive evolution of PRAME genes thus may be thought not to have phenotypic consequences at larger anatomical scales. Nevertheless, a gene with a similar function, and a similar evolutionary history, to PRAME genes, has been proposed to having contributed to anatomical adaptations during recent hominid evolution. The Abnormal spindle-like microcephaly associated (ASPM) gene, which has evolved rapidly in the great apes [45], has roles in both spermatogenesis and oogenesis in Drosophila [46,47]. When mutated in humans, this results in primary microcephaly, which is manifested by a greatly reduced brain size [6]. Further examination of human PRAME genes' functions should assist in our understanding of the cellular and physiological consequences of its recent and rapid evolution.
Conclusion
Whatever the selective advantages conferred by the PRAME genes discussed here, it is apparent from their recent introduction to an ancestral chromosome of primates and rodents that they benefited only these lineages. Moreover, the HSA1 PRAME gene family has expanded further, in particular by two large segmental duplications in the past 3 MY and further duplications that manifest themselves as copy number polymorphisms in the human population. These extremely rapid duplications, taken together with strong evidence for darwinian adaptation at approximately 30 sites among human PRAME genes, indicates that this family has experienced sustained episodes of positive selection during recent hominin history.
Methods
Survey of human CT-Antigen genes
We recently identified 41 human gene families that have experienced multiple gene duplications since the divergence of rodent and primate lineages [11]. Among these were 6 families, SSX-like, MAGE, GAGE, XAGE, SAGE and SPAN-X, all encoded on the X chromosome, which exhibit the tissue expression profiles of CT-antigens (CTAs). A seventh CTA family, recently duplicated in our lineage, encode PRAME genes that are located on Homo sapiens chromosome 1 (HSA1). In an initial survey, only 2 of these 7 families yielded evidence of positive selection at individual sites using codeml (data not shown), using methods and criteria described below. Subsequently, we chose to perform more comprehensive analyses of the PRAME gene family because of its larger size (13 human ENSEMBL predicted genes) and its high number of positively selected sites identified.
Prediction of human PRAME genes and exons
The amino acid sequences of 5 PRAME homologues were aligned using CLUSTALW [48]. These are genes that were identified by Ensembl [11,49] and lie in a cluster between bases 12550000 and 13100000 of human chromosome 1 (assembly NCBI35). (Many of the remaining 8 ENSEMBL PRAME genes were mispredicted.) From this multiple alignment a hidden Markov model (HMM) was constructed [50]. On the basis of strong conservation of a translation initiating methionine codon, presumed non-coding sequence upstream of this codon was discarded.
In order to ensure gene prediction fidelity and completeness, we repredicted PRAME homologous genes and pseudogenes from this region of HSA1 (bases 12550000 and 13100000, which include the flanking non-homologous RefSeq genes, DHRS3 and T1A-2). Gene prediction employed Genewise [51], the PRAME HMM and default parameters. Upon building phylogenetic trees, the predicted PRAME homologues were found to be monophyletic, and are only distantly-related to other PRAME homologous genes located elsewhere on HSA1 and on HSA22 (data not shown). Pseudogenes were distinguished from genes on the basis of premature stop codons or frameshifts; pseudogenes that are functionally disrupted due only to mutations that occur outside of coding sequence are thus misassigned.
In an independent approach, we also predicted homologues of each of the three protein-coding exons of these PRAME genes within this region of HSA1 using HMMs of multiply aligned nucleotide exonic sequence. This procedure resulted in no predictions that were additional to those found using Genewise and full-length gene sequence.
Prediction of mouse PRAME genes
Mouse PRAME genes were predicted as described above for human genes, except for the use of known mouse 'PRAME-like' (PRAMEL) genes to derive the HMM used in the Genewise step. Initially a CLUSTALW alignment of three known mouse RefSeq genes (PRAMEL1 [RefSeq code: NM_031377.1], PRAMEL3 [NM_031390.1] and PRAMEL4 [NM_178248.2]) was used for the HMM query template. The 3' ends of the PRAMEL genes, however, were found to be relatively divergent. The alignment was thus trimmed back to exclude the ends of the third, and final, exons. Thus the alignment of mouse PRAME proteins is 58 amino acids shorter than that of human PRAME proteins. Mouse genes were predicted within the orthologous region of the mouse genome (Mus musculus chromosome 4 [MMU4]; May 2004 assembly; bases 141,850,000–142,800,000) between mouse DHRS3 and T1A-2. 29 mouse PRAMEL genes and pseudogenes were predicted from this approach. An HMM was then derived from a multiple alignment of these sequences and used to query this region of MMU4 in a second round of searches. Four additional predictions were found.
Prediction of chimpanzee PRAME genes and exons
Chimpanzee (Pan troglodytes) PRAME genes were predicted as for human genes, as described above, using the HMM derived from the alignment of human PRAME amino acid sequences as query to search a region lying between orthologous DHRS3 and T1A-2 genes on chimpanzee chromosome 1 (PTR1, bases 10240000–13450000, Nov 2003 assembly). This method identified 17 full-length chimpanzee PRAME genes. This gene count was substantially fewer than for the human genome assembly, and may be a consequence of the low (four-fold) statistical coverage of the chimpanzee genome assembly. We reasoned, therefore, that additional non-full-length PRAME gene exons might be represented in the chimpanzee genome assembly. We thus predicted homologues of each of the three PRAME gene exons in this region using the protocol (described above) that was used for predicting human PRAME gene exons.
Prediction of introns
Human and chimp intron sequences were identified as the sequence intervening between adjacent exons. PRAME genes contain 3 coding exons (labelled A, B, C) and 2 intervening introns, labelled intron a and intron b. For human sequence, these introns were all complete, without gaps, but for chimpanzee sequence only 9 intact intron a and 5 intron b chimpanzee introns could be identified.
Exon, intron and splice site predictions from these three species were all consistent with the gene structures apparent from available cDNAs, in particular 14 cDNAs mapped to human HSA1 bases 12769277–1349033, 29 cDNAs mapped to mouse MMU4 bases 141852757–142731257, and cDNAs from the eponymous PRAME gene on HSA22 (bases 21,215,046–21,218,065). Conservation of gene structures between mammals as diverse as human and mouse, together with conservation of splice sites, indicates that chimpanzee PRAME genes also possess an identical gene structure.
Sequence alignments
Conceptual translations of PRAME genes and pseudogenes were aligned using CLUSTAL W [48] and then modified to minimise gaps (see Additional files 1, 2, 3). Stop characters were replaced by 'X'. Estimates of KA and KS for sequence pairs (see below) were calculated from cDNA sequences aligned according to these amino acid multiple alignments.
Human and chimpanzee nucleotide intronic sequences were aligned using DIALIGN-2 [52].
Exonic evolutionary rates
Codeml [23] was used to conduct site-specific KA/KS analysis on the human and mouse full length PRAME predictions. An amino acid alignment and corresponding cDNA alignment were prepared for each analysis. Identified pseudogenes were removed from the alignment because they are likely to be no longer subject to selective constraints.
The maximum likelihood approach of Yang [40] was used to predict sites in a group of cDNA sequences that have been subject to positive selection. Pairs of models were compared by calculating log likelihood values (l), which were then compared for significant differences using a Likelihood Ratio Test. The first of each pair of models compared is a simple model where sites are predicted to be associated with KA/KS ratios between 0 and 1. The second is a more complex model that allows adaptive sites: for these, ratios can be greater than 1. If the complex model indicates an estimated KA/KS ratio that is greater than one, and the test statistic (2Δl) is greater than critical values of the Chi square (χ2) distribution with the appropriate degree of freedom [53], then positive selection can be inferred. Bayesian probabilities are used to predict which codons in the original data have most likely been subjected to positive selection.
The pairs of simple and complex models we used were: M0 (one-ratio) [54] versus M3 (discrete) [23]; M1 (neutral) versus M2 (selection) [55]; and M7 (beta) versus M8 (beta + ω) [23,40]. Only non-conserved alignment positions predicted to be under positive selection with a posterior probability > 0.90 by all three codeml models were mapped onto a homologous protein structure (Figure 6).
Intronic evolutionary rates
Using the DIALIGN-2 alignment of intronic sequences, we calculated their genetic distances using the TN93 nucleotide substitution model [56]. We then derived a phylogenetic tree based on the distance matrix using neighbour-joining methods (1000 bootstrap iterations). Numbers of nucleotide substitutions per intronic site between sequence pairs (KI) were then estimated using BASEML [38,39], a maximum likelihood method, and the TN93 nucleotide substitution model. This analysis was implemented using the DAMBE package [57].
Structure
In order to gain insight into the functional relevance of positively selected sites, we searched the protein sequences of known tertiary structure using PSI-BLAST [25] using a human PRAME sequence (Homo_7; UniProt: YA03_HUMAN) at NCBI using default parameters. Significant sequence similarity (E = 1 × 10-12) was found after three search iterations to porcine ribonuclease inhibitor, RNI (PDB code 2BNH). Two types of alignment guided the assignment of leucine-rich repeats (LRRs) to human PRAME sequences. First, the BLAST alignment, and second the optimal and suboptimal alignments of PRAME sequence against the SMART [58] LRR HMM. RNI was first aligned to Homo_7 using these methods, and adjusted manually, and then aligned to the full alignment of all human PRAMEs guided by the Homo_7 alignment. This allowed human PRAME positively selected sites (as identified by the method above) to be mapped to RNI residues. This procedure was also followed to align full-length mouse PRAMEL sequences against RNI. Protein tertiary structure was viewed, manipulated and annotated in Swiss Pdbviewer [59].
Evolutionary relationships
Phylogenetic relationships were deduced from dendrograms constructed from three types of neutral rate estimates: either (i) KS values of pairwise alignments of full-length coding sequences; (ii) KS values of pairwise alignments of coding sequences from exons A, B or C; or (iii) KI values of pairwise alignments of intronic sequences from introns a or b. Dendrograms were constructed using code based on PHYLIP [60] which uses the Fitch-Margoliash criteria to build trees with contemporaneous tips. The dendrograms were visualised in njplot and treeview [61].
Copy number polymorphisms (CNPs)
The database of Genomic variants ([62,12,36]) was queried to determine whether the human PRAME region has been determined previously to harbour CNPs.
List of abbreviations
BLAST – Basic Local Alignment Search Tool
CNP – copy number polymorphism
CTA – cancer testis antigen
HMM – hidden Markov model
HSA1 – human chromosome 1
MYA – million years ago
NCBI – The National Centre for Biotechnology Information, USA
PDB – protein data bank
PRAME – preferentially expressed antigen of melanoma
PSI-BLAST – Position specific iterative BLAST
SNP – single nucleotide polymorphism
Authors' contributions
ZB predicted and aligned PRAME genes, identified the positively selected sites and drew the trees. ZB also wrote the first draft of the manuscript. LG retrieved the initial gene sequences and developed an analytical pipeline which proved critical to this project. CP participated in sequence alignment, analysis of the data and writing the manuscript, in particular the discussion. All Authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Phylogenetic relationships of exons B of human and chimpanzee PRAME homologues, inferred using KS as a distance metric
Click here for file
Additional File 2
Phylogenetic relationships of exons C of human and chimpanzee PRAME homologues, inferred using KSas a distance metric
Click here for file
Additional File 3
Phylogenetic relationships of introns b of human and chimpanzee PRAME homologues, inferred using KI as a distance metric, and a neighbour-joining tree. Percentage bootstrap support (1000 iterations) is shown on branches where the support was less than 50%. In the case of Homo_18_ib and Homo_9_ib, the branches leading to these sequences were too small to display the bootstrap values clearly. The bootstrap support values for these branches are 25% and 45% respectively.
Click here for file
Additional File 4
Human_prames_pep.aln • Clustal format • Human PRAME predictions • This contains an alignment of human predicted PRAME polypeptides. This alignment (without pseudogenes) was used to identify positively-selected sites.
Click here for file
Additional File 5
Human_prames_cdna.fa • FASTA format • Human PRAME cdnas • This contains the human PRAME cDNAs, as predicted by Genewise.
Click here for file
Additional File 6
Human_mouse_2BNH.aln • Clustal format • Human and mouse PRAME alignment with positively selected sites. • This shows all human and mouse PRAMEs (including pseudogenes) aligned. Two lines, homo_sites and mus_sites, indicate positively selected sites ('X') in the human PRAMEs and mouse PRAMEs respectively. This alignment (without pseudogenes) was used to identify positively-selected sites.
Click here for file
Additional File 7
Human_chimp_exonA_cdna.fa • FASTA format • Human and chimpanzee exon A cDNA • This contains the human and chimp exon A cDNA as predicted by Genewise.
Click here for file
Additional File 8
Human_chimp_exonB_cdna.fa • FASTA format • Human and chimpanzee exon B cDNA • This contains the human and chimp exon B cDNA as predicted by Genewise.
Click here for file
Additional File 9
Human_chimp_exonC_cdna.fa • FASTA format • Human and chimpanzee exon C cDNA • This contains the human and chimp exon C cDNA as predicted by Genewise.
Click here for file
Additional File 10
Human_chimp_exonA_pep.aln • Clustal format • Human and chimpanzee exon A peptide alignment • This contains human and chimp exons A peptide sequences aligned. This alignment was used to calculate the KS distances between sequences.
Click here for file
Additional File 11
Human_chimp_exonB_pep.aln • Clustal format • Human and chimpanzee exon B peptide alignment • This contains human and chimp exons B peptide sequences aligned. This alignment was used to calculate the KS distances between sequences.
Click here for file
Additional File 12
Human_chimp_exonC_pep.aln • Clustal format • Human and chimpanzee exon C peptide alignment • This contains human and chimp exons C peptide sequences aligned. This alignment was used to calculate the KS distances between sequences.
Click here for file
Additional File 13
Chimp_exonA_cdna.fa • FASTA format • Chimpanzee exon A cDNA • This contains the (unaligned) chimp exons A cDNA as predicted by Genewise.
Click here for file
Additional File 14
Chimp_exonB_cdna.fa • FASTA format • Chimpanzee exon B cDNA • This contains the (unaligned) chimp exons B cDNA as predicted by Genewise.
Click here for file
Additional File 15
Chimp_exonC_cdna.fa • FASTA format • Chimpanzee exon C cDNA • This contains the (unaligned) chimp exons C cDNA as predicted by Genewise.
Click here for file
Additional File 16
Human_prames_info.xls • Excel spreadsheet format • Human PRAME information • This excel spread sheet contains details of human PRAMEs, including translational start and end positions, length, chimpanzee orthologues and pseudogene assignment.
Click here for file
Additional File 17
chimp_prames_info.xls • Excel spreadsheet format • Full length chimpanzee PRAME information • This excel spread sheet contains details of full length chimpanzee PRAMEs. It indicates exons which constitute each PRAME and contains details of pseudogene assignment.
Click here for file
Additional File 18
Human_chimp_introna.aln • Clustal format • Human and chimpanzee introns a sequence • This contains human and chimpanzee introns a sequence aligned.
Click here for file
Additional File 19
Human_chimp_intronb.aln • Clustal format • Human and chimpanzee introns b sequence • This contains human and chimpanzee intron b sequences aligned.
Click here for file
Acknowledgements
We thank Ponting group members for helpful discussions and two anonymous reviewers for their helpful comments. This work was funded by the UK Medical Research Council whose support is gratefully acknowledged.
==== Refs
Brunet M Guy F Pilbeam D Mackaye HT Likius A Ahounta D Beauvilain A Blondel C Bocherens H Boisserie JR De Bonis L Coppens Y Dejax J Denys C Duringer P Eisenmann V Fanone G Fronty P Geraads D Lehmann T Lihoreau F Louchart A Mahamat A Merceron G Mouchelin G Otero O Pelaez Campomanes P Ponce De Leon M Rage JC Sapanet M Schuster M Sudre J Tassy P Valentin X Vignaud P Viriot L Zazzo A Zollikofer C A new hominid from the Upper Miocene of Chad, Central Africa Nature 2002 418 145 151 12110880 10.1038/nature00879
Ruvolo M Comparative primate genomics: the year of the chimpanzee Curr Opin Genet Dev 2004 14 650 656 15531160 10.1016/j.gde.2004.08.007
Enard W Przeworski M Fisher SE Lai CS Wiebe V Kitano T Monaco AP Paabo S Molecular evolution of FOXP2, a gene involved in speech and language Nature 2002 418 869 872 12192408 10.1038/nature01025
Stedman HH Kozyak BW Nelson A Thesier DM Su LT Low DW Bridges CR Shrager JB Minugh-Purvis N Mitchell MA Myosin gene mutation correlates with anatomical changes in the human lineage Nature 2004 428 415 418 15042088 10.1038/nature02358
Winter H Langbein L Krawczak M Cooper DN Jave-Suarez LF Rogers MA Praetzel S Heidt PJ Schweizer J Human type I hair keratin pseudogene phihHaA has functional orthologs in the chimpanzee and gorilla: evidence for recent inactivation of the human gene after the Pan-Homo divergence Hum Genet 2001 108 37 42 11214905 10.1007/s004390000439
Bond J Roberts E Mochida GH Hampshire DJ Scott S Askham JM Springell K Mahadevan M Crow YJ Markham AF Walsh CA Woods CG ASPM is a major determinant of cerebral cortical size Nat Genet 2002 32 316 320 12355089 10.1038/ng995
Ponting CP Jackson A Evolution of primary microcephaly genes and the enlargement of primate brain Curr Opin genet dev 2005 15 241 8 15917198 10.1016/j.gde.2005.04.009
Ebersberger I Metzler D Schwarz C Paabo S Genomewide comparison of DNA sequences between humans and chimpanzees Am J Hum Genet 2002 70 1490 1497 11992255 10.1086/340787
Shi J Xi H Wang Y Zhang C Jiang Z Zhang K Shen Y Jin L Yuan W Lin J Hua Q Wang F Xu S Ren S Zhao G Chen Z Huang W Divergence of the genes on human chromosome 21 between human and other hominoids and variation of substitution rates among transcription units Proc Natl Acad Sci U S A 2003 100 8331 8336 12826612 10.1073/pnas.1332748100
University of Washington in St Louis. Chimpanzee Sequencing project.
Consortium IHGS Finishing the euchromatic sequence of the human genome Nature 2004 431 931 945 15496913 10.1038/nature03001
Sebat J Lakshmi B Troge J Alexander J Young J Lundin P Maner S Massa H Walker M Chi M Navin N Lucito R Healy J Hicks J Ye K Reiner A Gilliam TC Trask B Patterson N Zetterberg A Wigler M Large-scale copy number polymorphism in the human genome Science 2004 305 525 528 15273396 10.1126/science.1098918
King MC Wilson AC Evolution at two levels in humans and chimpanzees Science 1975 188 107 116 1090005
Keightley PD Lercher MJ Eyre-Walker A Evidence for widespread degradation of gene control regions in hominid genomes PLoS Biol 2005 3 e42 15678168 10.1371/journal.pbio.0030042
Tuzun E Sharp AJ Bailey JA Kaul R Morrison VA Pertz LM Haugen E Hayden H Albertson D Pinkel D Olson MV Eichler EE Fine-scale structural variation of the human genome Nat Genet 2005 37 727 32 15895083 10.1038/ng1562
Scanlan MJ Gure AO Jungbluth AA Old LJ Chen YT Cancer/testis antigens: an expanding family of targets for cancer immunotherapy Immunol Rev 2002 188 22 32 12445278 10.1034/j.1600-065X.2002.18803.x
Ikeda H Lethe B Lehmann F van Baren N Baurain JF de Smet C Chambost H Vitale M Moretta A Boon T Coulie PG Characterization of an antigen that is recognized on a melanoma showing partial HLA loss by CTL expressing an NK inhibitory receptor Immunity 1997 6 199 208 9047241 10.1016/S1074-7613(00)80426-4
van Baren N Chambost H Ferrant A Michaux L Ikeda H Millard I Olive D Boon T Coulie PG PRAME, a gene encoding an antigen recognized on a human melanoma by cytolytic T cells, is expressed in acute leukaemia cells Br J Haematol 1998 102 1376 1379 9753074 10.1046/j.1365-2141.1998.00982.x
Emes RD Goodstadt L Winter EE Ponting CP Comparison of the genomes of human and mouse lays the foundation of genome zoology Hum Mol Genet 2003 12 701 709 12651866 10.1093/hmg/ddg078
Ross MT Grafham DV Coffey AJ Scherer S McLay K Muzny D Platzer M Howell GR Burrows C Bird CP Frankish A Lovell FL Howe KL Ashurst JL Fulton RS Sudbrak R Wen G Jones MC Hurles ME Andrews TD Scott CE Searle S Ramser J Whittaker A Deadman R Carter NP Hunt SE Chen R Cree A Gunaratne P Havlak P Hodgson A Metzker ML Richards S Scott G Steffen D Sodergren E Wheeler DA Worley KC Ainscough R Ambrose KD Ansari-Lari MA Aradhya S Ashwell RI Babbage AK Bagguley CL Ballabio A Banerjee R Barker GE Barlow KF Barrett IP Bates KN Beare DM Beasley H Beasley O Beck A Bethel G Blechschmidt K Brady N Bray-Allen S Bridgeman AM Brown AJ Brown MJ Bonnin D Bruford EA Buhay C Burch P Burford D Burgess J Burrill W Burton J Bye JM Carder C Carrel L Chako J Chapman JC Chavez D Chen E Chen G Chen Y Chen Z Chinault C Ciccodicola A Clark SY Clarke G Clee CM Clegg S Clerc-Blankenburg K Clifford K Cobley V Cole CG Conquer JS Corby N Connor RE David R Davies J Davis C Davis J Delgado O Deshazo D Dhami P Ding Y Dinh H Dodsworth S Draper H Dugan-Rocha S Dunham A Dunn M Durbin KJ Dutta I Eades T Ellwood M Emery-Cohen A Errington H Evans KL Faulkner L Francis F Frankland J Fraser AE Galgoczy P Gilbert J Gill R Glockner G Gregory SG Gribble S Griffiths C Grocock R Gu Y Gwilliam R Hamilton C Hart EA Hawes A Heath PD Heitmann K Hennig S Hernandez J Hinzmann B Ho S Hoffs M Howden PJ Huckle EJ Hume J Hunt PJ Hunt AR Isherwood J Jacob L Johnson D Jones S de Jong PJ Joseph SS Keenan S Kelly S Kershaw JK Khan Z Kioschis P Klages S Knights AJ Kosiura A Kovar-Smith C Laird GK Langford C Lawlor S Leversha M Lewis L Liu W Lloyd C Lloyd DM Loulseged H Loveland JE Lovell JD Lozado R Lu J Lyne R Ma J Maheshwari M Matthews LH McDowall J McLaren S McMurray A Meidl P Meitinger T Milne S Miner G Mistry SL Morgan M Morris S Muller I Mullikin JC Nguyen N Nordsiek G Nyakatura G O'Dell CN Okwuonu G Palmer S Pandian R Parker D Parrish J Pasternak S Patel D Pearce AV Pearson DM Pelan SE Perez L Porter KM Ramsey Y Reichwald K Rhodes S Ridler KA Schlessinger D Schueler MG Sehra HK Shaw-Smith C Shen H Sheridan EM Shownkeen R Skuce CD Smith ML Sotheran EC Steingruber HE Steward CA Storey R Swann RM Swarbreck D Tabor PE Taudien S Taylor T Teague B Thomas K Thorpe A Timms K Tracey A Trevanion S Tromans AC d'Urso M Verduzco D Villasana D Waldron L Wall M Wang Q Warren J Warry GL Wei X West A Whitehead SL Whiteley MN Wilkinson JE Willey DL Williams G Williams L Williamson A Williamson H Wilming L Woodmansey RL Wray PW Yen J Zhang J Zhou J Zoghbi H Zorilla S Buck D Reinhardt R Poustka A Rosenthal A Lehrach H Meindl A Minx PJ Hillier LW Willard HF Wilson RK Waterston RH Rice CM Vaudin M Coulson A Nelson DL Weinstock G Sulston JE Durbin R Hubbard T Gibbs RA Beck S Rogers J Bentley DR The DNA sequence of the human X chromosome Nature 2005 434 325 337 15772651 10.1038/nature03440
Wang PJ McCarrey JR Yang F Page DC An abundance of X-linked genes expressed in spermatogonia Nat Genet 2001 27 422 426 11279525 10.1038/86927
Minami N Aizawa A Ihara R Miyamoto M Ohashi A Imai H Oogenesin is a novel mouse protein expressed in oocytes and early cleavage-stage embryos Biol Reprod 2003 69 1736 1742 12890732 10.1095/biolreprod.103.018051
Yang Z Nielsen R Estimating synonymous and nonsynonymous substitution rates under realistic evolutionary models Mol Biol Evol 2000 17 32 43 10666704
Dade S Callebaut I Mermillod P Monget P Identification of a new expanding family of genes characterized by atypical LRR domains. Localization of a cluster preferentially expressed in oocyte FEBS Lett 2003 555 533 538 14675769 10.1016/S0014-5793(03)01341-3
Altschul SF Madden TL Schaffer AA Zhang J Zhang Z Miller W Lipman DJ Gapped BLAST and PSI-BLAST: a new generation of protein database search programs Nucleic Acids Res 1997 25 3389 3402 9254694 10.1093/nar/25.17.3389
Springer MS Murphy WJ Eizirik E O'Brien SJ Placental mammal diversification and the Cretaceous-Tertiary boundary Proc Natl Acad Sci U S A 2003 100 1056 1061 12552136 10.1073/pnas.0334222100
Collier S Tassabehji M Sinnott P Strachan T A de novo pathological point mutation at the 21-hydroxylase locus: implications for gene conversion in the human genome Nat Genet 1993 3 260 265 8485582 10.1038/ng0393-260
Urabe K Kimura A Harada F Iwanaga T Sasazuki T Gene conversion in steroid 21-hydroxylase genes Am J Hum Genet 1990 46 1178 1186 1971153
Waterston RH Lindblad-Toh K Birney E Rogers J Abril JF Agarwal P Agarwala R Ainscough R Alexandersson M An P Antonarakis SE Attwood J Baertsch R Bailey J Barlow K Beck S Berry E Birren B Bloom T Bork P Botcherby M Bray N Brent MR Brown DG Brown SD Bult C Burton J Butler J Campbell RD Carninci P Cawley S Chiaromonte F Chinwalla AT Church DM Clamp M Clee C Collins FS Cook LL Copley RR Coulson A Couronne O Cuff J Curwen V Cutts T Daly M David R Davies J Delehaunty KD Deri J Dermitzakis ET Dewey C Dickens NJ Diekhans M Dodge S Dubchak I Dunn DM Eddy SR Elnitski L Emes RD Eswara P Eyras E Felsenfeld A Fewell GA Flicek P Foley K Frankel WN Fulton LA Fulton RS Furey TS Gage D Gibbs RA Glusman G Gnerre S Goldman N Goodstadt L Grafham D Graves TA Green ED Gregory S Guigo R Guyer M Hardison RC Haussler D Hayashizaki Y Hillier LW Hinrichs A Hlavina W Holzer T Hsu F Hua A Hubbard T Hunt A Jackson I Jaffe DB Johnson LS Jones M Jones TA Joy A Kamal M Karlsson EK Karolchik D Kasprzyk A Kawai J Keibler E Kells C Kent WJ Kirby A Kolbe DL Korf I Kucherlapati RS Kulbokas EJ Kulp D Landers T Leger JP Leonard S Letunic I Levine R Li J Li M Lloyd C Lucas S Ma B Maglott DR Mardis ER Matthews L Mauceli E Mayer JH McCarthy M McCombie WR McLaren S McLay K McPherson JD Meldrim J Meredith B Mesirov JP Miller W Miner TL Mongin E Montgomery KT Morgan M Mott R Mullikin JC Muzny DM Nash WE Nelson JO Nhan MN Nicol R Ning Z Nusbaum C O'Connor MJ Okazaki Y Oliver K Overton-Larty E Pachter L Parra G Pepin KH Peterson J Pevzner P Plumb R Pohl CS Poliakov A Ponce TC Ponting CP Potter S Quail M Reymond A Roe BA Roskin KM Rubin EM Rust AG Santos R Sapojnikov V Schultz B Schultz J Schwartz MS Schwartz S Scott C Seaman S Searle S Sharpe T Sheridan A Shownkeen R Sims S Singer JB Slater G Smit A Smith DR Spencer B Stabenau A Stange-Thomann N Sugnet C Suyama M Tesler G Thompson J Torrents D Trevaskis E Tromp J Ucla C Ureta-Vidal A Vinson JP Von Niederhausern AC Wade CM Wall M Weber RJ Weiss RB Wendl MC West AP Wetterstrand K Wheeler R Whelan S Wierzbowski J Willey D Williams S Wilson RK Winter E Worley KC Wyman D Yang S Yang SP Zdobnov EM Zody MC Lander ES Initial sequencing and comparative analysis of the mouse genome Nature 2002 420 520 562 12466850 10.1038/nature01262
Clark AG Glanowski S Nielsen R Thomas PD Kejariwal A Todd MA Tanenbaum DM Civello D Lu F Murphy B Ferriera S Wang G Zheng X White TJ Sninsky JJ Adams MD Cargill M Inferring nonneutral evolution from human-chimp-mouse orthologous gene trios Science 2003 302 1960 1963 14671302 10.1126/science.1088821
Chen FC Li WH Genomic divergences between humans and other hominoids and the effective population size of the common ancestor of humans and chimpanzees Am J Hum Genet 2001 68 444 456 11170892 10.1086/318206
Fujiyama A Watanabe H Toyoda A Taylor TD Itoh T Tsai SF Park HS Yaspo ML Lehrach H Chen Z Fu G Saitou N Osoegawa K de Jong PJ Suto Y Hattori M Sakaki Y Construction and analysis of a human-chimpanzee comparative clone map Science 2002 295 131 134 11778049 10.1126/science.1065199
Sachidanandam R Weissman D Schmidt SC Kakol JM Stein LD Marth G Sherry S Mullikin JC Mortimore BJ Willey DL Hunt SE Cole CG Coggill PC Rice CM Ning Z Rogers J Bentley DR Kwok PY Mardis ER Yeh RT Schultz B Cook L Davenport R Dante M Fulton L Hillier L Waterston RH McPherson JD Gilman B Schaffner S Van Etten WJ Reich D Higgins J Daly MJ Blumenstiel B Baldwin J Stange-Thomann N Zody MC Linton L Lander ES Altshuler D A map of human genome sequence variation containing 1.42 million single nucleotide polymorphisms Nature 2001 409 928 933 11237013 10.1038/35057149
Reich DE Schaffner SF Daly MJ McVean G Mullikin JC Higgins JM Richter DJ Lander ES Altshuler D Human genome sequence variation and the influence of gene history, mutation and recombination Nat Genet 2002 32 135 142 12161752 10.1038/ng947
Stringer C Modern human origins: progress and prospects Philos Trans R Soc Lond B Biol Sci 2002 357 563 579 12028792 10.1098/rstb.2001.1057
Iafrate AJ Feuk L Rivera MN Listewnik ML Donahoe PK Qi Y Scherer SW Lee C Detection of large-scale variation in the human genome Nat Genet 2004 36 949 951 15286789 10.1038/ng1416
Carter NP As normal as normal can be? Nat Genet 2004 36 931 932 15340426 10.1038/ng0904-931
Yang Z Maximum-likelihood estimation of phylogeny from DNA sequences when substitution rates differ over sites Mol Biol Evol 1993 10 1396 1401 8277861
Yang Z PAML: a program package for phylogenetic analysis by maximum likelihood Comput Appl Biosci 1997 13 555 556 9367129
Yang Z Bielawski JP Statistical methods for detecting molecular adaptation Trends Ecol Evol 2000 15 496 503 11114436 10.1016/S0169-5347(00)01994-7
Massingham T Goldman N Detecting amino acid sites under positive selection and purifying selection Genetics 2005 169 1753 1762 15654091 10.1534/genetics.104.032144
Nei M Gu X Sitnikova T Evolution by the birth-and-death process in multigene families of the vertebrate immune system Proc Natl Acad Sci U S A 1997 94 7799 7806 9223266 10.1073/pnas.94.15.7799
Liao D Concerted evolution: molecular mechanism and biological implications Am J Hum Genet 1999 64 24 30 9915939 10.1086/302221
Goriely A McVean GA Rojmyr M Ingemarsson B Wilkie AO Evidence for selective advantage of pathogenic FGFR2 mutations in the male germ line Science 2003 301 643 646 12893942 10.1126/science.1085710
Evans PD Anderson JR Vallender EJ Gilbert SL Malcom CM Dorus S Lahn BT Adaptive evolution of ASPM, a major determinant of cerebral cortical size in humans Hum Mol Genet 2004 13 489 494 14722158 10.1093/hmg/ddh055
Casal J Gonzalez C Wandosell F Avila J Ripoll P Abnormal meiotic spindles cause a cascade of defects during spermatogenesis in asp males of Drosophila Development 1990 108 251 260 2112454
Riparbelli MG Massarelli C Robbins LG Callaini G The abnormal spindle protein is required for germ cell mitosis and oocyte differentiation during Drosophila oogenesis Exp Cell Res 2004 298 96 106 15242765 10.1016/j.yexcr.2004.03.054
Thompson JD Higgins DG Gibson TJ CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice Nucleic Acids Res 1994 22 4673 4680 7984417
Birney E Andrews D Bevan P Caccamo M Cameron G Chen Y Clarke L Coates G Cox T Cuff J Curwen V Cutts T Down T Durbin R Eyras E Fernandez-Suarez XM Gane P Gibbins B Gilbert J Hammond M Hotz H Iyer V Kahari A Jekosch K Kasprzyk A Keefe D Keenan S Lehvaslaiho H McVicker G Melsopp C Meidl P Mongin E Pettett R Potter S Proctor G Rae M Searle S Slater G Smedley D Smith J Spooner W Stabenau A Stalker J Storey R Ureta-Vidal A Woodwark C Clamp M Hubbard T Ensembl 2004 Nucleic Acids Res 2004 32 D468 70 14681459 10.1093/nar/gkh038
Eddy SR Hidden Markov models Curr Opin Struct Biol 1996 6 361 365 8804822 10.1016/S0959-440X(96)80056-X
Birney E Durbin R Using GeneWise in the Drosophila annotation experiment Genome Res 2000 10 547 548 10779496 10.1101/gr.10.4.547
Morgenstern B DIALIGN 2: improvement of the segment-to-segment approach to multiple sequence alignment Bioinformatics 1999 15 211 218 10222408 10.1093/bioinformatics/15.3.211
Yang Z Nielsen R Hasegawa M Models of amino acid substitution and applications to mitochondrial protein evolution Mol Biol Evol 1998 15 1600 1611 9866196
Goldman N Yang Z A codon-based model of nucleotide substitution for protein-coding DNA sequences Mol Biol Evol 1994 11 725 736 7968486
Nielsen R Yang Z Likelihood models for detecting positively selected amino acid sites and applications to the HIV-1 envelope gene Genetics 1998 148 929 936 9539414
Tamura K Nei M Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees Mol Biol Evol 1993 10 512 526 8336541
Xia X Xie Z DAMBE: software package for data analysis in molecular biology and evolution J Hered 2001 92 371 373 11535656 10.1093/jhered/92.4.371
Letunic I Copley RR Schmidt S Ciccarelli FD Doerks T Schultz J Ponting CP Bork P SMART 4.0: towards genomic data integration Nucleic Acids Res 2004 32 D142 4 14681379 10.1093/nar/gkh088
Guex N Peitsch MC SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modeling Electrophoresis 1997 18 2714 2723 9504803 10.1002/elps.1150181505
Felsenstein J PHYLIP (Phylogeny Inference Package) version 3.6 2004
Page RD TreeView: an application to display phylogenetic trees on personal computers Comput Appl Biosci 1996 12 357 358 8902363
Database of Genomic Variants
Sonnhammer EL Durbin R A dot-matrix program with dynamic threshold control suited for genomic DNA and protein sequence analysis Gene 1995 167 GC1 10 8566757 10.1016/0378-1119(95)00714-8
|
16159394
|
PMC1262708
|
CC BY
|
2021-01-04 16:32:48
|
no
|
BMC Genomics. 2005 Sep 13; 6:120
|
utf-8
|
BMC Genomics
| 2,005 |
10.1186/1471-2164-6-120
|
oa_comm
|
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-1291617152010.1186/1471-2164-6-129Research ArticleDiversity in domain architectures of Ser/Thr kinases and their homologues in prokaryotes Krupa A [email protected] N [email protected] Molecular Biophysics Unit, Indian Institute of Science, Bangalore 560 012, India2 Cell Cycle Control Laboratory, London Research Institute, Cancer Research – UK, South Mimms, Hertfordshire, EN6 3LD UK2005 19 9 2005 6 129 129 2 10 2004 19 9 2005 Copyright © 2005 Krupa and Srinivasan; licensee BioMed Central Ltd.2005Krupa and Srinivasan; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Ser/Thr/Tyr kinases (STYKs) commonly found in eukaryotes have been recently reported in many bacterial species. Recent studies elucidating their cellular functions have established their roles in bacterial growth and development. However functions of a large number of bacterial STYKs still remain elusive. The organisation of domains in a large dataset of bacterial STYKs has been investigated here in order to recognise variety in domain combinations which determine functions of bacterial STYKs.
Results
Using sensitive sequence and profile search methods, domain organisation of over 600 STYKs from 125 prokaryotic genomes have been examined. Kinase catalytic domains of STYKs tethered to a wide range of enzymatic domains such as phosphatases, HSP70, peptidyl prolyl isomerases, pectin esterases and glycoproteases have been identified. Such distinct preferences for domain combinations are not known to be present in either the Histidine kinase or the eukaryotic STYK families. Domain organisation of STYKs specific to certain groups of bacteria has also been noted in the current anlaysis. For example, Hydrophobin like domains in Mycobacterial STYK and penicillin binding domains in few STYKs of Gram-positive organisms and FHA domains in cyanobacterial STYKs. Homologues of characterised substrates of prokaryotic STYKs have also been identified.
Conclusion
The domains and domain architectures of most of the bacterial STYKs identified are very different from the known domain organisation in STYKs of eukaryotes. This observation highlights distinct biological roles of bacterial STYKs compared to eukaryotic STYKs. Bacterial STYKs reveal high diversity in domain organisation. Some of the modular organisations conserved across diverse bacterial species suggests their central role in bacterial physiology. Unique domain architectures of few other groups of STYKs reveal recruitment of functions specific to the species.
==== Body
Background
Extracellular signals perceived by the cells invoke an appropriate intracellular response triggering cascades of events that alter the activities of various signalling molecules. The variations in functional states of the proteins are brought about by covalent and non-covalent molecular interactions. Intracellular signalling in prokaryotes and eukaryotes is often achieved by phosphorylation of target proteins. The phosphorylated proteins undergo changes in their three-dimensional (3-D) shapes resulting in altered functional states. Phosphorylation events also lead to changes in the electrostatic interactions and ionisation states of residues at the catalytic site as exemplified by regulation of isocitrate dehydrogenases and HMG-CoA reductase [1,2]. Covalent attachment of the phosphoryl groups to various proteins is catalysed by protein kinases. These enzymes transfer a phosphate group from adenosine tri phosphate (ATP) onto an acceptor amino acid in a substrate protein. The acceptor groups on amino acid residues could be alcoholic as in Ser and Thr, phenolic group as in tyrosine or basic amino acid such as histidine (Nπ or Nτ), arginine (guanidino group), and the ε-NH2 group of lysine, carboxyl groups of aspartate and glutamate or thiol group of cysteine [3,4]. In addition to nucleoside triphosphates as donors of the phosphate groups, certain class of phosphorylating enzymes makes use of either phospho-enzyme intermediates or low molecular weight metabolites such as phosphoenol pyruvate, acetyl phosphate or carbamoyl phosphates and polyphosphates as donors [5-7].
Despite the occurrences of various phosphorylating enzymes, cellular signalling mechanisms are dependent on a subset of the list of enzymes mentioned above, for intra-cellular communications. In eubacteria and archaea, two component signal transduction systems [8], also referred to as His-Asp phospho-relay systems, mediate adaptive responses to changes in environmental conditions. The two-component system is composed of a histidine kinase (HK) and a cognate regulatory protein referred to as a response regulator (RR). HKs consist of an ATP-binding kinase domain and the H-box domain carrying the histidine which is the site of phosphorylation. HKs have also been identified in eukaryotes, such as yeast, Neurospora and Arabidopsis thaliana, and their functional roles have been studied [9,10].
However, in eukaryotes, Ser/Thr and Tyr kinases play a key role in cellular signal transduction. The Ser/Thr and Tyr kinases differ completely from the HKs both in sequence and structure. The Ser/Thr and the Tyr kinase share a common catalytic core of about 270 residues made up of N-terminal lobe consisting of mainly β-strands and a C-terminal lobe composed predominantly of α-helices [11]. The eukaryotes also possess two other distinct classes of protein kinases, the myosin heavy chain kinase/EF-2 kinase [12] and the mitochondrial protein kinases that include the pyruvate dehydrogenase kinase and branched-chain alpha-ketoacid dehydrogenase kinase [13] that differ from the Ser/Thr kinases in sequence and structure. Protein phosphorylation is hence considered to have originated independently in prokaryotes and eukaryotes in order to meet their diverse cellular and environmental requirements.
However studies on the eukaryotic protein-Ser/Thr/Tyr kinase like sequences (STYKs) in the prokaryotes emerged with the identification of proteins phosphorylated on Ser/Thr and Tyr residues in bacteria [14]. Phosphorylation of Ser/Thr residues in archaeal proteins has also been reported for Sulfolobus acidocaldarius by earlier studies [15]. Subsequent identification and characterisation of genes encoding STYKs in Myxococcus xanthus [16] and Streptomyces species [17] confirmed the presence of homologues of eukaryotic protein kinases in bacteria. Following these reports of STYKs, and the release of large number of bacterial genome sequences, a number of groups have surveyed for the occurrences of STYKs in them [18-22]. Their surveys suggested the occurrences of significant number of STYKs in both archaea and eubacterial genomes and they have been further classified based on their sequence similarity to various subfamilies [19]. A previous study [20] proposed that the Ser/Thr/Tyr protein kinases in bacteria have bacterial origin. Kinase activity in the STYKs has been demonstrated in a few species. These studies have suggested the involvement of STYKs in secondary metabolism [17], morphology of the aerial hyphae in Streptomyces granticolor [23], regulation of fruiting body formation in M.xanthus [24]. STYKs in Pseudomonas aeruginosa and Yersinia pseudotuberculosis [25,26] are shown to be critical for progression of infection and expression of virulence factors during certain stages of infection. The role of STYKs in the repression of proline utilisation operon and degradation of proline in S.typhimurium has been suggested [27].
The recently reported 3-D crystal structures of Mycobacterial PKN-B [28,29] has revealed an overall similarity of their structure to the eukaryotic Ser/Thr protein kinases with variations in the nucleotide binding regions. The prokaryotic homologues also include aminoglycoside kinases which phosphoryate antibiotics, and share remote sequence similarity with the STYKs. The 3-D fold of their catalytic domain is highly similar to the structures of eukaryotic STYKs [30]. Studies on the evolution of STYKs in Synechocystis sp. PCC 6803 have proposed the origin of STYKs before the divergence between prokaryotes and eukaryotes during evolution [20]. The conservation of these sequences in significant number of bacterial genomes despite limitation on their genomic size during evolution have therefore suggested their functions are essential in regulation of cellular activities in prokaryotes, as in their eukaryotic counterparts.
The present analysis is focussed on understanding the biological roles of the prokaryotic STYKs which have been shown to be encoded by significant number of genes in a large number of bacterial genomes by various groups. The present study is based on representative domain architectures of STYKs identified from the completely sequenced genomes of 125 prokaryotes which includes 35 genomes compiled in our early version of KinG database [31]. The domain architectures of STYKs and their homologues in 90 additional genomes included in this paper are recently added to the KinG database .
About 600 STYKs and their homologues have been identified in the current study in 125 bacterial genomes, of complete genomic data, using various sequence and profile search methods. Manual analysis of each of these STYKs and homologues has been made in order to check for the presence or absence of residues known to be critical for the activity of eukaryotic protein kinases. Using profile based search method (See Materials and Methods) the extent of similarity of bacterial homologues with different subfamilies of eukaryotic protein kinases has been analysed. The analysis suggests that the bacterial homologues form a distinct class of protein kinases as indicated by the highest similarity (20–30% sequence identity) of their catalytic kinase domain with the generic kinase profiles rather than any individual subfamily profile of eukaryotic protein kinases. i.e., the percent sequence identities of bacterial kinases with specific sub-families of eukaryotic STYKs is lower.
During the course of current analysis we encountered lipopolysaccharide kinases (LPSKs) which are well known for their remote relationship with STYKs [32]. RIO1 class of STYKs have been previously identified in eukaryotes [33,34] and in archaea [19] and our earlier work reported the occurrence of RIO1 proteins in all the three major taxons namely, Archaea, Bacteria and Eukaryota [32].
The domain organisation of STYKs identified from various genomes has been studied in an attempt to understand their probable biological roles. Groups of STYKs conserved only among specific classes of bacteria have been identified. Diverse domain arrangements observed in bacterial STYKs are radically different from those observed in eukaryotic Ser/Thr or Tyr kinases and prokaryotic HKs. These domain arrangements unique to bacterial STYKs are therefore suggestive of their functional roles distinct from that of other well-known families of protein kinases. Consideration of known functions of domains tethered to catalytic kinase domain of bacterial STYKs enabled obtaining clues about gross functional roles of various bacterial STYKs. Involvement of STYKs in stress response, protein folding, extensive protein-protein interactions and sugar transport has been suggested by covalent tethering of catalytic kinase domains of STYKs with other enzyme and non-enzyme modules. Homologues of bacterial PKs that have been previously well characterised in terms of biochemical activity and function have also been identified in the study. Putative substrates for a small set of bacterial STYKs have also been identified based on their similarity to bacterial kinases with well characterised substrates.
Results
The STYK-like sequences have been identified in various completely sequenced bacterial genomes and have been analysed for the occurrence of other functional domains. The domains have been identified using different profile search methods with stringent search parameters as described in the section on Materials and Methods and have also been manually checked for the conservation of key motifs. It should be noted that we have used stringent e-value cut-off of 10-8 and 0.01 respectively in the PSSM (Position Specific Scoring Matrix) matching and HMM (Hidden Markov Model) matching procedures. All the functional domain assignments discussed below are detected within the e-value limit of 0.01 by HMMER and are ensured, manually, that there is no obvious erroneous domain assignment.
Distribution of the STYKs and STYK-like sequences in bacterial genomes
The data set of bacterial genomes considered in the current analysis includes predicted ORFs from the genomes of sixteen archaeal and 109 bacterial genomes representing major division of these prokaryotic phylogenetic domains. A list of these completed genomes and number of Protein Kinase-like sequences (PKLS) that includes STYKs, RIO1 and ABC1-like kinases is given in Table 1. Table 1 also includes information on number of protein kinase-like sequences lacking the catalytic base residue (aspartate) and hence are likely to be non-functional as kinase. All the archaeal genomes analysed have at least one STYK, which belongs to the RIO1 family [19]. The RIO1s share a significant sequence similarity to the eukaryotic protein kinases and have recently been shown to possess a Ser/Thr autophosphorylation activity [35]. These RIO1 sequences have been initially observed in yeast [33,34] and subsequently detected in other eukaryotic genomes as well [36]. We have previously identified RIO1 like sequences in two of the eubacterial species, Deinococcus radiodurans and Pseudomonas aeruginosa [32].
Occurrence of RIO1 in many other eubacterial genomes has been reported in this paper for the first time. RIO1 sequences have been identified in other bacterial species including Bradyrhizobium japonicum, Pseudomonas putida, Pseudomonas syringae, Shewanella oneidensis and Yersinia pestis. The ubiquitous nature of RIO1 across archaea, bacteria and eukaryota is thus evident. Examination for the occurrences of HKs among the genome analysed reveal their absence in significant number of archaeal genomes, as well as in mycoplasma, Buchnera and Onion yellows phytoplasma genomes.
Domain arrangements in bacterial homologues of eukaryotic protein kinases
Interactions of multi-modular signalling proteins with cognate ligands are often mediated by the constituent modules. The domain organisation of the STYKs has been analysed extensively by various profile-search and fold recognition methods described in a later section. The details of the domain organisation of the bacterial kinases analysed in this paper have been included in the KinG database [31,37]. Incorporation of information about STYKs in 125 prokaryotic genomes in the KinG database represents a major update in the database compared to the previous version with STYKs only from 35 prokaryotic genomes. A comprehensive study on the domain organisation of the previously compiled repertoires of STYKs and those encoded in the additional genomes has been described here. Modular organisation of STYKs, newly identified in the current analysis have been listed in Table 2
Influence of the diverse domain composition of STYKs and their probable biological functions are discussed in the following sections.
Table 2 Representative modular organisation of bacterial STYKs identified in the current analysis.
Representative STYK Domain organisation as identified in the current analysis Bacterial species encoding STYKs of similar domain organisation
gi|9947763 Phosphatase-2C (PP2C) + Kinase Pseudomonas aeruginosa, Pseuomonas putida. Deinococcus radiodurans.
gi|32474490 Kinase + HSP70 Pirellula sp.
gi|31793271 Kinase + Hydrophobin Mycobacterium bovis
gi|3261694 Kinase + MalT-like ABC transporter domain Mycobacterium tuberculosis H37Rv
gi|41407147 Kinase + trans-membrane segment + peptidyl-prolyl cis-trans isomerase domain Mycobacterium avium sub sp. paratuberculosis
gi|23020879 Kinase + 6 TPR repeats + GGDEF Clostridium thermocellum
gi|17229994 Kinase + pectinesterase Nostoc sp. PCC 7120
gi|17134087 Kinase + GUN4 Nostoc sp. PCC 7120
gi|9945898 Kinase + von Willebrand Type A domain Pseudomonas aeruginosa
gi|17129689 Kinase + ANF Receptor Nostoc sp. PCC 7120
gi|21223285 Kinase + bacterial extracellular solute binding protein Streptomyces coelicolor
gi|1006577 Kinase + SH3b Synechocystis species
gi|17129893| FHA + Kinase Nostoc sp. PCC 7120, Thermodesmium erythraeum, Thermosynechococcuselongatus, Chloroflexus aurantia, Nostoc punctiforme.
gi|1709642 (PKN2) Kinase + Guanylate cyclase + Trans-membrane segment Myxococcus xanthus
Transmembrane bacterial Ser/Thr kinases
The amino acid sequences of kinase catalytic domains of bacterial STYKs are generally more similar to eukaryotic Ser/Thr kinases than the eukaryotic tyrosine kinases. The profiles of Ser/Thr kinase subfamilies match the bacterial kinases with better similarity (<1e-32) in contrast to the tyrosine kinases (>1e-17). However a significant number of bacterial STYKs contain trans-membrane spanning segments unlike the eukaryotic Ser/Thr kinases. The known rare exceptions of eukaryotic transmembrane protein with Ser/Thr kinase domain are TGF-beta receptor and the IRE1. Trans-membrane STYKs are absent in certain Gram-negative species Haemophilus influenzae, Rickettsiae, Helicobacter pylori, Chlamydiae and in the hyperthermophilic bacteria, Aquifex aeolicus. The Gram-positive species analysed in the data set have at least one trans-membrane STYKs in them. The presence of significant number of trans-membrane kinases (TM-kinases) in most bacterial species suggests many of them could play as receptor like kinases and are probably involved in direct interaction with the extra-cellular ligands. Further the TM kinases with extracellular PASTA (Fig 1i) domain such as PKNB [28,29] of Mycobacterium tuberculosis are encoded in diverse groups of Gram-positive bacteria analysed in the current study. The selective conservation of PASTA domains in STYKs encoded by Gram-positive bacteria implies their role in signalling pathways conserved across diverse Gram positive species. Currently the nature of ligands stimulating the trans-membrane STYKs is not known.
Figure 1 The representative domain combinations of the bacterial homologues of eukaryotic protein kinases are shown. The accession numbers of the protein domain families classified by PFAM are given below. (a) PP2c: phosphatase 2C [PF00481]; (b) Pectinesterase [PF01095]; (c) Glycoprotease [Peptidase_M22; PF00814]; (d) GAF[PF01590], PAS[PF00989], HisK[PF00512], HATPase[PF02518]. (e) FHA: Fork head associated domain [PF00498]; (f) WD-repeats [PF00040]; (g) and (h) amino acid binding domain [ANF_receptor; PF01094]; (i) peptidoglycan binding domain [PASTA, PF03973] k: trans-membrane segment [TM] shown in red.
Archaeal multi-domain STYKs
The multidomain STYKs of archaea include the kinases with O-sialoglycoprotein endopeptidases (OSGP) domain (Fig: 1c) present at the N-terminal of the kinase domain. The OSGP domains belong to the glycoprotease family (PFAM00814). The members of this family are known to cleave proteins that are heavily sialylated on serine or threonine residues. The OSGP containing STYKs described in earlier studies [19] are also encoded only in archaeal genomes. The archaeal species with the OSGP domains in their STYKs include Methanococcus janaschii, gi|2826367), Methanothermobacter thermoautotrophicus (gi|2622538), Halobacterium sp (gi|1058146), and Thermoplasma acidophilum (gi|10639497). OSGP domains of archaeal STYKs bear highest sequence similarity with the OSGP domains of the yeast gene product YK18 which is a putative glyco protease.
Phosphatase associated protein kinases
We report for the first time the tethering of the PP2C domains with STYK domain in bacteria. This unique class of STYKs has not been characterised so far. The PP2C like domains (Fig 1a) are tethered to the N-terminal of kinase domains in few STYKs encoded in Pseudomonas putida and Pseudomonas syringae. In addition to the well-characterised STYK of P.aeruginosa [38] a protein kinase (gi|9947763) with phosphatase-like domain described above has been identified in the current study. Earlier studies have biochemically characterised Stp1 of Pseudomonas aeruginosa, a single domain protein homologous to PP2C [38]. The regulation of Stp1 by its neighbouring gene encoding the protein kinase Stk1 has been suggested to be similar to the regulation of phosphatase and Ser/Thr kinases, YopH and YpkA of Y.pseudotuberculosis that serve as critical virulence determinants. Phosphatase like domain of Pseuodomonas species of the composite STYK shares 35% identity with Stp1 and about 30% identity with PP2C domains of Gram positive bacteria.
Protein kinases associated with nitrate disimilation
A group of STYKs (gi| 7190465, gi|4376414, gi|8978521, gi|7189538) specifically conserved in different species of Chlamydiae has a highly conserved C-terminal domain. A domain family previously referred to as NirV in Conserved Domain Database [39] has been assigned to this C-terminal region (E-58). This homologous domain family comprises of NirV like proteins, methylase involved in chemotaxis and sulphatase modifying enzymes and is referred to as DUF323 (Domain of unknown Function) in the PFAM database. The crystal structure of DUF323 has been recently solved [40] for a human homolog of Formylglycine generating enzyme (pFGE), which is a non-functional counter part of human FGE. FGE modifies the active site residue glycine of the sulphatases. The NirV domain of the Chlamydial kinases shares a sequence identity of 37% over a stretch of 100 residues with the NirV protein from Rhodobacter spaeroides. The NirV gene is situated in the operon containing two other genes namely, NirK and ppaZ whose gene products regulate nitrate dissimilation. The disruption of NirV gene in other organisms does not affect the nitrate reduction [41].
Protein kinases with domains homologous to modules associated with chaperones
We have identified a protein kinase (gi|32474490) with C-terminal domain homologous to HSP70 family of proteins in Pirellula species. The HSP70 family includes proteins that act as chaperones and assist the folding of a wide range of proteins.
Another class of STYKs that are also potentially involved in interactions with chaperonic proteins has been identified. STYKs with Tetratrico peptide repeats (TPR) have been identified in a number of bacterial species including M.tuberculosis, Clostridium acetobutylicum, Gloeobacter violaceus, Leptospira interrogans, Nostoc sp., Pirellula, Streptomyces avermitilis, Sulfolobus solfataricus and Sulfolobus tokodaii. TPR is known to form a series of anti parallel amphipathic α-helices that bundle together through hydrophobic interactions to form a fold with deep concave groove. They are suggested to bind a number of proteins of regulatory importance such as HSP90 based chaperone system [42].
Stress-responsive protein kinases
In the current analysis we have identified STYKs (gi| 13475495, gi|27353958 and gi|30180007) with universal stress induced protein associated (USPA) domain encoded in the genomes of Mesorhizobium loti, Bradyrhizobium japonicum and Nitrosomonas_europaea. Protein kinases with USPA domains have been recently identified in STYKs of Arabidopsis thaliana [43] and HisK of bacteria [44].
Protein kinases with protein – protein interaction modules
Protein kinases considered in the analysis have also revealed their association with domains made up of repeats that serve as an interface for the interaction of protein modules. An other class of STYKs with PQQ family repeats have been identified in Deinococcus radiodurans, Streptomyces avermilitis and Streptomyces coelicolor that have been described in the greater detail in later section with respect to their role in regulation of secondary metabolism. The PQQ repeats are known to form β-propeller like structures similar to WD – repeats that could serve as sites of interactions with target proteins.
STYK (gi|9948555) with leucine-rich repeats (LRR) have also been identified in Pseudomonas aeruginosa. LRRs are known to serve as protein interaction domains that form non-globular structures made up repeated ba subunits [45].
NHL repeats are also identified in two STYKs of M. tuberculosis (gi|2078052) and D.radiodurans (gi|6457717) which have 6 and 3 repeats following the kinase domain. They are known to serve as protein interaction surfaces and some times possess enzymatic activity as exemplified by the bifunctional peptidyl glycine α-amidylating monoxygenase protein.
Ubiquinone biosynthesis associated protein kinases
Significant number of archaeal and bacterial genomes encode STYKs with regions most similar to ubiB gene products flanking the catalytic kinase domain. Ubiquinone, is a critical component of the respiratory chain of most of organisms. The gene product of ubiB is involved in the production of an intermediate of Ubiquinone or the co-enzyme Q [46]. Based on the sequence comparisons, this group of kinases has been predicted to be involved in the regulation of ubiquinone biosynthesis [19]. This group of STYKs had been grouped into ABC1 family and was initially believed to have restricted occurrence in archaea [19]. The currently available genomic data reveals a wider representation of the ABC1 kinases in significant number of archael species.
They include Halobacterium salinarium, Thermoplasma acidophilum, Methanosarcina acetovirans, Methanosarcina mazei, Sulfolobus sulfotaricus, Sulfolobus tokadii, Thermoplasma volcanium All the ABC1 kinases encoded in archeal species analysed in the current study have an N-terminal ABC1 kinase domain and a putative trans-membrane spanning segments at the C_terminal. Thermoplasma acidophilum (gi|10639567) also has an unique ABC1 kinase with N-terminal RI01 like domain and C-terminal ABC1 Kinase-like domain followed by a putative trans-membrane spanning segment. ABC1 family of STYKs has a wide representation in large number of bacterial species. As described before Nitrosomonas europaea encodes a STYK of ABC1 family that has a N-terminal Usp domain.
Phosphorylation and structural components of bacterial cells
A set of closely related protein kinases is conserved in Mycobacterium bovis subsp. bovis AF2, Mycobacterium tuberculosis CDC1551 and Mycobacterium tuberculosis H37Rv. This new group of kinases identified in the current analysis has a N-terminal protein kinase domain and a C-terminal hydophobin like regions. The hydrophobins are low molecular weight cysteine rich proteins that form a critical component of the hydrophobic sheaths covering the fungal spores.
STYKs conserved among all the Gram-positive bacteria considered in the current analysis, have a C-terminal region that is highly similar to the C-terminal region of high molecular weight penicillin binding proteins. This domain is believed to play a role in the recognition of D-alanyl D-alanine dipeptides used to build up the peptidoglycan layers and is classified into the family called PASTA [47] in Pfam database due to their occurrence in bacterial Ser/Thr protein kinases and penicillin binding proteins.
Other domain combinations observed in bacterial protein kinases
A STYK found in the M.tuberculosis (gi|3261694) is comprised of an N-terminal kinase domain followed by C-terminal domain that shares high sequence similarity to distinct group of ABC transporters, MalT (E-value:10-180). MalT group of transporters are unique in terms of their direct participation of the transporter in transcriptional regulation of mal regulon [48]. These observations suggest the possible role of the STYK in the regulation of uptake of sugars. This domain has been observed only in M.tuberculosis STYK and is not found in the closely related species, M.leprae. This protein kinase is likely to be critical for nutrient uptake.
M.tuberculosis also contains a PK (gi|2131007) with a domain that shares significant similarity (e-value 10-4) with the periplasmic oxido reductase of thioredoxin family (DSBA). The thioredoxins are known for their interaction with a broad range of protein as a general protein disulphide oxido reductase. The potential role of this kinase for its possible involvement in disulphide isomerisation of various proteins during protein folding is therefore suggested. The occurrence of the DSBA domain in the extracellular region of the PK emphasises their role in reduction of periplasmic mycobacterial proteins. This is unique to Mycobacterium tuberculosis and is not found in M.leprae.
The genome of Mycobacterium avium subsp. paratuberculosis also encodes another STYK probably assisting the folding of proteins. The domain orgainisation of this novel kinase includes an N-terminal kinase domain, a central transmembrane domain and a C-terminal peptidyl-prolyl cis-trans isomerase (Pro_isomerase) domain. Among all available completely sequenced genomes of various species of Mycobacteria, this novel protein kinase is encoded in only in the M.avium genome.
Clostridium thermocellum ATCC 27405 encodes one of the largest STYK of 1826 amino acids encompassing a N-terminal kinase domain with six TPR repeats followed by the GGDEF domain at the C-terminal end. The domains of sensor kinases and response regulators of the His-Asp phospho relay systems have been studied previously [49]. The GGDEF domain is suggested and subsequently shown to be homologous to the catalytic domain of adenylyl cyclases [50,51]. Recent studies have also suggested their role in the regulation of bacterial growth and development through novel response regulators [52] by turnover of cyclic diguanosine monophosphate. Studies have also underlined the importance of adjacent sensor domains in the regulation of the catalytic activity of GGDEF [53]. The occurrence of GGDEF with STYKs implies the sharing of down stream signalling components by the HisKs and STYKs and their role in production of cyclicdiguanylate.
The Nostoc sp. PCC 7120 encodes a protein kinase (gi|17229994) with a pectinesterase domain (Fig 1b). The pectin esterases catalyse the hydrolysis of pectin, a component of extra cellular capsules and cell walls. The association of two such enzymatic activities is unique to this organism and suggests the probable role of phosphorylation in maintenance of integrity of bacterial cellular structural components.
A novel protein kinase (gi|17134087) with a C-terminal GUN4 domain has been identified in the Nostoc sp. PCC 7120. The GUN4 proteins in Arabidopsis thaliana are known to bind to magnesium-protoporphyrin IX, an intermediate in chlorophyll biosynthesis and there by regulating the enzyme Mg-chelatase involved in the biosynthesis [54]. The novel protein kinase is therefore suggested to be involved in the regulation of chlorophyll production in Nostoc species.
Pseudomonas aeruginosa, known to be an opportunistic pathogen encodes a total of 11 STYK-like sequences. Seven of these are STYKs and the rest are KDO kinases. We have identified the von Willebrand A domain located to the C-terminal of kinase domain in one of the STYKs that has been previously referred to as Ppka [55]. The glycoprotein binding vWA domain [56] might be playing an important role in its interaction with the host proteins as described in the later sections.
A STYK (gi|3328716) in Chlamydia species has a N-terminal kinase domain, and a C-terminal domain with similarity to TPR domain. Using remote homology detection methods as described in the 'Materials and Methods' section, the similarity of this region with the TPR domain of O-linked N-acetylglucosamine (O-GlcNAc) transferases [57] has been detected. The TPR domain of O-GlcNAc transferase (OGT) has been recently shown to be responsible for its targeting OGT and in determining its substrate specificity [58]. The domain homologous to TPR in the STYK is therefore likely to mediate its substrate interactions.
Ser/Thr kinases of bacteria with periplasmic protein domains
The periplasmic solute binding proteins are a part of the ATP binding cassette (ABC) transporter that transfers the amino acids and sugars into the cytoplasm. Ser/Thr kinases from Streptomyces coelicolor (gi|21223285) and Nostoc sp. 7120 (gi|17227840|) contain a solute binding domain for Glu, Gln (Fig 1g) and ANF receptor like domain (Fig 1h), which is involved in the binding of branched chain amino acids respectively.
Bacterial STYKs with non catalytic domains homologous to domains associated to eukaryotic PKs
This section describes the possible functions of a few bacterial STYKs inferred based on their similarity in domain combination with the well characterised eukaryotic protein kinases.
Synechocystis sp. is a cyanobacteria that encodes 12 STYKs. One of the STYK (gi|1006577) has a SH3b (bacterial SH3 domain) following the N-terminal kinase domain. Occurrences of SH3 domains in bacteria have been previously well documented [59]. They have also been classified into a different family in SMART database [60].
Protein kinases with WD repeats (Fig 1f) were identified in many bacterial species including Thermospora curvatum, Thermodesmium erythraeum, Chloroflexus aurantia, Nostoc punctiforme, Nostoc sp. PCC 7120 and Thermobifida fusca. Most of these kinases contain 7 WD repeats and are identified in cyanobacteria. The number of repeats varied in few from 4 to 14 and all the species mentioned above contain more than one such protein kinase with the exception of Thermobifida fusca. The WD repeats are associated with beta-propeller like folds and the most extensively studied are the Gβ-subunit of the heterotrimeric G-proteins in eukaryotes.
The FHA domain containing PKs (Fig: 1e) have been identified in the current analysis in the cyanobacterial species, Thermodesmium erythraeum (gi|23042923, gi|23043171, gi|23041283), Thermosynechococcus elongatus BP-1 (gi|22298671, gi|22299030|) Chloroflexus aurantia (gi|22970578), Nostoc punctiforme (gi|23127167), and Nostoc sp. PCC 7120 (gi|17232446, gi|17228044, gi|23043171). The phospho – threonine binding resiudes GR, SxxH and NG are well conserved in all these FHA domain containing protein kinases (Fig. 2). In eukaryotes the FHA containing protein kinases are conserved from yeast to mammals and are involved in the regulation of the various stages of cell cycle and development [61].
Figure 2 Multiple sequence alignment of FHA domains seen in bacterial Ser/Thr kinases with the conserved motifs involved in phospho-threonine binding, 'GR', 'SXXH', 'NG' shaded in yellow. gi|17228044|ref|NP_484592.1| (Nostoc sp. PCC 7120), gi|23127167|gb|ZP_00109042.1| (Nostocpunctiforme), gi|22299030|ref|NP_682277.1| (Thermosynechococcus elongatus BP-1), gi|23041283|gb|ZP_00072750.1| (Trichodesmium erythraeum IMS101), gi|22298671|ref|NP_681918.1| (Thermosynechococcus elongatus BP-1) gi|17232446|ref|NP_488994.1| (Nostoc sp. PCC 7120) gi|23043171|gb|ZP_00074493.1| (Trichodesmium erythraeum IMS101) gi|23042923|gb|ZP_00074272.1| (Trichodesmium erythraeum IMS101).
The Myxococcus xanthus protein kinase, PKN2 has an interesting domain organisation. The PK consists of an N-terminal kinase domain that extends as a stretch of Ala/Ser/Pro/Thr rich segments which further continues as a trans membrane segment as described in earlier studies [62,63]. We have identified a Adenylyl/ Guanylyl cyclase (AC/GC) domain in the region between the Ala/Ser/Pro/Thr rich regions and the trans membrane (TM)- region. Regions homologous to cyclases have been also observed in previous studies [64,65]. Cyclasess of this nature also occur in the eukaryotes. A stretch of nearly 200 residues with no similarity to any extra cellular domains known so far lies on the external side of the TM segment. However the protein kinase like domain (PKLD) in the eukaryotic GCs lack the catalytic residues. In the case of PKN2 the PK domain has all the catalytic residues and is also known to phosphorylate on Ser/Thr residues [62]. The conservation of catalytic residues of cyclase domain and the topology of the gene product has been analysed in the current study. The key residues at the catalytic sites of the GC, Glu and Cys confer nucleotide specificity to the enzyme. Mutation of Glu and Cys to Lys and Asp, the corresponding residues of the adenylyl cylases, changes the nucleotide specificity of GC to adenylate [66]. Among the residues known to confer the nucleotide specificity of the GCs the glutamate residue is conserved (Fig. 3) while the other residue involved in specific nucleotide binding, Cys, is substituted by Thr which is found in equivalent positions in few the adenylate cyclases. The nucleotide specificity of the GC-like domain of PKN2 sharing less than 15% identity with receptor GC is unclear. Trans membrane region of PKN2 follows the GC domain unlike eukaryotic transmembrane GCs.
Figure 3 Multiple sequence alignment of the human guanylate cyclase, rat adenylate cyclase and GC-like domain of PKN2 of Myxococcus xanthus is shown. The residues conferring nucleotide specificity are shaded yellow. Conserved hydrophobic and charged residues are shaded green and grey respectively.
A wide representation of eukaryotic signalling domains in the bacterial kingdom has been described in previous studies [67-69] and is also manifested in the bacterial Ser/Thr kinase family. Such domain recruitments across diverse signalling protein families imply the versatility in the nature of interactions of these domains.
Dual protein kinases in bacteria
It has been observed in the yeast, an unicellular eukaryote, that the osmo – regulation is brought about by both HisK and mitogen activated protein kinases (MAPKs) suggesting the signal transduction from HisK to the MAPKs to elicit the appropriate response [70,71]. The occurrences of such composite signalling pathways have been suggested in bacteria from other very early studies [72] and in studies on STYKs in Anabaena sp. PCC 7120 [72]. The STYKs with HisK, HATPase (the catalytic domain of HisKs) and GAF (regulatory domains found in phosphodiesterases, adenylate cyclases and pytochromes) have been referred to as dual protein kinases (Fig 1d) as they contain catalytic domain of two distinct kinases fused to form a single gene product. Hybrid kinases such as those described above have so far been identified in cyanobacterial species [64,65,73,74]. The cyanobacterial species Nostoc sp. PCC 7120 and Nostoc punctiforme have 12 and 11 dual protein kinases respectively. In the current study, the occurrences of such composite kinases have been noted in 5 different species. The other organisms that encode similar STYKs include Trichodesmium erythraeum, Leptospira interrogans, Ralstonia metallidurans and Rhodopsuedomonas plustris and Psuedomonas putida. The nature of signals these composite kinases perceive and the cellular processes they influence is intriguing. Some of these STYKs also have the PAS domains in the same polypeptide chain which is likely to serve as redox sensors and thereby suggesting the roles of such dual protein kinases in signalling pathways sensitive to redox potentials.
Protein kinases involved in bacterial cell division
Protein kinases encoded in the genomes of Bacillus subtilis and Bacillus halodurans, yabT (gi|2632333gnlPIDe1181999) and BH0080 (gi|10172692dbjBAB03799.1) respectively, are located in the same gene clusters as spoIIE and FtsH/MesJ. These STYKs are therefore suggested to be involved in the control of cell division like other genes in the same cluster. Among the many genes known so far to control the cell division in bacteria are the spoIIE and the FtsH/MesJ genes [75]. This group of STYK that are likely to be associated with cell division are analogous in their broader function to cyclin dependent kinases of the eukaryotes that control cell cycle.
Bacterial protein kinases and their substrates
Despite the identification of bacterial STYKs by genome-wide analysis and demonstration of in vitro phosphorylation activity on exogenous substrates of significant number of STYKs, very limited information is available regarding their natural substrates and their specificity.
Putative homologues of AfsK, a well characterised Ser/Thr kinase S.coelicolor, has been identified in Deinococcus radiodurans and Streptomyces avermitilis. In S. coelicolor, AfsK phosphorylates of a global regulator of secondary metabolism AfsR that also controls the production of Actinorhodin [17,76,77]. This is brought about by the Ser/Thr kinase AfsK. AfsK, encoded in the genomes of Streptomyces coelicolor and S. lividans is a membrane associated phosphokinase encoded by the afsK gene. The analysis of the domain composition of AfsK revealed the presence of nine bac_PQQ repeats located at the C-terminal of the protein.
A search for the AfsR homologues in Deinococcus radiodurans that are likely to be the putative targets of Afsk revealed a gene product (gi|15807540) with N-terminal transcription regulator domain (trans_reg) and Bacterial trans-activation domain (BTAD) similar to the domains observed in AfSR. This observation suggests that the gene product could probably be the target of the Afsk like kinase (gi|6460339) of D.radiodurans. Trans_reg or the BTAD domain conserved among all the transcription regulators of AFSR family are therefore likely to harbour sites for phosphorylation by AfsK.
The STYK sharing highest sequence similarity with AfsK in the catalytic domain (38% identity) in the M.tuberculosis genome is PKNB. A search for AfsR like gene products in the M.tuberculosis genome revealed the gene product of embR as its closest homologue. A further sequence analysis of embR gene product is known to be involved in the resistance to etambutol revealed the presence of BTAD, transactivation domain. This common domain shared by the AfsR and its homologue in M.tuberculosis and D.radiodurans further suggests that the BTAD, transactivation domain is critical for regulation of AfsR like gene products and as the most likely region of the phosphorylation mediated control of its activity. Recently embR gene product has been demonstrated to be phosphorylated by PKN-H [78].
Substrate identification studies on PKN2 of Myxococcus xanthus have revealed histone like proteins (HUα and HUβ)[63] and beta-lactamase [79] as the target proteins phosphorylated by the recombinant PKN2 expressed in E.coli. The phosphorylation of beta-lactamase inhibits its secretion [79] while the phosphorylation of HUα and HUβ+ [63] has been shown to prevent their binding to DNA resulting in toxicity to cells. PKN4 has been further been shown to phosphorylate phosphofructo kinase [80,81].
Phosphorylation as a mode of regulation of bacterial STYKs
Among the completed genomes analysed in this study, a significant number of STYKs are 'RD' protein kinases. The protein kinases in the eukaryotes are in turn regulated in a number of ways. Phosphorylation of Ser/Thr residues in the activation segment of the catalytic domains of these kinases is a common mode of regulation [82]. The PKs that undergo phosphorylation in the activation segment have a characteristic 'RD' doublet sequence motif in the catalytic loop wherein the Arg helps in neutralisation of the negative charge on the phosphorylated residue to suitably poise the catalytic aspartate for phosphorylation. These protein kinases are referred to as 'RD' kinases. A multiple sequence alignment of the putative activation segment in the catalytic domains of the bacterial 'RD' kinases have revealed the strong conservation of threonines in their activation segments (Fig: 4) These threonines are therefore suggested to be the sites of phosphorylation. Most of these kinases have two threonines that are well conserved and hence multiple phosphorylation for activation of these kinases as exemplified by the PKN-B of M.tuberculosis [83] is suggested.
Figure 4 Multiple sequence alignment of the activation segment of 'RD' protein kinases in bacteria with the canonical 'DFG', 'APE' shaded in blue and conserved threonines, the potential autophosphorylation sites shaded yellow. gi2911096, gi1370255, gi2131011, gi3261596, gi2131007, gi2078052, (M.tuberculosis); gi13092426 gi13092427 gi13092968 (M.leprae); gi8978468 (C.pneumoniae); gi2633949 (B.subtilis); gi10175124 (B.halodurans); gi12724921 (Lactococcus lactis); gi13622698 (S.pyogenes); gi13701020 gi13701020 (S.aureus); gi6460339 gi6457717 gi6460744 gi6459630 (D.radiodurans); gi13472161 (M.melitoti); gi1652313, gi1006577 gi1653955 gi1653478 gi1652588 (Synechocystis.sp); gi3844698 (M.genitalium); gi1674285 (M.pneumoniae); gi9947763 (P.aeruginosa); gi6899183 (U.ureolyticum).
Discussion
The genome-wide analyses reveal the occurrences of a large number of STYKs in bacteria. However the available knowledge about their biological functions is limited.
The STYKs show a lower representation compared to HKs in most of the genomes analysed (Data not shown). HKs are encoded by higher number of genes in comparison with STYK encoding genes in the prokaryotes. Nearly two percent of the genomes of eukaryotic organisms encodes STYKs as suggested from the genome-wide analysis by various groups [34,84-87]. Difference in the relative abundance of kinases belonging to the two classes, namely HKs and STYKs in the eukaryotes and the prokaryotes probably arises from the selection of specific class of PKs during the course of evolution, as an adaptation to distinct cellular environment. Low occurrence of STYKs compared to HKs in most bacterial species with the exception of Pirellula (Rhodopirellula baltica SH1) suggests, in general, the predominant role of HKs in prokaryotes.
HisK in archaeal genomes have been suggested to have arisen from lateral gene transfer events explaining their limited representation [88,89]. With the exception of Buchnera, Onion phytoplasma and Thermococcus kodakaraensis, all the above mentioned genomes have at least one of the two kinds of protein kinases, namely HisK and STYK. STYKs have not been identified in genomes of the two spirochaetes, Treponema pallidum and Borrelia burgdorferi, Chlorobium tepidum, Helicobacter hepaticus, Thermotoga maritima and Xanthomonas species. The absence of either classes of protein kinases in Buchnera species, an obligate symbiont of the Aphids and Onion phytoplasma (intra-cellular plant pathogen) suggests the loss of such enzymes in the above species is likely to be an adaptation to the host-intracellular environment. Our analysis also identifies an anaerobic obligate heterotrophic thermophile, T.kodakaraensis as the only species with a non-symbiotic or non-parasitic habitat that lacks HisK and STYKs. Further studies on influence of phosphorylation on response to external stimuli will therefore be able to provide insight into the nature of enzymes and significance of phosphorylation in these organsims.
The number of STYKs in phylogenetic domains of archaea and bacteria increase with the growing size of bacterial genomic data. The assignment of probable biochemical functions and biological roles to large number of sequences are often dependent on the occurrence of homologous gene products in other phylogenetic domains despite the vast divergence between the species. Earlier studies have also indicated that the gross biological function and interaction network of multi-domain proteins are conserved by the preservation of domain composition and arrangement [90-92]. Therefore domain organisation of the bacterial STYKs have been studied here to get an insight into their putative biological roles in comparison with other proteins of similar domain organisation.
Enzymatic and non-enzymatic domains described in the earlier sections have been found covalently tethered to the kinase catalytic domain. Among the archaeal multi-domain STYKs, the OSGP-containing STYKs are widely represented. Knowledge of substrates of the OSGP-containing STYKs and their sites of phosphorylation is required to understand their role in regulation of functions of glycoproteins. The Ser/Thr residues serve as sites for both glycosylation and the phosphorylation [93,94]. In many instances the phosphorylation and glycosylation occurs at common sites in a competitive manner. The phosphorylation would therefore prevent the glycosylation of the proteins. OSGP domains are involved in the cleavage of those proteins at the sites containing glycosylated Ser/Thr residues. The phosphorylation of the target proteins by the associated kinase domains of the OSGP containing STYKs might therefore serve as a regulatory mechanism to inhibit the proteolysis of glycoproteins.
Eubacterial STYKs are associated with a large variety of signalling domains. Chase 2 domains containing STYKs identified by recent studies [64,65,73] are also found in other classes of trans membrane receptors including histidine kinases, diguanylate cyclases and methyl accepting proteins. Presence of widely occurring sensory domain in Ser/Thr kinases of Nostoc further emphasises on its role in the perception of some critical signals that are likely to trigger cellular pathways synergistically.
STYK of P.aeruginosa with phosphatase like domain has significant similarity with spoIIE protein family in the PFAM database. Recent studies [95,96] have shown the involvement of Ser/Thr kinase-Phosphatase pair, PrkC and PrpC in the sporulation of B.subtilis. Psuedomonas aeruginosa however encodes a composite kinase-phosphatase in the same gene. The phosphatase-2C (PP2C) domains are known to selectively dephosphorylate protein phosphorylated on Ser/Thr residues and their homologues have been previously identified and characterised in various bacteria [97-101]. The PP2C domains have been predicted to be functionally active in SpoIIE [102,103] of Bacillus subtilis, a Gram positive bacterial species. SpoIIE controls the activation of sporulation transcription factors and critical for the development of the bacteria.
The bacterial STYKs with TPR repeats identified here are suggested to be influenced by the chaperone like protein during their trafficking, folding or interactions with substrate proteins. The involvement of HSP90 in hormonal and growth factor signal transduction mediated by protein kinases through proteins containing TPR repeats is well characterised in eukaryotic cells [42]. Further phosphatase pp5 of Plasmodium falciparum with TPR domain is known to act as co-chaperone for hsp90 and it has been therefore suggested that protein dephosphorylation might play a regulatory role in protein folding [104,105]. Despite their high sequence divergence the TPR repeats and their constituting domain form a general interface of interactions with HSP90 like chaperones [104].
STYKs with stress response domains such as USPA have been identified. The USPA proteins are suggested to be autophosphorylated to switch on to active state. Subsequent to the induction of USPA proteins, many phosphorylated proteins have been detected [106]. The two-component system involved in nitrate assimilation is known to induce bacterial universal stress proteins of the nucleotide binding class [44]. Majority of these proteins are tightly regulated and expressed under various conditions of stress in the bacteria.
The USPA containing kinases of plants have been implicated in disease resistance signalling pathway based on their similarity to other protein kinases like Pto and Pto interacting proteins of tomatoes that are involved in disease resistance [43,107]. Based on the role of USPA containing proteins in stress response and their regulation by the some two component system, we suggest the bacterial STYKs with USPA domains are likely to be induced during stress response. In the eukaryotes, MAP kinases are also known to trigger stress induced signalling pathways. This group of bacterial kinases could therefore be functionally analogous to their eukaryotic counterparts that are induced during stress.
Protein kinase containing gene products with C-terminal pentapeptide repeats have been identified initially by an early study [73] in Nostoc sp. PCC 7120 (gi|17132363) and Synechocystis sp. PCC 7803 (gi|1653478). Nitrate assimilation and nitrogen fixation are well characterised in various species of Cyanobacteria. The key enzyme involved in this pathway, the nitrogenase, is highly sensitive to oxygen. Some multi-cellular cyanobacterial species like Nostoc sp. form specialised cells called heterocysts wherein the glycolipid layer prevents the diffusion of oxygen [77]. The accumulation of glycolipids into the heterocysts is brought about by a number of proteins. Some of these proteins have the pentapeptide repeats, which are known to be critical for the localisation of glycolipids [108].
PASTA domains conserved in a group of STYKs encoded by Gram positive bacteria are found in proteins comprising the divisome complex. The representation of this group of STYKs in all gram positive bacteria, including the Lactococcus lactis (gi|12724921), and Streptococcus pyogenes (gi|13622698), which contain only one STYK suggest their role to be critical in their cell division. The role of these protein kinases in cell division is reiterated with their genes clustering with rodA involved in the regulation of cell morphology during cell division. The phosphatase 2C encoding genes also lie in the same cluster supporting the role of reversible phophorylation during cell division.
Lipopolysaccharide kinases (LPSK) engaged in the phopshorylation of the lipopolysaccharides in the outer membrane of Gram-negative bacteria share remote similarity with eukaryotic STYKs [32]. They could be grouped into two classes based on their sites of phosphorylation into waaP gene products and Kdo kinases (KdoK) [109,110]. A remote similarity between STYKs and the two family of LPSKs has been recently reported by our group [32]. The similarity in the potential 3-D structure and catalytic residues of LPSKs with the eukaryotic protein kinases suggests analogous catalytic functions to the residues that are well conserved across the two distinct families of enzymes. The occurrence of Kdo kinase is so far restricted to pathogenic Gram-negative species and is also known to be influencing the virulence in H.influenzae [109]. Inferences drawn on the structure and the catalytic mechanism of LPSK based on their similarity to eukaryotic protein kinases can therefore serve as a guide to the design of inhibitors to the LPSKs in virulent organisms.
The genomic organisation of all STYKs encoded in M.tuberculosis H37Rv genome has been analysed in an earlier study [111]. The gene encoding the protein kinase PKN-G of M.tuberculosis is clustered with the genes, glnH for periplasmic Gln binding proteins. Therefore PKN-G is likely to be the functional homologue of the other Gln binding periplasmic Ser/Thr protein kinase of Streptomyces coelicolor (gi|21223285). These protein kinases therefore suggests the intake of amino acids in to the cell is also likely to be influenced by phosphorylation, in addition to the sugar transport in M.tuberculosis as described previously. Recent studies have established the involvement of PKN-G in the regulation of cellular levels of Gln/Glu [112,113] and in evading phagocytosis by the host cells.
Eukaryotic signalling domains have been observed in a large number of bacterial proteins including in the STYKs [67-69]. The cyanobacterial STYK with a SH3b domain has a transmembrane segment between the kinase domain and the SH3 domain. The nature of signals perceived by this trans-membrane STYK is not clear. These SH3 domains are distantly related to their eukaryotic counterparts. Eukaryotic Ser/Thr kinases with SH3 domains have been well studied which is exemplified by mitogen activated protein kinase kinase [114,115].
WD-repeats, a versatile protein-protein interaction module has also been observed bacterial protein kinases [64,73,116,117] and in human protein kinases [46]. Pkwa of Themospora curvata, a member of WD-repeat kinase family has been shown to be phosphorylated by exogenous protein kinase form S. granticolor [100].
The fork head associated (FHA) domain was initially found in transcription factors and has been recently identified as protein – protein interaction modules known to bind to phospho-threonine residues on target proteins. Hence the interactions of the FHA domains with their cognate proteins are largely influenced by phosphorylation. Many bacterial proteins have been recently identified with FHA domains [117]. Previously it has been observed that the genes encoding FHA containing proteins cluster along with Ser/Thr protein kinases and phosphatases in the M. tuberculosis genome [111]. Thus the identification of the PKs with FHA domains in cyanobacteria suggests synergistic actions of FHA domains containing proteins and enzymes involved in reversible phosphorylation as observed in eukaryotes.
Very few studies have shed light on the nature of substrates of bacterial kinase [17]. The AfsR, the substrate of AfsK has an N-terminal transcription regulator domain (trans_reg domain), trans-activation domain (bacterial trans-activation domain, BTAD) followed by a signalling domain NB-ARC and a C-terminal tetratrico peptide repeat region which is known to be involved in protein-protein interactions. The BTAD domain is shared by all the members DNRI/REDD/AFSR family of transcription regulators involved in secondary metabolism. Other proteins in Streptomyces coelicolor and related species that control actinorhodin production share the trans_reg domain and the BTAD domain. A similar domain organisation among the homologues of AFSR is intriguing and is likely to contain sites of phosphorylation.
The phosphorylation site of HUα has been identified as a threonine residue which is well conserved across HUα of various species. RXXRTGR is the sequence pattern conserved at the phosphorylation site (P-site). The occurrence of R at the P-1 site suggests PKN2 to be a 'basic directed' protein kinase like Protein Kinase A (PKA) that specifically phosphorylates substrates with consensus motif RRXS/TZ' (X is any amino acid and Z is any hydrophobic amino acid). Based on the residue preferences of PKN2, at known phosphporylation sites on substrates, bacterial STYKs sharing significant similarity with PKN2 have been examined, that are likely to seek similar residue determinants on their substrates. The catalytic domain of PKN2 shares highest similarity (40% identity) with PKNB of M.tuberculosis and its homologue in M.leprae. The high sequence similarity between PKN2 and PKNB is likely to reflect in similar substrate binding modes such as preferences for N-terminal basic amino acid residues as also observed for PKA.
PKN2 of Myxococcus xanthus is known to autophosphorylate in the region following the kinase domain. The GC domain that lies C-terminal to the kinase domain contains 10 threonines. Further experimental studies are therefore necessary to ascertain if these sites could be the natural phosphorylation sites targeted by their respective enzymes.
A significant number of bacterial STYKs have been biochemically characterised previously to elucidate their biological roles. STYKs of the PKN subfamily have been identified and well characterised in Myxococcus xanthus [79-81]Streptomycetes coelicolor [17], Strepyomycetes granticolor [86], Synechocystis sp [118-120], Yersinia tuberculosis [33], Mycobacterium tuberculosis [111,122] and Psuedomonas aeruginosa [55]. The PKN representatives of Myxococcus xanthus and Streptomyces species are involved in the specialised developmental cycles characteristic to these species. They are also shown to be capable of autophosphorylation although it has been suggested to be unimportant for their kinase activity in vitro [17,79-82].
The YpkA of Yersinia pseudotuberculosis is a secreted STYK that has been previously shown to interact with small GTPases like RhoA and Rac1 of the host [26]. We have identified the homologue of YpkA, the only STYK in the completely sequenced genome of Yersinia pestis CO92 sharing 99% sequence identity with YpkA. However the genome of Yersinia pestis KIM does not encode any YpkA homologues. In addition to ABC1 and RIO1 class of STYKs, a single gene product containing only the kinase domain is identified.
PpkA of Pseudomonas aeruginosa have been implicated in their virulence and are suggested to be induced during the process of infection [55]. An unusual domain arrangement of PpkA identified in the current analysis comprises of an N-terminal kinase domain followed by a stretch of 300 residues of unknown function and a C-terminal von Willebrand factor Type A (vWA) domain. The vWA domain, known to bind to glycoproteins and growth factor receptors, in PpkA is therefore suggested to aid in interactions at the surface of the host cell.
Mbk, is a Ser/Thr kinase from Mycobacterium tuberculosis encoded by the gene mbk [121] that lies in pst operon that controls phosphate transport. Mbk corresponds to STYK gi|2078052 in M.tuberculosis genome, which has an N-terminal kinase domain, a transmembrane region and a C-terminal NHL-repeats. The occurrence of membrane spanning region and extra-cellular NHL-repeats that forms an interaction surface suggests the Mbk to serve as receptors for unknown ligands regulating the phosphate transport.
The STYK, referred to as SpkA in Synechocystis is previously shown to be required for the motility of the cyanobacteria. SpkA corresponds to STYK (gi|1652312), in the genome of Synechocystis sp that has an N-terminal kinase domain and a C-terminal extension of 200 residues to which no known functional domain could be associated. The role of the C-terminal extension in the regulation or interaction with other proteins controlling the motility of the organism remains to be elucidated
The attempt made in the current analysis to study the bacterial STYKs by In silico analysis is hoped to aid in the further understanding of STYKs encoded in various genomes. The diversity of the functional domains occurring in these protein kinase like gene products provide clues to their biological functions that are yet to be explored experimentally. The involvement of bacterial STYKs in various cellular processes described in the previous sections suggests a very little overlap with the Histidine kinases and the protein kinases of the eukaryotes.
Large-scale genomic analysis are fraught with uncertainties related to the inheritance of function among remotely homologous proteins. Nevertheless search procedures employed in the current study for the study of domain organisation of gene products has been extensively bench marked [122] An accuracy level of 96% is achieved by use of stringent search parameters described in 'Materials and Methods' section. The functional roles of STYKs and related systems encoded in various prokaryotic organisms deduced using computational approaches in the present analysis and in a recent independent analysis [123] is hoped to provide useful leads for further studies on their regulation.
Experimental studies [eg., [124]] elucidating the mechanisms of action STYKs and their homologues in various cellular processes as suggested from the analysis would therefore have implications in understanding their influence on the growth and development of the bacteria.
Methods
The complete set of predicted protein sequences from the ORFs of the bacterial genomes has been obtained from NCBI [125] Using sensitive sequence profile matching algorithms, STYK like sequences have been identified in the genomes.
We have employed multiple sensitive sequence search and analysis methods PSI-BLAST [126], IMPALA [127] and HMMer which matches Hidden Markov Models (HMMs) [128]. These programs have been previously benchmarked [122,129,130] and we have used a stringent cutoff for e-values (0.0005 in PSI_BLAST, 10-5 in IMPALA and 0.01 for HMMer) for identifying homologues of kinases. The list of predicted STYKs used for further analysis has been arrived at after careful cross-referencing between the results of these methods as well as manual scrutiny for a variety of factors such as length of the kinase domains and presence of critical functional residues. Kinase catalytic domains of the PKLA were aligned using CLUSTALW. Manual decisions are taken considering the presence/absence of functional motifs and the lengths of putative kinase domains. Those kinase-like sequences without functional residues, such as an aspartate in the catalytic base position, are segregated separately and are not considered for the detailed analysis. Number of such sequences in various genomes is provided in Table 2.
Domain assignment to the non-catalytic regions of the kinase containing genes has been carried out using the HMM search methods by querying each of the kinase containing sequences against the 6190 protein family HMMs available in the Pfam database [131]. Trans-membrane segments were detected using TMHMM [132].
The 3-dimensional structures were superimposed using STAMP [133].
Table 1 Prokaryotes of complete genomic data analysed and the distribution of encoded STYKs and homologues. The list includes ABC1, RI01 and kinases that share significant sequence similarity with STYK family. Number in parenthesis indicates the sequences lacking catalytic Asp and hence are unlikely to function as a kinase.
Organism STYK ABC1 RIO1
ARCHAEA
Aeropyrum pernix 7 - 2
Archaeoglobus fulgidus 2 - 2
Halobacterium salinarium 5 (1) 1 2
Methanococcus janaschii 4 - 2
Methanopyrus kandleri 1 - 1
Methanosarcina acetivorans 5 2 2
Methanosarcina mazei Goe1 4 1 2
Methanothermobacter thermoautotrophius 3 1 1
Nanoarchaeum equitans 3 - 2
Pyrobaculum aerophilum 4 - 2
Pyrococcus abysii 4 - 2
Pyrococcus horikoshii 4 - 2
Sulfolobus solfataricus 7 1 2
Sulfolobus tokodaii 11 1 2
Thermoplasma acidophilum 3 1 1
Thermoplasma volcanium 4 1 2
BACTERIA
Agrobacterium tumefaciens str. C58 2 1 -
Aquifex aeolicus 2 - 1
Bacillus anthracis str. Ames 3 - -
Bacillus cereus ATCC 14579 6 2 -
Bacillus halodurans 4 1 -
Bacillus subtilis 3 - -
Bacteroides thetaiotaomicron VPI-5482 2 - -
Bifidobacterium longum NCC2705 7 1 -
Bordetella bronchiseptica 2 1 -
Bordetella parapertussis 2 1 -
Bordetella pertussis 2 1 -
Borrelia burgdorferi - - -
Bradyrhizobium japonicum 5 1 1
Brucella melitensis - - -
Brucella suis 2 2 -
Buchnera aphidicola - - -
Buchnera sp. APS - - -
Candidatus Blochmannia floridanus 1 1 -
Caulobacter crescentus 1 1 -
Chlamydia muridarium 1(1) - -
Chlamydophila pneumoniae J138 3 (1) - -
Chlamydia trachomatis 3 (1) - -
Chlamydophila caviae GPIC 3 (1) - -
Chlamydophila pneumoniae AR39 3 (1) - -
Chlorobium tepidum TLS - - -
Chromobacterium violaceum ATCC 12472 3 2 -
Clostridium acetobutylicum 3 1 -
Clostridium perfringens 2 1 -
Clostridium tetani E88 1 - -
Corynebacterium diphtheriae 4 - -
Corynebacterium efficiens YS-314 4 - -
Corynebacterium glutamicum ATCC 13032 4 - -
Coxiella burnetii RSA 493 4 1 -
Deinococcus radiodurans 10 - 1
Enterococcus faecalis V583 2 - -
Escherichia coli K12 4 1 -
Escherichia coli O157:H7 EDL933 4 1 -
Fusobacterium nucleatum subsp. nucleatum ATCC 25586 1 - -
Geobacter sulfurreducens PCA 1 - -
Gloeobacter violaceus[B]4.7NC_005125 19 (1) 4 -
Haemophilus ducreyi 2 1 -
Haemophilus influenzae 1 - -
Helicobacter hepaticus ATCC 51449 - - -
Helicobacter pylori J99 1 - -
Lactobacillus plantarum WCFS1 3 1 -
Lactococcus lactis 1 - -
Leptospira interrogans 5 (2) 2 -
Mesorhizobium meliloti 3 1 -
Mycobacterium avium subsp. paratuberculosis 12 2 -
Mycobacterium bovis 10 2 -
Mycobacterium leprae 6 2 -
Mycobacterium tuberculosis 13 2 -
Mycoplasma gallisepticum 1 - -
Mycoplasma genitalium 1 - -
Mycoplasma penetrans 1 - -
Mycoplasma pneumoniae 1 - -
Mycoplasma pulmonis 1 - -
Neisseria meningititis 1 1 -
Nitrosomonas europaea 2 2 -
Nostoc sp. PCC 7120 53 4 -
Oceanobacillus iheyensis 1 - -
Onion yellows phytoplasma - - -
Pasteurella multocida 2 1 -
Photorhabdus luminescens subsp. laumondii 1 1 -
Pirellula sp. 59 (1) 2 -
Porphyromonas gingivalis W83 1 - -
Prochlorococcus marinus MED4 3 3 -
Prochlorococcus marinus MIT 9313 4 3 -
Prochlorococcus marinus subsp. pastoris str. 3 3 -
Pseudomonas aeruginosa 11 1 1
Pseudomonas putida KT2440 14 (2) - 2
Pseudomonas syringae 22 2 2
Ralstonia solanacearum 3 1 -
Rhodopseudomonas palustris 5 (3) 1 -
Rickettsia conorii 1 1 -
Rickettsia prowazekii 1 1 -
Salmonella enterica subsp. enterica serovar Typhi Ty2 4 1 -
Salmonella typhimurium LT2 2 1 -
Shewanella oneidensis MR-1 8 (2) 1 1
Shigella flexneri 2a str. 2457T 3 1 -
Sinorhizobium meliloti 1 1 -
Staphylococcus aureus 2 - -
Staphylococcus epidermidis ATCC 12228 1 - -
Streptococcus mutans UA159 1 - -
Streptococcus pyogenes M1 GAS 1 - -
Streptococcus pyogenes SSI-1 1 - -
Streptomyces avermitilis 34 (2) 1 -
Streptomyces coelicolor A3 18 - -
Synechococcus sp. WH 8102 3 3 -
Synechocystis 12 (1) 5 -
Thermoanaerobacter tengcongensis 1 - -
Thermosynechococcus elongatus 14 - -
Thermotoga maritima 1 (1) 1 -
Treponema pallidum - - -
Tropheryma whipplei str Twist 4 (2) - -
Ureaplasma ureolyticum 1 - -
Vibrio cholerae 2 1 -
Vibrio parahaemolyticus RIMD 2210633 6 1 -
Vibrio vulnificus CMCP6 5 1 -
Wigglesworthia brevipalpis 1 1 -
Wolinella succinogenes 1 1 -
Xanthomonas axonopodis pv. citri str. 306 5 - -
Xanthomonas campestris pv. campestris str. ATCC 33913 1 - -
Xylella fastidiosa 3 1 1
Xylella fastidiosa Temecula1 3 1 1
Yersinia pestis 4 1 1
Authors' contributions
AK performed the computational sequence analysis and modelling. AK and NS conceived the study and participated in its design and coordination. Both the authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Data files comprising of the description of protein kinases and homologues encoded in genomes of organisims considered in the current analysis are provided as supplementary information accompanying this article. Each additional data file lists the gene identifiers, length, and domain arrangement of protein kinases and homologues identified in the current analysis.
Click here for file
Acknowledgements
Authors thank Dr. Galperin and anonymous referees for their critical comments and suggestions. A.K. is supported by a fellowship from the Council of Scientific and Industrial Research, India and Wellcome Trust, UK. This research is supported by International Senior Fellowship in biomedical sciences to N.S. by the Wellcome Trust, UK, National Bioscience award by the Department of Biotechnology, India and by the computational genomics project supported by the Department of Biotechnology, India.
==== Refs
Hurley JH Dean AM Thorsness PE Koshland DE JrStroud RM Regulation of isocitrate dehydrogenase by phosphorylation involves no long-range conformational change in the free enzyme J Biol Chem 1990 265 3599 3602 2406256
Kim DY Stauffacher CV Rodwell VW Engineering of Sulfolobus solfataricus HMG-CoA reductase to a form whose activity is regulated by phosphorylation and dephosphorylation Biochemistry 2000 39 2269 2275 10694393 10.1021/bi991749t
Hunter T Protein kinase classification Methods Enzymol 1991 200 3 37 1835513
Duclos B Marcandier S Cozzone AJ Chemical properties and separation of phosphoamino acids by thin-layer chromatography and/or electrophoresis Methods Enzymol 1991 201 10 21 1943759
Saier MH JrWu LF Reizer J Regulation of bacterial physiological processes by three types of protein phosphorylating systems Trends Biochem Sci 1990 15 391 395 2251731 10.1016/0968-0004(90)90238-7
Lukat GS McCleary WR Stock AM Stock JB Phosphorylation of bacterial response regulator proteins by low molecular weight phospho-donors Proc Natl Acad Sci U S A 1992 89 718 722 1731345
Skorko R Polyphosphate as a source of phosphoryl group in protein modification in the archaebacterium Sulfolobus acidocaldarius Biochimie 1989 71 1089 1093 2513000 10.1016/0300-9084(89)90115-6
Dutta R Qin L Inouye M Histidine kinases: diversity of domain organization Mol Microbiol 1999 34 633 640 10564504 10.1046/j.1365-2958.1999.01646.x
Thomason P Kay R Eukaryotic signal transduction via histidine-aspartate phosphorelay J Cell Sci 2000 113 3141 150 10954413
Wolanin PM Thomason PA Stock JB Histidine protein kinases: key signal transducers outside the animal kingdom Genome Biol 2002 3 3013 12372152 10.1186/gb-2002-3-10-reviews3013
Zheng J Knighton DR ten Eyck LF Karlsson R Xuong N Taylor SS Sowadski JM Crystal structure of the catalytic subunit of cAMP-dependent protein kinase complexed with MgATP and peptide inhibitor Biochemistry 1993 32 2154 2161 8443157 10.1021/bi00060a005
Ryazanov AG Ward MD Mendola CE Pavur KS Dorovkov MV Wiedmann M Erdjument-Bromage H Tempst P Parmer TG Prostko CR Germino FJ Hait WN Identification of a new class of protein kinases represented by eukaryotic elongation factor-2 kinase Proc Natl Acad Sci U S A 1997 94 4884 4889 9144159 10.1073/pnas.94.10.4884
Thomson M Evidence of undiscovered cell regulatory mechanisms: phosphoproteins and protein kinases in mitochondria Cell Mol Life Sci 2002 59 213 219 11915939 10.1007/s00018-002-8417-7
Cozzone AJ Protein phosphorylation in prokaryotes Annu Rev Microbiol 1988 42 97 125 2849375 10.1146/annurev.mi.42.100188.000525
Skorko R Protein phosphorylation in the archaebacterium Sulfolobus acidocaldarius Eur J Biochem 1984 145 617 622 6439562 10.1111/j.1432-1033.1984.tb08601.x
Munoz-Dorado J Inouye S Inouye M A gene encoding a protein serine/threonine kinase is required for normal development of M. xanthus, a gram-negative bacterium Cell 1991 67 995 1006 1835671 10.1016/0092-8674(91)90372-6
Matsumoto A Hong SK Ishizuka H Horinouchi S Beppu T Phosphorylation of the AfsR protein involved in secondary metabolism in Streptomyces species by a eukaryotic-type protein kinase Gene 1994 146 47 56 8063104 10.1016/0378-1119(94)90832-X
Shi L Potts M Kennelly PJ The serine, threonine, and/or tyrosine-specific protein kinases and protein phosphatases of prokaryotic organisms: a family portrait FEMS Microbiol Rev 1998 22 229 253 9862122 10.1016/S0168-6445(98)00015-1
Leonard CJ Aravind L Koonin EV Novel families of putative protein kinases in bacteria and archaea: evolution of the "eukaryotic" protein kinase superfamily Genome Res 1998 8 1038 1047 9799791
Han G Zheng CC On the origin of Ser/Thr kinases in a prokaryote FEMS Microbiol lett 2001 200 79 84 11410353 10.1016/S0378-1097(01)00206-3
Kennelly PJ Protein kinases and protein phosphatases in prokaryotes: a genomic perspective FEMS Microbiol Lett 2002 206 1 8 11786249 10.1016/S0378-1097(01)00479-7
Kenelly PJ Archaeal protein kinases and protein phosphatases: insights from genomics and biochemistry Biochem J 2003 370 373 389 12444920 10.1042/BJ20021547
Nadvornik R Vomastek T Janecek J Technikova Z Branny P Pkg2, a novel transmembrane protein Ser/Thr kinase of Streptomyces granaticolor J Bacteriol 1999 181 15 23 9864307
Udo H Inouye M Inouye S Effects of overexpression of Pkn2, a transmembrane protein serine/threonine kinase, on development of Myxococcus xanthus J Bacteriol 1996 178 6647 6649 8932326
Barz C Abahji TN Trulzsch K Heesemann J The Yersinia Ser/Thr protein kinase YpkA/YopO directly interacts with the small GTPases RhoA and Rac-1 FEBS Lett 2000 482 139 143 11018537 10.1016/S0014-5793(00)02045-7
Hakansson EE Forsberg SA Wolf-Watz H A secreted protein kinase of Yersinia pseudotuberculosis is an indispensable virulence determinant Nature 1993 361 730 732 8441468 10.1038/361730a0
Ostrovsky PC Maloy S Protein phosphorylation on serine, threonine, and tyrosine residues modulates membrane-protein interactions and transcriptional regulation in Salmonella typhimurium Genes Dev 1995 9 2034 2041 7544316
Ortiz-Lombardia M Pompeo F Boitel B Alzari PM Crystal structure of the catalytic domain of the PknB serine/threonine kinase from Mycobacterium tuberculosis J Biol Chem 2003 278 13094 13100 12551895 10.1074/jbc.M300660200
Young TA Delagoutte B Endrizzi JA Falick AM Alber T Structure of Mycobacterium tuberculosis PknB supports a universal activation mechanism for Ser/Thr protein kinases Nat Struct Biol 2003 10 168 174 12548283 10.1038/nsb897
Hon WC McKay GA Thompson PR Sweet RM Yang DS Wright GD Berghuis AM Structure of an enzyme required for aminoglycoside antibiotic resistance reveals homology to eukaryotic protein kinases Cell 1997 89 887 895 9200607 10.1016/S0092-8674(00)80274-3
Krupa A Abhinandan KR Srinivasan N Kin G: A database of protein kinases in genomes Nucl Acids Res 2004 32 D153 155 14681382 10.1093/nar/gkh019
Krupa A Srinivasan N Lipopolysaccharide phosphorylating enzymes encoded in the genomes of Gram-negative bacteria are related to the eukaryotic protein kinases Protein Sci 2002 11 1580 1584 12021457 10.1110/ps.3560102
Angermayr M Roidl A Bandlow W Yeast Rio1p is the founding member of a novel subfamily of protein serine kinases involved in the control of cell cycle progression Mol Microbiol 2002 44 309 324 11972772 10.1046/j.1365-2958.2002.02881.x
Hunter T Plowman GD The protein kinases of budding yeast: six score and more Trends Biochem Sci 1997 22 18 22 9020587 10.1016/S0968-0004(96)10068-2
Angermayr M Bandlow W RIO1, an extraordinary novel protein kinase FEBS Lett 2002 524 31 36 12135737 10.1016/S0014-5793(02)02993-9
Plowman GD Sudarsana S Bingham J Whyte D Hunter T The protein kinases of Caenorhabditis elegans: a model for signal transduction in multicellular organisms Proc Natl Acad Sci U S A 1999 96 13603 13610 10570119 10.1073/pnas.96.24.13603
KinG: Kinases in genomes
Mukhopadhyay S Kapatral V Xu W Chakrabarty AM Characterisation of a Hank's type serine/threonine kinase and ser/threonine phosphoprotein phosphatase in Pseudomonas aeruginosa J Bacteriol 1999 181 6615 6622 10542161
Marchler-Bauer A Anderson JB DeWeese-Scott C Fedorova ND Geer LY He S Hurwitz DI Jackson JD Jacobs AR Lanczycki CJ Liebert CA Liu C Madej T Marchler GH Mazumder R Nikolskaya AN Panchenko AR Rao BS Shoemaker BA Simonyan V Song JS Thiessen PA Vasudevan S Wang Y Yamashita RA Yin JJ Bryant SH CDD: a curated Entrez database of conserved domain alignments Nucleic Acids Res 2003 31 383 387 12520028 10.1093/nar/gkg087
Dickmanns A Schmidt B Rudolph MG Mariappan M Dierks T von Figura K Ficner R Crystal structure of human pFGE, the paralog of the Calphaformylglycine-generating enzyme J Biol Chem 2005 280 15180 15187 15687489 10.1074/jbc.M414317200
Jain R Shapleigh JP Characterization of nirV and a gene encoding a novel pseudoazurin in Rhodobacter sphaeroides 2.4.3 Microbiology 2001 147 2505 2515 11535790
Pratt WB The hsp90-based chaperone system: involvement in signal transduction from a variety of hormone and growth factor receptors Proc Soc Exp Biol Med 1998 217 420 434 9521088
Kerk D Bulgrien J Smith DW Gribskov M Arabidopsis proteins containing similarity to the universal stress protein domain of bacteria Plant Physiol 2003 131 1209 1219 12644671 10.1104/pp.102.016006
Kvint K Nachin L Diez A Nystrom T The bacterial universal stress protein: function and regulation Curr Opin Microbiol 2003 6 140 145 12732303 10.1016/S1369-5274(03)00025-0
Kobe B Deisenhofer J The leucine-rich repeat: a versatile binding motif Trends Biochem Sci 1994 19 415 421 7817399 10.1016/0968-0004(94)90090-6
Poon WW Davis DE Ha HT Jonassen T Rather PN Clarke CF Identification of Escherichia coli ubiB, a gene required for the first monooxygenase step in ubiquinone biosynthesis J Bacteriol 2000 182 5139 5146 10960098 10.1128/JB.182.18.5139-5146.2000
Yeats C Finn RD Bateman A The PASTA domain: a beta-lactam-binding domain Trends Biochem Sci 2002 27 438 12217513 10.1016/S0968-0004(02)02164-3
Boos W Shuman H Maltose/maltodextrin system of Escherichia coli: transport, metabolism, and regulation Microbiol Mol Biol Rev 1998 62 204 229 9529892
Galperin MY Nikolskaya AN Koonin EV Novel domains of the prokaryotic two-component signal transduction systems FEMS Microbiol Lett 2001 203 11 21 11557134 10.1016/S0378-1097(01)00326-3
Pei J Grishin NV GGDEF domain is homologous to adenylyl cyclase Proteins 2001 42 210 216 11119645 10.1002/1097-0134(20010201)42:2<210::AID-PROT80>3.0.CO;2-8
Chan C Paul R Samoray D Amiot NC Giese B Jenal U Schirmer T Structural basis of activity and allosteric control of diguanylate cyclase Proc Natl Acad Sci U S A 2004 101 17084 17089 15569936 10.1073/pnas.0406134101
Paul R Weiser S Amiot NC Chan C Schirmer T Giese B Jenal U Cell cycle-dependent dynamic localization of a bacterial response regulator with a novel di-guanylate cyclase output domain Genes Dev 2004 18 715 727 15075296 10.1101/gad.289504
Ryjenkov DA Tarutina M Moskvin OV Gomelsky M Cyclic diguanylate is a ubiquitous signaling molecule in bacteria: insights into biochemistry of the GGDEF protein domain J Bacteriol 2005 187 1792 1798 15716451 10.1128/JB.187.5.1792-1798.2005
Larkin RM Alonso JM Ecker JR Chory J GUN4, a regulator of chlorophyll synthesis and intracellular signaling Science 2003 299 902 906 12574634 10.1126/science.1079978
Motley ST Lory S Functional characterization of a serine/threonine protein kinase of Pseudomonas aeruginosa Infect Immun 1999 67 5386 5394 10496921
Whittaker CA Hynes RO Distribution and evolution of von Willebrand/integrin A domains: widely dispersed domains with roles in cell adhesion and elsewhere Mol Biol Cell 2002 13 3369 3387 12388743 10.1091/mbc.E02-05-0259
Jinek M Rehwinkel J Lazarus BD Izaurralde E Hanover JA Conti E The superhelical TPR-repeat domain of O-linked GlcNAc transferase exhibits structural similarities to importin alpha Nat Struct Mol Biol 2004 11 1001 1007 15361863 10.1038/nsmb833
Iyer SP Hart GW Roles of the tetratricopeptide repeat domain in O-GlcNAc transferase targeting and protein substrate specificity J Biol Chem 2003 278 24608 24616 12724313 10.1074/jbc.M300036200
Whisstock JC Lesk AM SH3 domains in prokaryotes Trends Biochem Sci 1999 24 132 133 10322416 10.1016/S0968-0004(99)01366-3
Schultz J Copley RR Doerks T Ponting CP Bork P SMART: a web-based tool for the study of genetically mobile domains Nucleic Acids Res 2000 28 231 234 10592234 10.1093/nar/28.1.231
Li J Williams BL Haire LF Goldberg M Wilker E Durocher D Yaffe MB Jackson SP Smerdon SJ Structural and functional versatility of the FHA domain in DNA-damage signalingby the tumor suppressor kinase Chk2 Mol Cell 2002 9 1045 1054 12049740 10.1016/S1097-2765(02)00527-0
Udo H Inouye M Inouye S Biochemical characterization of Pkn2, a protein Ser/Thr kinase from Myxococcus xanthus, a Gram-negative developmental bacterium FEBS Lett 1997 400 188 192 9001395 10.1016/S0014-5793(96)01384-1
Udo H Lam CK Mori S Inouye M Inouye S Identification of a substrate for Pkn2, a protein Ser/Thr kinase from Myxococcus xanthus by a novel method for substrate identification J Mol Microbiol Biotechnol 2000 2 557 563 11075932
Ohmori M Ikeuchi M Sato N Wolk P Kaneko T Ogawa T Kanehisa M Goto S Kawashima S Okamoto S Yoshimura H Katoh H Fujisaw T Ehira S Kamei A Yoshihara S Narikawa R Tabata Satoshi Characterisation of genes encoding Multi-domain proteins in the genome of the filamentous nitrogen-fixing cyanobacterium Anabena sp. Strain PCC 7120 DNA Res 2001 8 271 284 11858227 10.1093/dnares/8.6.271
Wang L Sun YP Chen WL Li JH Zhang CC Genomic analysis of protein phosphatases and two-component regulatory systems of the cyanobacterium Anabaena sp. Strain PCC 7120 FEMS Microbiol Lett 2002 217 155 165 12480098 10.1016/S0378-1097(02)01052-2
Tucker CL Woodcock SC Kelsell RE Ramamurthy V Hunt DM Hurley JB Biochemical analysis of a dimerization domain mutation in RetGC-1 associated with dominant cone-rod dystrophy Proc Natl Acad Sci U S A 1999 96 9039 9044 10430891 10.1073/pnas.96.16.9039
Bakal CJ Davies JE No longer an exclusive club: eukaryotic signalling domains in bacteria Trends Cell Biol 2000 10 32 38 10603474 10.1016/S0962-8924(99)01681-5
Zhulin IB Nikolskaya AN Galperin MY Common extracellular sensory domains in transmembrane receptors for diverse signal transduction pathways in bacteria and archaea J Bacteriol 2003 185 285 294 12486065 10.1128/JB.185.1.285-294.2003
Koonin EV Makarova KS Aravind L Horizontal gene transfer in prokaryotes: quantification and classification Annu Rev Microbiol 2002 55 709 742 11544372 10.1146/annurev.micro.55.1.709
Hughes DA Signal transduction. Histidine kinases hog the limelight Nature 1994 369 187 188 8183337 10.1038/369187a0
Maeda T Wurgler-Murphy SM Saito H A two-component system that regulates an osmosensing MAP kinase cascade in yeast Nature 1994 369 242 245 8183345 10.1038/369242a0
Phalip V Li JH Zhang CC HstK, a cyanobacterial protein with both a serine/threonine kinase domain and a histidine kinase domain: implication for the mechanism of signal transduction Biochem J 2001 360 639 644 11736654 10.1042/0264-6021:3600639
Zhang CC Gonzalez L Phalip V Survey, anlaysis and genetic organisation of genes encoding eukaryotic proteins on a cyanobacterial genome Nucl Acids Res 1998 26 3619 3625 9685474 10.1093/nar/26.16.3619
Zhang CC Bacterial signalling involving eukaryotic-type protein kinases Mol Microbiol 1996 20 9 15 8861199
Wehrl W Niederweis M Schumann W The FtsH protein accumulates at the septum of Bacillus subtilis during cell division and sporulation J Bacteriol 2000 182 3870 3873 10851010 10.1128/JB.182.13.3870-3873.2000
Umeyama T Horinouchi S Autophosphorylation of a bacterial serine/threonine kinase, AfsK, is inhibited by KbpA, an AfsK-binding protein J Bacteriol 2001 183 5506 5512 11544211 10.1128/JB.183.19.5506-5512.2001
Hong SK Matsumoto A Horinouchi S Beppu T Effects of protein kinase inhibitors on in vitro protein phosphorylation and cellular differentiation of Streptomyces griseus Mol Gen Genet 1993 236 347 354 8437579 10.1007/BF00277132
Molle V Kremer L Girard-Blanc C Besra GS Cozzone AJ Prost JF An FHA phosphoprotein recognition domain mediates protein EmbR phosphorylation by PknH, a Ser/Thr protein kinase from Mycobacterium tuberculosis Biochemistry 2003 42 15300 15309 14690440 10.1021/bi035150b
Udo H Munoz-Dorado J Inouye M Inouye S Myxococcus xanthus, a gram-negative bacterium, contains a transmembrane protein serine/threonine kinase that blocks the secretion of beta-lactamase by phosphorylation Genes Dev 1995 9 972 983 7774814
Nariya H Inouye S Modulating factors for the Pkn4 kinase cascade in regulating 6-phosphofructokinase in Myxococcus xanthus Mol Microbiol 2005 56 1314 1328 15882423 10.1111/j.1365-2958.2005.04619.x
Nariya H Inouye S Activation of 6-phosphofructokinase via phosphorylation by Pkn4, a proteinSer/Thr kinase of Myxococcus xanthus Mol Microbiol 2002 46 1353 1366 12453221 10.1046/j.1365-2958.2002.03251.x
Johnson LN Noble ME Owen DJ Active and inactive protein kinases: structural basis for regulation Cell 1996 85 149 158 8612268 10.1016/S0092-8674(00)81092-2
Boitel B Ortiz-Lombardia M Duran R Pompeo F Cole ST Cervenansky C Alzari PM PknB kinase activity is regulated by phosphorylation in two Thr residues and dephosphorylation by PstP, the cognate phospho-Ser/Thr phosphatase, in Mycobacterium tuberculosis Mol Microbiol 2003 49 1493 508 12950916 10.1046/j.1365-2958.2003.03657.x
Morrison DK Murakami MS Cleghon V Protein kinases and phosphatases in the Drosophila genome J Cell Biol 2000 150 F57 62 10908587 10.1083/jcb.150.2.F57
Kostich M English J Madison V Gheyas F Wang L Qiu P Greene J Laz TM Human members of the eukaryotic protein kinase family Genome Biol 2002 3 RESEARCH0043 12225582 10.1186/gb-2002-3-9-research0043
Manning G Whyte DB Martinez R Hunter T Sudarsanam S The protein kinase complement of the human genome Science 2002 298 1912 1934 12471243 10.1126/science.1075762
Krupa A Srinivasan N The repertoire of protein kinases encoded in the draft version of the human genome: atypical variations and uncommon domain combinations Genome Biol 2002 3 RESEARCH0066 12537555 10.1186/gb-2002-3-12-research0066
Kim D Forst S Genomic analysis of the histidine kinase family in bacteria and archaea Microbiology 2001 147 1197 1212 11320123
Koretke KK Lupas AN Warren PV Rosenberg M Brown JR Evolution of two-component signal transduction Mol Biol Evol 2000 17 1956 1970 11110912
Bashton M Chothia C The geometry of domain combination in proteins J Mol Biol 2002 315 927 939 11812158 10.1006/jmbi.2001.5288
Teichmann SA Murzin AG Chothia C Determination of protein function, evolution and interactions by structural genomics Curr Opin Struct Biol 2001 11 354 363 11406387 10.1016/S0959-440X(00)00215-3
Park J Lappe M Teichmann SA Mapping protein family interactions: intramolecular and intermolecular protein family interaction repertoires in the PDB and yeast J Mol Biol 2001 307 929 938 11273711 10.1006/jmbi.2001.4526
Kamemura K Hart GW Dynamic interplay between O-glycosylation and O-phosphorylation of nucleocytoplasmic proteins: a new paradigm for metabolic control of signal transduction and transcription Prog Nucleic Acid Res Mol Biol 2003 73 107 136 12882516
Wells L Vosseller K Hart GW Glycosylation of nucleocytoplasmic proteins:signal transduction and O-GlcNAc Science 2001 291 2376 2378 11269319 10.1126/science.1058714
Madec E Stensballe A Kjellstrom S Cladiere L Obuchowski M Jensen ON Seror SJ Mass spectrometry and site-directed mutagenesis identify several autophosphorylated residues required for the activity of PrkC, a Ser/Thr kinase from Bacillus subtilis J Mol Biol 2003 330 459 472 12842463 10.1016/S0022-2836(03)00579-5
Madec E Laszkiewicz A Iwanicki A Obuchowski M Seror S Characterization of a membrane-linked Ser/Thr protein kinase in Bacillus subtilis, implicated in developmental processes Mol Microbiol 2002 46 571 586 12406230 10.1046/j.1365-2958.2002.03178.x
Bork P Brown NP Hegyi H Schultz J The protein phosphatase 2C (PP2C) superfamily: detection of bacterial homologues Protein Sci 1996 5 1421 1425 8819174
Irmler A Forchhammer K A PP2C-type phosphatase dephosphorylates the PII signaling protein in the cyanobacterium Synechocystis PCC 6803 Proc Natl Acad Sci U S A 2001 98 12978 12983 11687619 10.1073/pnas.231254998
Dutta S Lewis RJ Crystallization and preliminary crystallographic analysis of the kinase-recruitment domain of the PP2C-type phosphatase RsbU Acta Crystallogr D Biol Crystallogr 2003 59 191 193 12499568 10.1107/S0907444902020723
Delumeau O Dutta S Brigulla M Kuhnke G Hardwick SW Volker U Yudkin MD Lewis RJ Functional and structural characterization of RsbU, a stress signaling protein phosphatase 2C J Biol Chem 2004 279 40927 40937 15263010 10.1074/jbc.M405464200
Lai SM Le Moual H PrpZ, a Salmonella enterica serovar Typhi serine/threonine protein phosphatase 2C with dual substrate specificity Microbiology 2005 151 1159 1167 15817783 10.1099/mic.0.27585-0
Duncan L Alper S Arigoni F Losick R Stragier P Activation of cell-specific transcription by a serine phosphatase at the site of asymmetric division Science 1995 270 641 644 7570023
Feucht A Magnin T Yudkin MD Errington J Bifunctional protein required for asymmetric cell division and cell-specific transcription in Bacillus subtilis Genes Dev 1996 10 794 803 8846916
Russell LC Whitt SR Chen MS Chinkers M Identification of conserved residues required for the binding of a tetratricopeptide repeat domain to heat shock protein90 J Biol Chem 1999 274 20060 20063 10400612 10.1074/jbc.274.29.20060
Chen MS Silverstein AM Pratt WB Chinkers M The tetratricopeptide repeat domain of protein phosphatase 5 mediates binding to glucocorticoid receptor heterocomplexes and acts as a dominant negative mutant J Biol Chem 1996 271 32315 32320 8943293 10.1074/jbc.271.50.32315
Freestone P Nystrom T Trinei M Norris V The universal stress protein, UspA, of Escherichia coli is phosphorylated in response to stasis J Mol Biol 1997 274 318 324 9405142 10.1006/jmbi.1997.1397
Sessa G Martin GB Signal recognition and transduction mediated by the tomato Pto kinase: a paradigm of innate immunity in plants Microbes Infect 2000 2 1591 1597 11113378 10.1016/S1286-4579(00)01315-0
Black K Buikema WJ Haselkorn R The hglK gene is required for localization of heterocyst-specific glycolipids in the cyanobacterium Anabaena sp. strain PCC 7120 J Bacteriol 1995 177 6440 6448 7592418
White KA Lin S Cotter RJ Raetz CR A Haemophilus influenzae gene that encodes a membrane bound 3-deoxy-D-manno-octulosonic acid (Kdo) kinase. Possible involvement of kdo phosphorylation in bacterial virulence J Biol Chem 1999 274 31391 31400 10531340 10.1074/jbc.274.44.31391
Yethon JA Whitfield C Purification and characterization of WaaP from Escherichia coli, a lipopolysaccharide kinase essential for outer membrane stability J Biol Chem 2001 276 5498 5504 11069912 10.1074/jbc.M008255200
Av-Gay Y Everett M The eukaryotic-like Ser/Thr protein kinases of Mycobacterium tuberculosis Trends Microbiol 2000 8 238 244 10785641 10.1016/S0966-842X(00)01734-0
Walburger A Koul A Ferrari G Nguyen L Prescianotto-Baschong C Huygen K Klebl B Thompson C Bacher G Pieters J Protein kinase G from pathogenic mycobacteria promotes survival within macrophages Science 2004 304 1800 1804 15155913 10.1126/science.1099384
Cowley S Ko M Pick N Chow R Downing KJ Gordhan BG Betts JC Mizrahi V Smith DA Stokes RW Av-Gay Y The Mycobacterium tuberculosis protein serine/threonine kinase PknG is linked to cellular glutamate/glutamine levels and is important for growth in vivo Mol Microbiol 2004 52 1691 702 15186418 10.1111/j.1365-2958.2004.04085.x
Dorow DS Devereux L Tu GF Price G Nicholl JK Sutherland GR Simpson RJ Complete nucleotide sequence, expression, and chromosomal localisation of human mixed-lineage kinase 2 Eur J Biochem 1995 234 492 500 8536694 10.1111/j.1432-1033.1995.492_b.x
Katoh M Hirai M Sugimura T Terada M Cloning and characterization of MST, a novel (putative) serine/threonine kinase with SH3 domain Oncogene 1995 10 1447 1451 7731697
Joshi B Janda L Stoytcheva Z Tichy P PkwA, a WD-repeat protein, is expressed in spore-derived mycelium of Thermomonospora curvata and phosphorylation of its WD domain could act as a molecular switch Microbiology 2000 146 3259 3267 11101684
Ponting CP Aravind L Schultz J Bork P Koonin EV Eukaryotic signalling domain homologues in archaea and bacteria. Ancient ancestry and horizontal gene transfer J Mol Biol 1999 289 729 745 10369758 10.1006/jmbi.1999.2827
Kamei A Yoshihara S Yuasa T Geng X Ikeuchi M Biochemical and functional characterization of a eukaryotic-type protein kinase, SpkB, in the cyanobacterium, Synechocystis sp. PCC 6803 Curr Microbiol 2003 46 296 301 12732981 10.1007/s00284-002-3887-2
Kamei A Yuasa T Geng X Ikeuchi M Biochemical examination of the potential eukaryotic-type protein kinase genes in the complete genome of the unicellular Cyanobacterium synechocystis sp. PCC 6803 DNA Res 2002 9 71 78 12168951 10.1093/dnares/9.3.71
Kamei A Yuasa T Orikawa K Geng XX Ikeuchi M A eukaryotic-type protein kinase, SpkA, is required for normal motility of the unicellular Cyanobacterium synechocystis sp. strain PCC 6803 J Bacteriol 2001 183 1505 1510 11160079 10.1128/JB.183.5.1505-1510.2001
Peirs P De Wit L Braibant M Huygen K Content J A serine/threonine protein kinase from Mycobacterium tuberculosis Eur J Biochem 1997 244 604 612 9119030 10.1111/j.1432-1033.1997.00604.x
Namboori S Srinivasan N Pandit SB Recognition of remotely related structural homologues using sequence profiles of aligned homologous protein structures In Silico Biol 2004 4 0037
Galperin MY A census of membrane-bound and intracellular signal transduction proteins in bacteria: bacterial IQ, extroverts and introverts BMC Micro Biol 2005 5 35 10.1186/1471-2180-5-35
Grundner C Gay LM Alber T Mycobacterium tuberculosis serine/threonine kinases PknB, PknD, PknE, and PknF phosphorylate multiple FHA domains Protein Sci 2005 14 1918 1921 15987910 10.1110/ps.051413405
FTP site at NCBI
Altschul SF Madden TL Schaffer AA Zhang J Zhang Z Miller Lipman DJ Gapped BLAST and PSI-BLAST: a new generation of protein database search programs Nucleic Acids Res 1997 25 3389 3402 9254694 10.1093/nar/25.17.3389
Schaffer AA Wolf YI Ponting CP Koonin EV Aravind L Altschul SF IMPALA: matching a protein sequence against a collection of PSI-BLAST-constructed position specific score matrices Bioinformatics 1999 15 1000 1011 10745990 10.1093/bioinformatics/15.12.1000
Eddy SR Profile hidden Markov models Bioinformatics 1998 14 755 763 9918945 10.1093/bioinformatics/14.9.755
Muller A MacCallum RM Sternberg MJ Benchmarking PSI-BLAST in genome annotation J Mol Biol 1999 293 1257 1271 10547299 10.1006/jmbi.1999.3233
Lindahl E Elofsson A Identification of related proteins on family, superfamily and fold level J Mol Biol 2000 295 613 625 10623551 10.1006/jmbi.1999.3377
Bateman A Birney E Cerruti L Durbin R Etwiller L Eddy SR Griffiths-Jones S Howe KL Marshall M Sonnhammer EL The Pfam protein families database Nucleic Acids Res 2002 30 276 280 11752314 10.1093/nar/30.1.276
Krogh A Larsson B von Heijne G Sonnhammer EL Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes J Mol Biol 2001 305 567 580 11152613 10.1006/jmbi.2000.4315
Russell RB Barton GJ Multiple protein sequence alignment from tertiary structure comparison: assignment of global and residue confidence levels Proteins 1992 2 309 323 1409577 10.1002/prot.340140216
|
16171520
|
PMC1262709
|
CC BY
|
2021-01-04 16:32:48
|
no
|
BMC Genomics. 2005 Sep 19; 6:129
|
utf-8
|
BMC Genomics
| 2,005 |
10.1186/1471-2164-6-129
|
oa_comm
|
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-1321617430410.1186/1471-2164-6-132Research ArticleMolecular signature of clinical severity in recovering patients with severe acute respiratory syndrome coronavirus (SARS-CoV) Lee Yun-Shien [email protected] Chun-Houh [email protected] Angel [email protected] En-Shih [email protected] Min-Li [email protected] Lung-Kun [email protected] Kuender D [email protected] Meng-Chih [email protected] Yi-Hsi [email protected] Jien-Wei [email protected] Hock-Liew [email protected] Ping-Cherng [email protected] Ting-Shu [email protected] Kuo-Chein [email protected] Chung-Guei [email protected] Yin-Jing [email protected] Tzu-Hao [email protected] Hsing-Shih [email protected] Ying-Shiung [email protected] Genomic Medicine Research Core Laboratory, Chang Gung Memorial Hospital (CGMH), Tao-Yuan, Taiwan2 Department of Biotechnology, Ming Chuan University, Tao-Yuan, Taiwan3 Institute of Statistical Science, Academia Sinica, Taipei, Taiwan4 Department of Obstetrics and Gynecology, Lin-Kou Medical Center, CGMH, Tao-Yuan, Taiwan5 Graduate Institute of Clinical Medical Sciences, College of Medicine, Chang Gung University, Tao-Yuan, Taiwan6 Department of Pediatrics, Kaohsiung Medical Center, CGMH, Kaohsiung, Taiwan7 Department of Internal Medicine, Division of Pulmonary and Critical Care Medicine, Kaohsiung Medical Center, CGMH, Kaohsiung, Taiwan8 Department of Internal Medicine, Division of Infectious Diseases, Kaohsiung Medical Center, CGMH, Kaohsiung, Taiwan9 Department of Pathology, Kaohsiung Medical Center, CGMH, Kaohsiung, Taiwan10 Department of Internal Medicine, Division of Infectious Diseases, Lin-Kou Medical Center, CGMH, Tao-Yuan, Taiwan11 Clinical Virology Laboratory, Department of Clinical Pathology, CGMH, Tao-Yuan, Taiwan12 Institute of Statistics, National Central University, Tao-Yuan, Taiwan2005 21 9 2005 6 132 132 11 3 2005 21 9 2005 Copyright © 2005 Lee et al; licensee BioMed Central Ltd.2005Lee et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Severe acute respiratory syndrome (SARS), a recent epidemic human disease, is caused by a novel coronavirus (SARS-CoV). First reported in Asia, SARS quickly spread worldwide through international travelling. As of July 2003, the World Health Organization reported a total of 8,437 people afflicted with SARS with a 9.6% mortality rate. Although immunopathological damages may account for the severity of respiratory distress, little is known about how the genome-wide gene expression of the host changes under the attack of SARS-CoV.
Results
Based on changes in gene expression of peripheral blood, we identified 52 signature genes that accurately discriminated acute SARS patients from non-SARS controls. While a general suppression of gene expression predominated in SARS-infected blood, several genes including those involved in innate immunity, such as defensins and eosinophil-derived neurotoxin, were upregulated. Instead of employing clustering methods, we ranked the severity of recovering SARS patients by generalized associate plots (GAP) according to the expression profiles of 52 signature genes. Through this method, we discovered a smooth transition pattern of severity from normal controls to acute SARS patients. The rank of SARS severity was significantly correlated with the recovery period (in days) and with the clinical pulmonary infection score.
Conclusion
The use of the GAP approach has proved useful in analyzing the complexity and continuity of biological systems. The severity rank derived from the global expression profile of significantly regulated genes in patients may be useful for further elucidating the pathophysiology of their disease.
==== Body
Background
SARS-CoV is a single-stranded, plus-sense RNA virus with a genome of ~30 kb. Its sequence does not closely resemble any of the previously characterized coronaviruses [1-4]. Before SARS-CoV was recognized as the cause of the deadly SARS [1-3,5-7], other human coronaviruses had only been known to account for 15–30% of colds [8]. SARS-CoV appears to be new to humans, as supported by the finding that human sera collected before the SARS outbreak did not contain antibodies against this virus [3,9]. After an incubation period from 2 to 10 days, SARS patients might develop fever (>38°C), headache, dry cough, and pneumonia [3,5,9-14]. Most patients gradually recovered while some progressed to respiratory distress syndrome with ~10% mortality rate. The genome-wide changes in human gene expression when challenged by this novel pathogen are essentially unknown.
Profiles of gene expression patterns help define the complex biological processes associated with both health and disease in vivo. Investigation of host responses to infection with in vitro models have offered insights into mechanisms of pathogenesis, and have highlighted the potential for applications of microarray technology to diagnose infection in vivo [15]. Whitney et al. observed that the variation in gene expression patterns in the blood of healthy subjects was strikingly smaller than the significant changes induced by diseases either in patients with cancer or with bacterial infections [15]. It was conceivable that microarray profiling of gene expression in whole bloods exhibits the potential in monitoring the patients' responses to a disease, especially a novel infection such as SARS.
Many discriminative methods have been developed for analysis of microarray gene expression data in cancer patients and the resulting classifications have been correlated closely with clinical parameters [16-19]. For instance, the discovery of signature genes for breast cancers through microarray analysis of gene expression has provided us with a more precise clinical staging that will improve the outcome of treatment [20,21]. However, clinical parameters are not always in a discrete pattern but more likely in a continuous fashion, where an absolute classification may not be achievable. Herein we present the use of cDNA microarray analysis of gene expression in whole blood from a cohort of recovering SARS patients, of whom the disease severity appeared to be a continuum. After we had identified the molecular signature of 52 genes that accurately discriminated acute SARS patients from non-SARS controls, we ranked the disease severity of these patients using a generalized association plot (GAP) elliptical seriation algorithm [22] based on the expression profiles of the 52 genes. The derived severity rank of the patients proved to be closely correlated with their clinical parameters, namely, the recovery period (in days) and the clinical pulmonary infection score.
Results
Patient information
Using the cDNA microarrays spotted with duplicated 7,334 cDNA clone, we analyzed RNA specimens successfully amplified in 44 peripheral blood collected from 25 confirmed SARS patients (age ranged from 23 to 80 years old, mean = 41.8, SD = 17.2, median = 34), of whom 24 survived. Except for one patient who died on the 4th day, duration of hospitalization in this cohort ranged from 12 to 51 days (n = 24, mean = 24.5, SD = 10.1, median = 21) (Additional file 1). We defined 11 specimens as acute SARS (AS) using the following criteria: (i) the whole blood RNA from a hospitalized patient was PCR positive for SARS-CoV, or (ii) the specimen was collected within 10 days after the disease onset in patients whose blood was later diagnosed ELISA-positive for anti-SARS IgG. The rest of 33 RNA specimens from SARS patients were labelled as recovering SARS (RS). Our study included 11 normal control (NC) volunteers and 11 patients with bacterial infections (IN) as healthy and non-SARS infection controls, respectively (Additional files 2 and 3).
cDNA microarray analysis
When we compared the gene expression profiles among acute SARS (AS), recovering SARS (RS), bacterial infection (IN), and normal control (NC) groups, we observed the variances of gene expression in both SARS (AS, AS+RS) and bacterial (IN) groups to be equally higher than that in healthy controls (NC) (Fig. 1a). This result indicates that gene expression profiles of either SARS or bacterial groups differed significantly from that of normal controls.
Figure 1 Significant differences in gene expression profiles in patients with SARS or bacterial infection. Using a probe set with 6,525 annotated genes, global gene expression was analyzed by (A) variation distribution in peripheral blood specimens from patients with acute SARS (AS), recovering SARS (RS), bacterial infections (IN), and normal controls (NC). (B) In the hierarchical clustering of relative change in gene expression using a probe set of 885 filtered genes (gene vector >0.5 SD), red indicates upregulation and green indicates downregulation in gene expression relative to a common reference that was the pooled amplified RNA from 11 normal controls. (C) Using the 885-gene set, singular value decomposition (SVD) analysis by two eigenvectors showed three distinguished clusters of AS (red ●), IN (green ▲), and NC (blue ▮) groups, with the RS (red ○) specimens scattering among AS and IN.
A probe set of 885 genes with standard deviations greater than 0.5 across 66 arrays was selected for further analyses. An average linkage hierarchical clustering tree with Pearson correlation proximity was built on the 33 arrays (11 NC, 11 IN, and 11AS) using these 885 genes (Fig. 1b). The AS and NC groups were well separated into two opposite coherent clusters. Singular value decomposition (SVD) analysis, a dimension reduction method to project gene expression profiles to fewer representative eigenvectors [23], also successfully separated AS, IN, and NC specimens into three clusters with first two eigenvectors (Fig. 1c). Interestingly, the recovering SARS (RS) samples are interspersed among the AS and IN samples.
To identify which genes were specifically regulated by SARS-CoV, we performed two sets of two-sample Student t-test for means with an unequal-variance assumption. In the first set, we contrasted 11 AS versus 22 non-AS (NC and IN) specimens on all 885 genes. The genes with significant testing results were considered to be specifically induced by SARS-CoV (Fig. 2a,b). For the second set of t-tests, we compared 11 NC with 22 non-NC (IN and AS) specimens. We considered that the change in significant genes identified by the second t-test was induced by both bacterial and viral infections (Fig. 2c,d). Genes identified from these two sets of test were then ranked separately according to the corresponding sets of P-values. Gene expression profiles for the top 20 and the bottom 20 genes from both sets are displayed as Figure 2.
Figure 2 The top 40 discriminating genes with the highest distinction values for AS or NC groups. Twenty genes that were specifically (A) upregulated or (B) downregulated in patients with SARS. Another twenty genes that were non-specifically (C) upregulated or (D) downregulated by both bacterial infection and SARS. Each column represents an individual sample and each row represents a gene. The color range reflected relative change according to the scale shown. NC, normal control; IN, bacterial infection; AS, acute SARS. GMRCL clone numbers of some ESTs are also included in the parentheses.
Unexpectedly, most of the genes specifically upreguated by SARS-CoV are ESTs (13/20 genes) that were not annotated previously (Fig. 2a). On the other hand, SARS-CoV stimulated the host innate immunity by upregulating genes including defensins [24,25] and eosinophil derived neurotoxin [26,27], similar to that of bacterial infections (Fig. 2c).
Signature genes and GAP algorithms
A simple k-nearest-neighbour method was used to obtain a near optimal number of 30 genes from the 885 filtered gene set for discriminating specimens between acute SARS (AS) and non-SARS (NC and IN) (Additional file 4). The selected top 30 upregulated (P < 6 × 10-6) and the top 30 downregulated genes (P < 4 × 10-7) from the AS versus non-AS (IN and NC) Student's t-test were used as the specific probe set to assess the status of SARS infection. Eight genes that were also significant in the NC versus non-NC (AS and IN) t-test were excluded, resulting in a specific AS probe set of 52 genes. For the GAP analysis, we calculated pair-wise Euclidean distances among 55 samples (11 AS, 33 RS, 11 NC) using these 52 genes, aiming to identify a one-dimensional order that could reflect the severity structure of the disease (Fig. 3a). Using this GAP elliptical arrangement of 55 specimens (columns), we observed a transition of gene expression patterns of 52 genes (rows) from the left side where NC clustered to the right side where AS accumulated (Fig. 3b). Hierarchical clustering trees guided by self-organized map (SOM) and other clustering methods were also performed to sort the SARS patient samples using the same 52 genes. The Robinson criterion [22,28,29] is often employed to assess the performances of different seriation algorithms. Table 1 (and Additional file 5) shows that the GAP algorithm derived a smoother transition pattern than other methods in the Robinson sense. Thereby, we have derived the SARS severity rank according to the expression profile of 52 signature genes as a whole in each patient, as demonstrated by the smooth transition of expression levels in each (row) of these genes from NC to RS to AS (Fig. 3b).
Figure 3 Generalized associated plots (GAP) analysis of SARS patients samples. (A) Pair-wise Euclidean distance matrix that was sorted by a GAP using 52 genes with the highest discriminating power for AS groups revealed the minimum anti-Robinson events in the matrix, resulting in a smooth transition order of the AS and RS specimens from severely diseased to healthy states. AS (red ●); RS (red ○); NC (blue ▮). (B) Gene expression profile for the 52 discriminating genes displayed in the order obtained from the GAP method.
Table 1 Performance of Robinson structure with different seriation algorithms.
Anti-Robinson Events
Seriation Algorithm Counts Scaled Counts (%)
Random Order (NC-AS-RS) 26,245 100.00
Original Order 23,030 87.75
Self Organizing Map (SOM) Order 11,499 43.81
Average Linkage Tree/Original 10,268 39.12
Average Linkage Tree/SOM 10,738 40.91
Average Linkage Tree/GAP 5,940 22.63
GAP Elliptical Order 5,022 19.14
For validation purposes, we further tested the stability of the rank (order) derived from GAP analysis on the 52 genes for the 55 specimens. The same GAP procedure was repeatedly applied to the top 20 to 200 genes (among the filtered 885 genes) with significant p-values (Student's t-test) between the AS versus non-AS (IN and NC). While the ranks for the 55 specimens obtained from the most significant 20 to 200 genes are highly correlated to each other, they are significantly different from the ranks derived from the 52-gene sets that were randomly selected from the 885 genes (data not shown).
We scrutinized the clinical courses of patients who donated the 10 RS specimens that were scattered among AS (Fig. 3a) and found evidence of underlying severity of the disease in the majority of patients. For example, sample RS43 from a patient who had been discharged from hospital for 2 weeks was still PCR-positive for SARS-CoV; RS54, a PCR-positive sample was not grouped as AS because of the negative ELISA result. RS38, RS40, and RS42 still represented acute SARS infections because they were collected only 1, 2, and 3 days after AS37, AS39, and AS41, respectively. Patients with RS78 and RS91 who had severe SARS courses were hospitalized for 41 and 51 days, respectively. The patient for RS8 was in the second week of disease. The only two unexplained specimens, RS18 and RS71 from the same patient, may represent a unique biological variability, accounting for the misclassification using this 52-gene molecular signature.
Molecular signature for severity and clinical correlations
To test the efficacy of using these 52 genes as the molecular signature for the severity of SARS patients, we identified a significant correlation (P < 1 × 10-6) between the derived rank of SARS severity and the number of days after the onset of disease (Fig. 4a). We further used this rank of SARS severity to examine the recovery trend in 17 recovering patients who had donated multiple specimens (Fig. 4b). Except for the one patient (5.3 % = 1/19, shown as the red line in Fig. 4b), similar trends existed in 18 out of 19 lines (94.7 %). Pugin et al. combined body temperature, white blood cell count, volume and appearance of tracheal secretions, oxygenation, chest X-ray, and tracheal aspirate cultures into a clinical pulmonary infection score (CPIS) as a diagnostic tool for pneumonia [30]. We observed that the rank of SARS severity was also significantly (P < 0.001) correlated with the CPIS (Fig. 4c). Collectively, these results demonstrate a correlation between the molecular severity rank and clinical factors, suggesting the usefulness of the molecular signature as a genome-wide parameter for gauging the severity of SARS patients.
Figure 4 Correlations between the GAP-derived rank for SARS severity and clinical parameters. (A) The scatter plot of all SARS specimens with the order obtained from the GAP method and the days after the onset of disease showed a significant correlation (P < 5 × 10-7). (B) Sixteen out of 17 SARS patients who submitted multiple blood specimens showed a similar trend of changes in the GAP-derived severity rank along with the recovery from the disease. Patients with 2 (n = 15) and 3 specimens (n = 2) were labeled with blue and green lines, respectively. (C) The scatter plot of all AS and RS specimens with the order obtained from the GAP method and clinical pulmonary infection score (CPIS) showed a significant correlation (P < 0.001).
Discussion
Diverse infections can induce a shared core gene expression involving the human innate immune system; each infection may also trigger a pathogen-specific immune response of the host. The innate immune genes were upregulated in both acute SARS (AS) and bacteria infection (IN) patients (Fig. 2c). SARS was a novel viral infection that had not been encountered by the humans in the history before 2003. Intriguingly, most of the genes specifically upreguated in SARS patients were ESTs (13/20 genes) (Fig. 2a), suggesting that the first human encounter with SARS-CoV might provoke a set of human genes that were poorly annotated due to disuse. Annotation of these ESTs may lead to the discovery of novel genes.
Given the high cost of microarray analyses, the detection of a comprehensive gene expression profile may not be cost-effective for clinical diagnosis and evaluation of patients with infectious diseases. However, in a complex system such as the human body where genes interplay through intricate circuitries, it is inadequate to examine only a few routine parameters in biochemistry and blood cell counts for the global physiochemical status of a patient at the time of blood collection. In this report, we applied the GAP method to derive a smooth transition pattern among samples based on the molecular signature consisting of 52 genes, which in turn were used to monitor the severity of clinical courses of SARS patients. Instead of clustering samples into discrete groups in a method similar to commonly-used microarray classifications [31], GAP focuses more on a global orientation of the sample-to-sample relationship. For instance, the AS and RS samples were seriation ranked (Fig. 3), and the rank order proved to correlate well with clinical parameters (Fig. 4).
The GAP-derived rank of severity also provided us with a unique way, where expression of most relevant genes were all considered, to decipher the meaning of the changes in other genes obtained from the same microarray experiment. For instance, we have identified the correlative change in matrix metalloproteinase MMP-7 and MMP-9 (Additional file 6): both can stimulate α-defensin [32]. Importantly, these correlations could not be revealed with other parameters alone, such as number of days after disease onset or clinical score CPIS (data not shown).
In this study, however, there might be technical limitations during RNA isolation from some clinical specimens as well as an unavoidable sample-collecting bias. First, both RNA isolation from SARS specimens and RNA amplification were performed in the Biosafety Level 3 laboratory, where the instrument for RNA quantitation was not available. This limitation resulted in the failed generation of aRNA from 10 out of 54 SARS specimens (Additional file 1). Unfortunately, these 10 specimens contained 7 specimens from patients at an early (i.e. first 2 weeks) stage [14]. Secondly, 25 SARS patients who donated blood specimens for this study may belong to the milder subgroup of a total of 44 SARS patients in Kaohsiung Medical Center of Chang Gung Memorial Hospital. According to a paper describing the complete cohort of SARS patients [33], intubation and mechanical ventilation were required in 20 out of these 44 patients. However, only two in our 25 patients needed intubation (Additional file 1). The aforementioned two potential limitations may account for why our microarray results could not detect a correlation with a possible worsening clinical course before recovering, which was described by Peiris et al [14].
In conclusion, we propose the use of a molecular signature reflecting the severity of SARS in order to interpret the trends of expression changes in groups of genes within particular functional categories. The use of GAP methodology proved to be instrumental in determining the severity of SARS. The derived severity ranking of SARS patients in turn formed a gradual basis for the analysis of the interaction patterns, providing us with a useful tool for understanding the molecular pathogenesis of this novel viral infection.
Conclusion
We illuminate the human gene expression profiles, in terms of gene expression in peripheral blood, to the unprecedented infection of SARS-CoV. We also discovered a smooth transition pattern of severity from normal controls to acute SARS patients based on the gene expression profiles by generalized associate plots (GAP). The rank of SARS severity was significantly correlated with other clinical parameters.
Methods
Patient information and specimen preparation
Blood specimens of 25 SARS patients (Additional file 1) were collected from 10 May to 4 July 2003 at Kaohsiung Medical Center of Chang Gung Memorial Hospital (CGMH) in Kaohsiung City of southern Taiwan. Two additional blood samples (RS94 and RS97) were collected from apparently healthy individuals who had recovered from SARS infection 3 months later. Diagnosis of SARS was based on the guidelines of World Health Organization (WHO) [34]. More comprehensive data of the SARS cohort were previously published [33]. This study was approved by the Institute Review Board of CGMH. Total RNA was isolated with the PAXgene Blood RNA System (Qiagen, USA) and stored at -80°C. After RNAs were further purified and concentrated into 15 μl BR5 solution with RNeasy MinElute kit (Qiagen, USA), 2 μl were used for linear RNA amplification using RiboAmp RNA Amplification Kit (Arcturus, California USA). Before the first Strand Nuclease Mix was added to the RNA samples, all of the RNA purification and amplification were performed inside a Biosafety Level 3 laboratory located in Lin-Kou Medical Center of CGMH. We analyzed the quality and quantity of amplified RNA with Bioanalyzer 2100 (Agilent, CA, USA).
Anti-SARS-CoV IgG ELISA and real-time quantitative PCR analysis
The antigen used for the SARS detection ELISA was the detergent-extracted and gamma irradiated Vero E6 cells infected with SARS-CoV. Identical preparations from uninfected Vero E6 cells were used as the control. Patients' sera were 1:10 diluted and added to the ELISA plates, and goat anti-human IgG antibody conjugated with horseradish peroxidase (DAKO, Cambridgeshire, UK) was added for enzymatic reaction. After adding the substrate, O-phenylenediamine, the optical density (O.D.) was measured at 450 nm wavelength. The cutoff value of O.D. for SARS-CoV IgG ELISA was 0.15. Sensitivity of this method was 100% (28/28 in confirmed SARS cases) and specificity was 98.4% (790/803) in the healthy control group.
Real-time quantitative PCR analysis for SARS-CoV was performed with Cor-p-F4, Cor-p-R4 and Cor-probe developed by CDC (GA, USA) with HT 7900 Sequence Detection System (Applied Biosystems, CA, USA).
Microarray procedures
In this study, we used the GMRCL Human 7K set, Version 2 chips as previously described [35]. Twelve amplified RNA samples from healthy donors (Additional file 3) were pooled as the common reference for every array in this study. A total of 66 aRNA samples including 11 acute SARS (AS), 33 recovering SARS (RS), 11 non-SARS infection (IN), and 11 normal controls (NC) were analyzed with cDNA microarrays as tests against the pooled aRNA (the common reference). Among 66 aRNA preparations, 28 were analyzed with the dye-swapping microarray design. We averaged the log ratios of the duplicated spots on each slide. In the dye swapping experiments, we further averaged the log ratios derived from two slides. We used 400 ng of aRNA for labeling and hybridization using a 3DNA Array 350RP Detection kit (Genisphere, PA, USA), and scanned slides with a confocal scanner ChipReader (Virtek, Canada). We acquired the spot and background intensities with GenePix Pro 4.1 software (Axon Instruments, Inc., CA, USA), and carried out within-slide normalization using programs written with MATLAB 6.0 software (The MathWorks, Inc., MA, USA). To assure the reproducibility of our microarray system, we got the similar gene expression profiles from replicated samples (RS88) using the hierarchical clustering analysis and also got the highly correlated results (r2 = 0.84) from two specimens (AS37 and RS38) that were collected from the same patient at a time interval of only one day. We consistently obtained identical results in each of 28 pairs dye-swapping experiments. The complete microarray data is available in Additional file 7.
Hierarchical clustering and singular value decomposition
We performed hierarchical clustering using Cluster and TreeView software [36] with the following parameters: (i) a standard deviation > 0.5 as the filtering cutoff point (885 genes with marked changes selected among 66 arrays), (ii) mean-centered genes and normalized genes, (iii) cluster analysis carried out with uncentered correlation of arrays. We also performed a singular value decomposition (SVD) [23] analysis of the correlation matrix for all 66 samples. The first two eigenvectors weighted by the corresponding singular values (eigenvalues) of the 66 samples were plotted against each other.
Euclidean distance matrix by generalized association plots
Robinson criterion [22,28] is frequently used to assess the performances of sorting algorithms on symmetric proximity matrices. A Robinson Matrix, R = [rij], is a symmetric matrix such that rij ≤ rik if j<k<i and rij ≥ rik if i<j<k. The GAP elliptical seriation [22] utilizing the ellipse structure from a singular value decomposition of a converged correlation coefficient matrix usually identifies permuted matrix with a near Robinson form. A brief review on GAP and some details of its applications are available [37].
Authors' contributions
Lee YS, Chen CH and Wang TH designed the study and prepared the manuscript. Lee YS, Tien-YJ and Chen CH carried out the statistical analysis. Lee YS, Wang TH, Chen ES, Wei ML, Chen LK carried out the microarray experiments. Chao A, Yang KD, Lin MC, Wang YH, Liu JW, Eng HW, Chiang PC, Wu TS, Tsao KC, Huang CG, Wang HS and Lee YS obtained the clinical materials and analyzed clinical information. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Demographics of SARS Patients.
Click here for file
Additional File 2
Demographics of patients with non-SARS infection.
Click here for file
Additional File 3
Demographics of healthy donor (information of the pooled reference).
Click here for file
Additional File 4
K-nearest-neighbour methods in evaluating the best discriminating (classifying) accuracy for AS and non-SARS specimens.
Click here for file
Additional File 5
Euclidean distance matrix of 55 specimens with 52 selected genes using different seriation algorithms.
Click here for file
Additional File 6
Analyses of gene expression in MMP-7 and MMP-9, both of which are involved in innate immunity.
Click here for file
Additional File 7
Complete microarray data
Click here for file
Acknowledgements
We thank Yalin Huang, Fong-Yee Chiu, Yu-Liey Tong, Wei-Hsiang Kong, Shihyee Mimi Wang (University of Illinois in Chicago), Hsiu-Chuan Liu, Rong-Fu Chen and Ling-Ling Huang for technical assistance, Shih-Tien Wang (Northwestern University) for editing the manuscript, and PC Huang (Johns Hopkins University) for critical comments. The authors also gratefully acknowledge the SARS team of Kaohsiung Chang Gung Memorial Hospital (Yun-Tze Chen, Ju-Hao Lee, Sui-Liong Wang, Tze-Yu Lee, Chao-Chien Wu, Sheung-Fat Ko, Chen-Hsiang Lee) and many more medical personnel who served courageously during the SARS episode. This study was supported by grants CMRPD32019S (YS Lee), CMRPG1008 (TH Wang), CMRPG32010S (TH Wang) from Chang Gung Memorial Hospital, NSC93-2320-B-130-001 (YS Lee) from National Science Council of Taiwan, and a generous donation of Mr. Yung-Ching Wang, Chairman of Formosa Plastic Corporation.
==== Refs
Rota PA Oberste MS Monroe SS Nix WA Campagnoli R Icenogle JP Penaranda S Bankamp B Maher K Chen MH Tong S Tamin A Lowe L Frace M DeRisi JL Chen Q Wang D Erdman DD Peret TC Burns C Ksiazek TG Rollin PE Sanchez A Liffick S Holloway B Limor J McCaustland K Olsen-Rasmussen M Fouchier R Gunther S Osterhaus AD Drosten C Pallansch MA Anderson LJ Bellini WJ Characterization of a novel coronavirus associated with severe acute respiratory syndrome Science 2003 300 1394 1399 12730500 10.1126/science.1085952
Marra MA Jones SJ Astell CR Holt RA Brooks-Wilson A Butterfield YS Khattra J Asano JK Barber SA Chan SY Cloutier A Coughlin SM Freeman D Girn N Griffith OL Leach SR Mayo M McDonald H Montgomery SB Pandoh PK Petrescu AS Robertson AG Schein JE Siddiqui A Smailus DE Stott JM Yang GS Plummer F Andonov A Artsob H Bastien N Bernard K Booth TF Bowness D Czub M Drebot M Fernando L Flick R Garbutt M Gray M Grolla A Jones S Feldmann H Meyers A Kabani A Li Y Normand S Stroher U Tipples GA Tyler S Vogrig R Ward D Watson B Brunham RC Krajden M Petric M Skowronski DM Upton C Roper RL The Genome sequence of the SARS-associated coronavirus Science 2003 300 1399 1404 12730501 10.1126/science.1085953
Peiris JS Lai ST Poon LL Guan Y Yam LY Lim W Nicholls J Yee WK Yan WW Cheung MT Cheng VC Chan KH Tsang DN Yung RW Ng TK Yuen KY Coronavirus as a possible cause of severe acute respiratory syndrome Lancet 2003 361 1319 1325 12711465 10.1016/S0140-6736(03)13077-2
Ruan YJ Wei CL Ee AL Vega VB Thoreau H Su ST Chia JM Ng P Chiu KP Lim L Zhang T Peng CK Lin EO Lee NM Yee SL Ng LF Chee RE Stanton LW Long PM Liu ET Comparative full-length genome sequence analysis of 14 SARS coronavirus isolates and common mutations associated with putative origins of infection Lancet 2003 361 1779 1785 12781537 10.1016/S0140-6736(03)13414-9
Drosten C Gunther S Preiser W van der Werf S Brodt HR Becker S Rabenau H Panning M Kolesnikova L Fouchier RA Berger A Burguiere AM Cinatl J Eickmann M Escriou N Grywna K Kramme S Manuguerra JC Muller S Rickerts V Sturmer M Vieth S Klenk HD Osterhaus AD Schmitz H Doerr HW Identification of a novel coronavirus in patients with severe acute respiratory syndrome N Engl J Med 2003 348 1967 1976 12690091 10.1056/NEJMoa030747
Fouchier RA Kuiken T Schutten M Van Amerongen G Van Doornum GJ Van Den Hoogen BG Peiris M Lim W Stohr K Osterhaus AD Aetiology: Koch's postulates fulfilled for SARS virus Nature 2003 423 240 12748632 10.1038/423240a
Kuiken T Fouchier RA Schutten M Rimmelzwaan GF van Amerongen G van Riel D Laman JD de Jong T van Doornum G Lim W Ling AE Chan PK Tam JS Zambon MC Gopal R Drosten C van der Werf S Escriou N Manuguerra JC Stohr K Peiris JS Osterhaus AD Newly discovered coronavirus as the primary cause of severe acute respiratory syndrome Lancet 2003 362 263 270 12892955 10.1016/S0140-6736(03)13967-0
Holmes KV SARS-associated coronavirus N Engl J Med 2003 348 1948 1951 12748314 10.1056/NEJMp030078
Ksiazek TG Erdman D Goldsmith CS Zaki SR Peret T Emery S Tong S Urbani C Comer JA Lim W Rollin PE Dowell SF Ling AE Humphrey CD Shieh WJ Guarner J Paddock CD Rota P Fields B DeRisi J Yang JY Cox N Hughes JM LeDuc JW Bellini WJ Anderson LJ A novel coronavirus associated with severe acute respiratory syndrome N Engl J Med 2003 348 1953 1966 12690092 10.1056/NEJMoa030781
Poutanen SM Low DE Henry B Finkelstein S Rose D Green K Tellier R Draker R Adachi D Ayers M Chan AK Skowronski DM Salit I Simor AE Slutsky AS Doyle PW Krajden M Petric M Brunham RC McGeer AJ Identification of severe acute respiratory syndrome in Canada N Engl J Med 2003 348 1995 2005 12671061 10.1056/NEJMoa030634
Tsang KW Ho PL Ooi GC Yee WK Wang T Chan-Yeung M Lam WK Seto WH Yam LY Cheung TM Wong PC Lam B Ip MS Chan J Yuen KY Lai KN A cluster of cases of severe acute respiratory syndrome in Hong Kong N Engl J Med 2003 348 1977 1985 12671062 10.1056/NEJMoa030666
Donnelly CA Ghani AC Leung GM Hedley AJ Fraser C Riley S Abu-Raddad LJ Ho LM Thach TQ Chau P Chan KP Lam TH Tse LY Tsang T Liu SH Kong JH Lau EM Ferguson NM Anderson RM Epidemiological determinants of spread of causal agent of severe acute respiratory syndrome in Hong Kong Lancet 2003 361 1761 1766 12781533 10.1016/S0140-6736(03)13410-1
Nicholls JM Poon LL Lee KC Ng WF Lai ST Leung CY Chu CM Hui PK Mak KL Lim W Yan KW Chan KH Tsang NC Guan Y Yuen KY Peiris JS Lung pathology of fatal severe acute respiratory syndrome Lancet 2003 361 1773 1778 12781536 10.1016/S0140-6736(03)13413-7
Peiris JS Chu CM Cheng VC Chan KS Hung IF Poon LL Law KI Tang BS Hon TY Chan CS Chan KH Ng JS Zheng BJ Ng WL Lai RW Guan Y Yuen KY Clinical progression and viral load in a community outbreak of coronavirus-associated SARS pneumonia: a prospective study Lancet 2003 361 1767 1772 12781535 10.1016/S0140-6736(03)13412-5
Whitney AR Diehn M Popper SJ Alizadeh AA Boldrick JC Relman DA Brown PO Individuality and variation in gene expression patterns in human blood Proc Natl Acad Sci USA 2003 100 1896 1901 12578971 10.1073/pnas.252784499
van de Vijver MJ He YD van't Veer LJ Dai H Hart AA Voskuil DW Schreiber GJ Peterse JL Roberts C Marton MJ Parrish M Atsma D Witteveen A Glas A Delahaye L van der Velde T Bartelink H Rodenhuis S Rutgers ET Friend SH Bernards R A gene-expression signature as a predictor of survival in breast cancer N Engl J Med 2002 347 1999 2009 12490681 10.1056/NEJMoa021967
Ramaswamy S Ross KN Lander ES Golub TR A molecular signature of metastasis in primary solid tumors Nat Genet 2003 33 49 54 12469122 10.1038/ng1060
Lossos IS Czerwinski DK Alizadeh AA Wechser MA Tibshirani R Botstein D Levy R Prediction of survival in diffuse large-B-cell lymphoma based on the expression of six genes N Engl J Med 2004 350 1828 1837 15115829 10.1056/NEJMoa032520
Spentzos D Levine DA Ramoni MF Joseph M Gu X Boyd J Libermann TA Cannistra SA Gene expression signature with independent prognostic significance in epithelial ovarian cancer J Clin Oncol 2004 22 4700 4710 15505275 10.1200/JCO.2004.04.070
Cleator S Ashworth A Molecular profiling of breast cancer: clinical implications Br J Cancer 2004 90 1120 1124 15026788 10.1038/sj.bjc.6601667
Robison JE Perreard L Bernard PS State of the science: molecular classifications of breast cancer for clinical diagnostics Clin Biochem 2004 37 572 578 15234238 10.1016/j.clinbiochem.2004.05.002
Chen CH Generalized association plots: information visualization via iteratively generated correlation matrices. Statistica Sinica 2002 12 7 29
Alter O Brown PO Botstein D Singular value decomposition for genome-wide expression data processing and modeling Proc Natl Acad Sci USA 2000 97 10101 10106 10963673 10.1073/pnas.97.18.10101
Zhang L Yu W He T Yu J Caffrey RE Dalmasso EA Fu S Pham T Mei J Ho JJ Zhang W Lopez P Ho DD Contribution of human alpha-defensin 1, 2, and 3 to the anti-HIV-1 activity of CD8 antiviral factor Science 2002 298 995 1000 12351674 10.1126/science.1076185
Ganz T Defensins: antimicrobial peptides of innate immunity Nat Rev Immunol 2003 3 710 720 12949495 10.1038/nri1180
Rosenberg HF Tenen DG Ackerman SJ Molecular cloning of the human eosinophil-derived neurotoxin: a member of the ribonuclease gene family Proc Natl Acad Sci USA 1989 86 4460 4464 2734298
Domachowske JB Dyer KD Bonville CA Rosenberg HF Recombinant human eosinophil-derived neurotoxin/RNase 2 functions as an effective antiviral agent against respiratory syncytial virus J Infect Dis 1998 177 1458 1464 9607820
Robinson WS A method for chronologically ordering archaeological deposits. American Antiquity 1951 16 191 301
Hurley CB Clustering visualization of multidimensional data J Computational & Graphical Statistics 2004 13 788 806 10.1198/106186004X12425
Pugin J Auckenthaler R Mili N Janssens JP Lew PD Suter PM Diagnosis of ventilator-associated pneumonia by bacteriologic analysis of bronchoscopic and nonbronchoscopic "blind" bronchoalveolar lavage fluid Am Rev Respir Dis 1991 143 1121 1129 2024824
Hampton GM Frierson HF Classifying human cancer by analysis of gene expression Trends Mol Med 2003 9 5 10 12524204 10.1016/S1471-4914(02)00006-0
Wilson CL Ouellette AJ Satchell DP Ayabe T Lopez-Boado YS Stratman JL Hultgren SJ Matrisian LM Parks WC Regulation of intestinal alpha-defensin activation by the metalloproteinase matrilysin in innate host defense Science 1999 286 113 117 10506557 10.1126/science.286.5437.113
Wang YH Lin AS Chao TY Lu SN Liu JW Chen SS Lin MC A cluster of patients with severe acute respiratory syndrome in a chest ward in southern Taiwan Intensive Care Med 2004 30 1228 31. Epub 2004 Apr 23. 15105985 10.1007/s00134-004-2311-8
The World Health Organization:Case Definitions for Surveillance of Severe Acute Respiratory Syndrome (SARS)
Wang TH Lee YS Chen ES Gong WH Chen LK Hsueh DW Wei ML Wang HS Lee YS Establishment of cDNA microarray analysis at the Genomic Medicine Research Core Laboratory (GMRCL) of Chang Gung Memorial Hospital Chang Gung Medical Journal 2004 27 243 260 15239191
Eisen MB Spellman PT Brown PO Botstein D Cluster analysis and display of genome-wide expression patterns Proc Natl Acad Sci USA 1998 95 14863 14868 9843981 10.1073/pnas.95.25.14863
Generalized association plots (GAP)
|
16174304
|
PMC1262710
|
CC BY
|
2021-01-04 16:32:48
|
no
|
BMC Genomics. 2005 Sep 21; 6:132
|
utf-8
|
BMC Genomics
| 2,005 |
10.1186/1471-2164-6-132
|
oa_comm
|
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-1361618536310.1186/1471-2164-6-136Research ArticleMitochondrial-encoded membrane protein transcripts are pyrimidine-rich while soluble protein transcripts and ribosomal RNA are purine-rich Bradshaw Patrick C [email protected] Anand [email protected] David C [email protected] Virginia Bioinformatics Institute, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061, USA2005 26 9 2005 6 136 136 12 5 2005 26 9 2005 Copyright © 2005 Bradshaw et al; licensee BioMed Central Ltd.2005Bradshaw et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Eukaryotic organisms contain mitochondria, organelles capable of producing large amounts of ATP by oxidative phosphorylation. Each cell contains many mitochondria with many copies of mitochondrial DNA in each organelle. The mitochondrial DNA encodes a small but functionally critical portion of the oxidative phosphorylation machinery, a few other species-specific proteins, and the rRNA and tRNA used for the translation of these transcripts. Because the microenvironment of the mitochondrion is unique, mitochondrial genes may be subject to different selectional pressures than those affecting nuclear genes.
Results
From an analysis of the mitochondrial genomes of a wide range of eukaryotic species we show that there are three simple rules for the pyrimidine and purine abundances in mitochondrial DNA transcripts. Mitochondrial membrane protein transcripts are pyrimidine rich, rRNA transcripts are purine-rich and the soluble protein transcripts are purine-rich. The transitions between pyrimidine and purine-rich regions of the genomes are rapid and are easily visible on a pyrimidine-purine walk graph. These rules are followed, with few exceptions, independent of which strand encodes the gene. Despite the robustness of these rules across a diverse set of species, the magnitude of the differences between the pyrimidine and purine content is fairly small. Typically, the mitochondrial membrane protein transcripts have a pyrimidine richness of 56%, the rRNA transcripts are 55% purine, and the soluble protein transcripts are only 53% purine.
Conclusion
The pyrimidine richness of mitochondrial-encoded membrane protein transcripts is partly driven by U nucleotides in the second codon position in all species, which yields hydrophobic amino acids. The purine-richness of soluble protein transcripts is mainly driven by A nucleotides in the first codon position. The purine-richness of rRNA is also due to an abundance of A nucleotides. Possible mechanisms as to how these trends are maintained in mtDNA genomes of such diverse ancestry, size and variability of A-T richness are discussed.
==== Body
Background
Mitochondria are the descendents of an early bacterium that developed a symbiotic relationship with another cell approximately 1.5 billion years ago [1]. Although the mitochondria still contain DNA, the mitochondrial genome has greatly simplified over its long history of symbiosis. Naturally, this simplification in the mitochondrial genome has taken different routes as life diverged into different kingdoms. Vertebrate mitochondrial genomes are among the most compact, gene-rich genomes, while some plant mitochondria have evolved to have a low percentage of coding region similar to that of nuclear DNA [2]. Features of the mitochondrial genomes that have persisted through the divergent evolution of eukaryotic life are likely to be due to fundamental limitations on the variation of that genome. In this paper we discuss three such features that are preserved across eukaryotic species.
Because of the relatively small size of mitochondrial DNA, it is ideally suited for analysis by n-dimensional DNA walks. One dimensional pyrimidine-purine walks were first used to find long-range correlations in nucleotide sequences [3]. Recently multi-fractal walks of mitochondrial DNA were used to find a nonlinear organization in the mitochondrial genome [4]. Combining this information with pyrimidine-purine walks and walks of G-C versus A-T content [5] gives a better understanding of the nucleotide organization of the genome.
Using these techniques we demonstrate certain features of mtDNA sequences which have been preserved by evolution. Greater understanding of the evolutionary selection pressures on mtDNA will allow the construction of more accurate phylogenetic trees based upon mtDNA gene sequences [6,7] as well as a better grasp of the root causes of mitochondrial DNA mutations responsible for many human diseases.
Results
A pyrimidine (C and T) – purine (A and G) walk of the (+) strand of human mtDNA (commonly called the "light" strand in vertebrates) is shown in Figure 1. For each pyrimidine in the sequence a step up is taken and for each purine a step down is taken. In vertebrates, all mitochondrial genes except ND6 and many tRNAs are encoded on the heavy strand. Therefore the mRNA species are predominantly (+) light strand synonymous. The first 3-kilobase section of the human mtDNA encodes two ribosomal RNAs and the pyrimidine-purine walk slopes downward in this region, indicating that the ribosomal RNAs are slightly purine-rich. The remainder of the genome predominately encodes mitochondrial oxidative phosphorylation proteins with small tRNAs interspersed between them. Each mitochondrial protein transcript (except ND6) is pyrimidine-rich giving the remainder of the graph an upward slope. Within this overall rise, small almost flat sections where tRNA genes are located can be seen between the protein coding regions. A particularly large group of these tRNAs is contained in the section of mtDNA around the origin of light strand replication (OL) (shown as an inset to Figure 1). The OL is dramatically clear in this plot as a large run of pyrimidines on one side of the OL followed by a long run of purines on the other side. This section is thought to form a stem-loop structure as an initiation event for light strand DNA synthesis [8,9]. However, we should note that a stem-loop structure does not require a dramatic separation of pyrimidines and purines as is seen here. The ND6 gene has a slightly upward slope in the walk indicating that the human ND6 mRNA is slightly purine-rich in contrast to the other protein-coding mRNAs encoded on the opposite strand.
Figure 1 A pyrimidine-purine walk of human (+ strand) mtDNA. Genes having mRNA synonymous with the (+) strand or (-) strand are indicated by color and also shown on the strand bars below the graph. An inset of a tRNA-containing section of the graph around the origin of light strand replication (OL) is shown.
Clear pyrimidine and purine rich genome segments can also be seen in the mtDNA from other eukaryotic species. The pyrimidine-purine walks of the mitochondrial genome of seven diverse species are shown in Figure 2. In these species many of the genes, the gene order, and the gene distribution over the two DNA strands are different. Figure 2A shows a pyrimidine-purine walk of the mitochondrial genome of the red algae Chondrus crispus (Irish Moss). The genes are color-coded based upon whether they encode membrane proteins, soluble proteins, or rRNA and upon the strand in which they are encoded. Figure 2B shows a mitochondrial genome walk of the red algae, Porphyra purpurea. The similar gene order in these two red algae species (Figures 2A and 2B) gives the walks a similar overall shape. Metazoan mitochondrial genomes such as those shown from Drosophila (Figure 2C) and sea urchin (Figure 2D) are pyrimidine rich overall (+ strands), especially in the sea urchin where all but one of the genes are encoded on the same strand. Plant, fungal, and protist genomes that are gene-poor, which encode oxidative phosphorylation proteins on both strands, and also encode soluble proteins are generally purine-rich. Small genomes (<25 kB), such as that of Schizosaccharomyces pombe (fission yeast) (Figure 2E), are pyrimidine-rich because they are gene-rich and encoded entirely on one strand.
Figure 2 Pyrimidine-purine walks of mitochondrial genome (+) strands of selected species. M, S, and R indicate membrane protein-coding, soluble protein-coding, and RNA-coding segments, respectively. Single tRNA genes are not shown due to their small size, but stretches of 2 or more consecutive tRNAs on the same strand are shown. The coloring scheme for colors not shown in the legend follows that of Figure 1.
No matter which strand encodes the genes, or whether the entire mitochondrial genome is pyrimidine or purine-rich, there are highly conserved features in these walks. Places in the genome where the DNA walk went down were locations where rRNA or soluble proteins were encoded on the (-) (heavy) strand or where membrane proteins were encoded on the (+) (light) strand. Locations of membrane proteins on the (-) strand or rRNA or soluble proteins on the (+) strand were associated with an upward slope in the pyrimidine-purine walk. Exceptions to these rules are indicated on the figure with an asterisk. The most notable exception that we found was the mtDNA from the slime mold Dictyostelium discoideum (Figure 2G), which has a very strand-asymmetric genome, being very purine-rich on the (+) strand (60%). Unlike other species, in Dictyostelium almost all oxidative-phosphorylation membrane-complex transcripts on the (+) strand were purine-rich, just like the rest of its genome. To place the extreme purine richness of Dictyostelium in context, the human mtDNA genome (+) strand is 44 % purine while that of the plant Arabidopsis thaliana is 50 % purine. The purine abundance of the mitochondrial genome of other species is listed in additional file 1: Table S1. Slime molds contain the most purine-rich (+) strand of any of the 23 organisms we examined, while mammals, birds, and a green algae species, Pedinomonas minor, contain the least amount of purine (44%).
We analyzed the pyrimidine and purine content of mitochondrial transcripts from many diverse eukaryotic species (Table 1). Unlike vertebrate mtDNA that lacks genes for soluble proteins, plant, fungi, and protist mtDNA encode genes for many ribosomal proteins and a few other soluble proteins. From this data we defined the following three rules that apply to the pyrimidine-purine richness of mitochondrial transcripts.
Table 1 The number of genes that obey the rules of mitochondrial pyrimidine-purine base composition. All oxidative phosphorylation complex protein genes were included as membrane protein genes. Soluble mitochondrial protein genes included those of ribosomal proteins, maturases and endonucleases from intronic ORFs, and polymerase-like proteins. Unknown ORFs, hypothetical proteins, and proteins of unknown localization were excluded from the analysis. Transcripts that do not follow Rule #1 include almost all Dictyostelium transcripts, Chondrus crispus SDH2, Porphyra purpurea SDH2, COX2, and ymf39, Marchantia polymorpha NAD7 and ATPa, and Arabidopsis NAD7, NAD9, and ATP1. The Chlamydomonas reinhardtii rtl transcript breaks Rule #3.
Organism Name Rule 1 Rule 2 Rule 3 Hypothetical Proteins or Unknown Localization
Genbank Accession Latin Common Holds (Fails) Holds (Fails) Holds (Fails)
[NC_001807] Homo sapiens Human 12 (1) 2 - -
[NC_005089] Mus musculus Mouse 12 (1) 2 - -
[NC_001913] Oryctolagus cuniculus Rabbit 12 (1) 2 - -
[NC_002008] Canis familiaris Dog 13 2 - -
[NC_000845] Sus scrofa Pig 13 2 - -
[NC_001323] Gallus gallus Chicken 13 2 - -
[NC_002784] Dromaius novaehollandiae Emu 13 2 - -
[NC_001573] Xenopus laevis Frog 13 2 - -
[NC_002333] Danio rerio Zebrafish 13 2 - -
[NC_001709] Drosophila melanogaster Fruit fly 13 2 - -
[NC_002074] Rhipicephalus sanguineus Brown dog tick 13 2 - -
[NC_003344] Thyropygus sp. DVL-2001 Giant millipede 13 2 - -
[NC_001453] Strongylocentrot. purpuratus Sea urchin 13 2 - -
[NC_000933] Metridium senile Brown sea anemone 13 2 1 -
[NC_001328] Caenorhabditis elegans Soil nematode 12 1 (1) - -
[NC_001224] Saccharomyces cerevisiae Baker's yeast 7 3 11 1
[NC_001326] Schizosaccharomyces pombe Fission yeast 6 2 - 4
[NC_000895] Dictyostelium discoideum Slime mold 2(13) 2 14 11
[NC_001677] Chondrus crispus Irish moss/Red algae 18(1) 2 (1) 5 1
[NC_002007] Porphyra purpurea Seaweed/Red algae 16(3) 2 6 6
[NC_001638] Chlamydomonas reinhardtii Green algae 7 11(3) 0(1) -
[NC_000892] Pedinomonas minor Green algae 11 3 - -
[NC_001660] Marchantia polymorpha Liverwort 13(2) 3 17 43
[NC_001284] Arabidopsis thaliana Thale cress 19(3) 2 (1) 8 87
Rule 1) Oxidative phosphorylation complex and other membrane protein transcripts are pyrimidine-rich.
Rule 2) Ribosomal RNA is purine-rich.
Rule 3) Soluble protein transcripts are purine-rich.
Table 1 lists the number of genes in each species that follow each rule, along with the number that fail. There were few exceptions to these rules. In some mammals (though not all) the ND6 transcript does not follow rule #1. The main exception for rule #2 is the large ribosomal RNA subunit in C. elegans, which has almost equal numbers of purines and pyrimidines. In other non-animal species the short 5S rRNA sometimes contains more pyrimidines than purines.
We examined the mtDNA from 8 species that encode genes for both soluble and membrane proteins. In Figure 3A we plot the percent pyrimidine in the transcripts versus the frequency at which transcripts of that type (soluble proteins, membrane proteins, or rRNA) occur in the 8 species. The membrane protein transcripts had a distinctive distribution with a peak at around 56 % pyrimidine. The ribosomal RNA and soluble protein transcripts had overlapping distributions with peaks near 45 % and 47 % pyrimidine respectively. These data explain the signals obtained in the pyrimidine-purine walks. It also gives an explanation as to why the soluble protein transcript walks are more variable than the membrane protein transcript walks, since the purine-rich signal is weaker in soluble protein transcripts than is the pyrimidine-rich signal in the membrane protein transcripts. The relative pyrimidine percentage at each codon position in membrane and soluble protein transcripts is shown in Figures 3B and 3C will be discussed later.
Figure 3 Pyrimidine abundance in mitochondrial-encoded rRNA and codon positions in membrane and soluble protein transcripts. (A) Complete transcripts (B) Codon positions in membrane protein transcripts (C) Codon positions in soluble protein transcipts. Mitochondrial genes from Arabidopsis thaliana, Marchantia polymorpha, Chlamydomonas reinhardtii, Chondrus crispus, Porphyra purpurea, Saccharomyces cerevisiae, and Metridium senile were analyzed. Unknown ORFS and hypothetical genes were excluded.
To clearly illustrate the relationship between protein hydrophobicity and the pyrimidine content of the genes, we plot in Figure 4 the percent pyrimidine in the protein transcript versus the grand average of hydropathicity (GRAVY) of the protein for four species with numerous mitochondrial-encoded soluble and membrane protein genes. A higher GRAVY score indicates a higher hydrophobicity of the protein. There was a strong correlation (P < 0.001) between the percent pyrimidine and the hydrophobicity of the encoded protein. This correlation has been shown previously for transcripts of nuclear-encoded proteins [10] and for the second codon position in mitochondrial transcripts from animals and other metazoan mitochondrial genomes that strictly encode membrane proteins [11]. We show that the correlation holds nicely for entire mitochondrial protein transcripts, whether the proteins are soluble or membrane-bound. At high GRAVY scores there is a consistent excursion of membrane proteins from the correlation line. Also, the membrane proteins having low GRAVY scores and low pyrimidine content in the transcripts are likely peripheral membrane proteins. Interestingly, the strong correlation in Figure 4 also holds for Dictyostelium where almost all mitochondrial transcripts are purine-rich.
Figure 4 The correlation between the hydrophobicity of a mitochondrial transcript and its pyrimidine content. GRAVY scores were calculated using the ExPASy ProtParam website. Linear fit P-values were less than 0.001 for all panels. Linear fit R-values were (A) 0.82 (B) 0.75 (C) 0.90, and (D) 0.88. The numbers of membrane, soluble, and unknown protein-coding genes for the species in panels A-C are found in Table 1.
It has been noted that the hydrophobicity of a protein is related to the pyrimidine content of position 2 in the codons of the gene [11,12]. If this is the cause of the pattern that we see in the mitochondrial protein genes, then by splitting the DNA walk into three separate walks, one for each codon position, we would expect that the walk using codon position 2 would be responsible for the signal, while the walks for codon positions 1 and 3 might be random. Fig. 5 shows a pyrimidine-purine walk of each codon position of the human COX1 membrane protein transcript and the Chondrus crispus S12 soluble ribosomal protein transcript. For comparison, the pyrimidine-purine walk of the human 16S ribosomal RNA is also shown. The base composition of mitochondrial genome sections encoding rRNA and tRNA from other species is given in additional file 1: Table S1. These walks are given as examples to show the uniformity of the signal along the length of the gene. The mitochondrial-encoded transcripts from other species have a similar pattern of pyrimidine-richness in the three codon positions (see Figures 3B and 3C and Table 2). In mtDNA-encoded membrane-protein transcripts, codon position 2 contains the most pyrimidines, as predicted (Figure 3B). However, codon position 3 also contributes slightly to the pyrimidine-rich signal while codon position 1 is often slightly purine-rich. In the soluble ribosomal protein transcript the purine-rich signal is driven mainly by codon position 1, while codon positions 2 and 3 contribute only slightly (see also Figure 3B). The eight known mtDNA-encoded soluble protein transcripts from Arabidopsis give a similar purine-rich signal in pyrimidine-purine walks (see additional file 1: Figure S1). Even the signals of the individual codon positions follow the same trends in all eight genes.
Figure 5 Pyrimidine-purine codon position walks of select mitochondrial-encoded transcripts. (A) Membrane protein transcript, human COX1 (C) Soluble protein transcript, Chondrus crispus ribosomal protein S12. For each codon position step, the x-axis was incremented by 3 for comparison to the complete transcript. A pyrimidine-purine walk of (B) human 16S ribosomal RNA is also shown for comparison.
Table 2 Base composition at each codon position in mtDNA-encoded membrane and soluble protein-coding transcripts. Analysis was performed on transcripts from humans and the species from Table 1 that encode soluble proteins in mtDNA.
Membrane Proteins Soluble Proteins
%A %G %C %T %Pu %AT %A %G %C %T %Pu %AT
Saccharomyces cerevisiae
Codon pos. 1 28.1 26.4 10.3 35.2 54.5 63.3 42.5 20 8.5 29.1 62.4 71.6
Codon pos. 2 21.3 12.6 22 44.1 33.9 65.3 36.4 13.3 15 35.3 49.6 71.6
Codon pos. 3 42.9 5.14 7.49 44.5 48 87.4 42.5 4.57 3.9 49 47.1 91.5
All 30.7 14.7 13.3 41.3 45.5 72 40.5 12.6 9.2 37.8 53.1 78.3
Arabidopsis thaliana
Codon pos. 1 26 27.6 20.5 26 53.6 51.9 33 26 22 19.1 59.1 52.1
Codon pos. 2 21.5 17.7 23.7 37.2 39.2 58.6 29 22 22 27.4 51 56.5
Codon pos. 3 26.5 17.8 17.5 38.2 44.3 64.8 29.8 22.2 22 26.3 51.9 56.1
All 24.7 21 20.5 33.8 45.7 58.4 30.6 23.4 22 24.3 54 54.9
Porphyra purpurea
Codon pos. 1 30.8 23.4 14.7 31.1 54.2 61.9 41.8 21.2 14 23.1 63 64.9
Codon pos. 2 21.5 14.6 20.9 43 36.1 64.4 37.6 14.9 18 29.4 52.5 67.1
Codon pos. 3 34.9 11.4 15.4 38.3 46.3 73.2 39.9 12.2 14 34 52.1 73.9
All 29.1 16.5 17 37.5 45.5 66.5 39.8 16.1 15 28.8 55.9 68.6
Dictyostelium discoideum
Codon pos. 1 35.9 28.3 7.92 27.9 64.2 63.8 45.4 23.6 11 20.4 69 65.8
Codon pos. 2 25.7 15.3 15.9 43.1 41 68.8 39.4 17.7 13 30.1 57.2 69.6
Codon pos. 3 57.8 7.24 4.83 30.2 65 87.9 61.2 10.1 4.1 24.6 71.3 85.8
All 39.8 16.9 9.55 33.7 56.7 73.5 48.7 17.2 9.1 25 65.8 73.7
Chondrus crispus
Codon pos. 1 30.5 20.3 13.8 35.4 50.8 65.9 41.8 16.1 14 28.5 57.9 70.3
Codon pos. 2 21.4 13.5 19.8 45.3 34.9 66.7 35.7 13.9 17 33.6 49.6 69.3
Codon pos. 3 35.9 7.92 7.75 48.4 43.8 84.3 46 6.81 7.2 40 52.8 86
All 29.3 13.9 13.8 43.1 43.2 72.3 41.2 12.3 13 34 53.4 75.2
Marchantia polymorpha
Codon pos. 1 27 28.5 16 28.6 55.5 55.5 35.2 23 19 22.9 58.2 58.1
Codon pos. 2 20.8 16.5 21.1 41.6 37.3 62.4 32.7 19.6 19 28.5 52.3 61.1
Codon pos. 3 28.5 14.6 13.9 43 43.1 71.5 36 15.3 13 35.3 51.3 71.3
All 25.4 19.8 17 37.7 45.3 63.2 34.6 19.3 17 28.9 53.9 63.5
Metridium senile
Codon pos. 1 21.7 18.5 19.8 40 40.2 61.7 30.4 28.1 17 24.6 58.5 54.9
Codon pos. 2 25.9 17.7 16.4 40 43.6 65.9 31.7 20.1 17 30.8 51.8 62.5
Codon pos. 3 26.4 25.1 14.7 33.8 51.5 60.2 39.3 13.4 11 36.6 52.7 75.9
All 24.6 20.4 17 37.9 45.1 62.6 33.8 20.5 15 30.7 54.3 64.4
Chlamydomonas reinhardtii
Codon pos. 1 24.7 30.5 17.5 27.2 55.2 51.9 25.2 24.4 32 18.7 49.6 43.9
Codon pos. 2 16.4 19.3 21.5 42.8 35.8 59.2 34.1 14.4 19 32.8 48.5 66.9
Codon pos. 3 15.5 19.8 26.3 38.4 35.3 53.9 21.7 23.6 24 31.2 45.3 52.8
All 18.9 23.2 21.8 36.1 42.1 55 27 20.8 25 27.6 47.8 54.6
Homo sapiens
Codon pos. 1 29.5 19.6 26.9 24 49.1 53.5
Codon pos. 2 22.8 11.4 29.9 36 34.2 58.8
Codon pos. 3 34.3 8.38 39 18.4 42.7 52.7
All 28.9 13.1 31.9 26.1 42 55
As an example of the robustness of this signal, pyrimidine-purine walks of each codon position from the other 12 human mtDNA protein-coding genes are shown in Figure 6. From the linearity of the walk of the entire genome in Figure 1, it is clear that the signal strength is almost constant through all protein-coding genes. The pyrimidine-rich signal is driven by codon position 2 in almost all cases, with position 3 contributing modestly and codon position 1 not contributing to an appreciable extent. The conservation of this pattern through the vast majority of the mitochondrial transcripts indicates the strong selective pressure for this signature. Unlike the other transcripts, the ND6 transcript is purine-rich, but it is also the only transcript encoded on the (+) light strand of mtDNA. So there does appear to be a strand-specific selective force present as well, in human mtDNA. However, the ND6 transcript is pyrimidine-rich in other mammals even though it is encoded on the (+) light strand (see Table 1). The percent occurrence of each individual nucleotide species in each codon position in mitochondrial genomes from many eukaryotic species is given in Table 2. It is shown that C more than T in codon positions 1 and 3 drives the pyrimidine-rich signal in human membrane protein transcripts, while T more than C at codon position 2 also contributes. The constancy of this pattern throughout the genes as well as the overall abundance of A over G and C over T can be observed in 2-dimensional walks of the individual codon positions and entire transcripts (see additional file 1: Figure S2). The pyrimidine-rich signal is driven by C over T only in birds, reptiles, and some mammalian species. In all other species examined, T drives the signal from all 3 codon positions.
Figure 6 Pyrimidine-purine codon position walks of human mtDNA-encoded protein transcripts. The COX1 transcript walk, absent in this figure, is shown in Figure 5A. All 13 genes have similar patterns in the codon positions of the pyrimidine-purine walks.
Discussion
Mitochondrial-encoded membrane protein transcripts are pyrimidine-rich
Protein transcripts with an abundance of U (T) in the second codon position encode hydrophobic amino acids [10,13] that tend to form membrane-spanning alpha helices [14] or beta strands [12]. This is likely the most important factor that contributes to the relative pyrimidine-richness of mitochondrial membrane complex transcripts. However it does not explain the entire signal in humans where large quantities of C in the third codon position also play an important role. In fact, the pyrimidine-rich signal in humans is mainly driven by C in the third codon position (Table 2). The signal is also partially driven by the lack of G in the transcripts. Mitochondrial DNA is replicated by a strand asymmetric mechanism [15] that is likely responsible for the unequal strand distribution of G nucleotides [16]. G is the most easily oxidized base, forming 8-hydroxy guanine [17]. A low percentage of G in the vertebrate mitochondrial transcripts has been hypothesized to contribute to mRNA stability in the oxidative environment of the matrix space [18]. However, we must emphasize that the low G abundance in the light strand in vertebrates is not the primary source of the pyrimidine richness in these transcripts, because membrane protein transcripts are also pyrimidine-rich in species where no mitochondrial strand asymmetry in G content is present (see Table 2).
The relative contribution of C versus T, and A versus G throughout the transcripts can be seen in the 2-dimensional walks of the genes using A-G on one axis and C-T on the other (additional file 1: Figure S2). The percentage of C vs. T has been shown to vary greatly in different mammalian lineages [7]. The greater abundance of C over T on the mitochondrial light (+) strand first appears evolutionarily in reptiles and is accompanied by a slightly more G-C rich mitochondrial genome (37 % in Xenopus compared to 44 % in humans (additional file 1: Table S1)). The development of GC-rich isochores also first occurred in the nuclear DNA of reptiles and may be one of the factors allowing the evolution of warm-blooded birds and mammals [19]. Based on the data presented here, some of the same selective pressures may be affecting both the nuclear and mitochondrial genomes.
Mitochondrial-encoded soluble protein transcripts are purine-rich
It has been suggested that purine-loading of transcripts may have evolved to prevent detrimental RNA-RNA interactions [20]. However this hypothesis does not explain the codon-specific pattern of purine-richness in mitochondrial soluble protein-coding transcripts. An A in the second codon position of nuclear-encoded transcripts often encodes relatively hydrophilic amino acids [12]. These amino acids have been shown to be abundant in the aperiodic secondary structure of soluble proteins. However, in the mitochondrial genome, purines (see Figure 3C), specifically A, in the first codon position (not the second) mainly drives the purine-richness of soluble proteins (see Table 2). To the best of our knowledge, purine abundance in the first codon position has not previously been associated with the hydrophilic nature of soluble proteins, even though this signature does occur in the vast majority of nuclear-encoded transcripts [21,22]. One hypothesis that could be tested is that ribosomes translate more efficiently when purines are present at the first codon position. Additionally, increased levels of specific tRNAs in the mitochondrial matrix may select for such a trend. However A in the second codon position also contributes to the signal. A decrease in T nucleotides accompanies the increase in A nucleotides in both positions. The result of this A for T substitution in the first two codon positions is the greater abundance of the hydrophilic amino acids lysine and asparagine (codon AAX) in soluble mitochondrial proteins and the decreased abundance of the hydrophobic amino acids phenylalanine and leucine (codons UUX and CUX). In fact much of the purine-rich signal in the soluble proteins is due to the 3–4 fold increase in positively charged lysine residues in these proteins compared to membrane proteins (data not shown). Mitochondrial ribosomal proteins use these residues to bind the negatively charged phosphate backbone of ribosomal RNA [23,24].
Mitochondrial ribosomal RNA is purine-rich
The selective pressure that maintains the slight purine richness of mitochondrial ribosomal RNA is not entirely clear [25,26]. It is known that ribosomal RNA interacts with ribosomal protein through hydrophobic interactions of unpaired A residues in the RNA loop regions with hydrophobic protein side chains [27,28]. Purine nucleotides are more hydrophobic than pyrimidine nucleotides [13,29]. Therefore this slight purine abundance in the loop regions may be conserved to facilitate this interaction. Mitochondrial introns are also purine-rich (additional file 1: Figure S3), likely conserving a hydrophobic interaction between splicing proteins and the loop structures in the RNA.
It is difficult to hypothesize how such small magnitutudes of purine and pyrimidine base skew can be conserved over the billion years of mitochondrial evolution. Skewed ribonucleoside triphosphate pools (highest in ATP) [30] may select for a high level of A (purine) in ribosomal RNA and soluble protein transcripts while the need for hydrophobicity in membrane proteins may overcome this pressure, resulting in pyrimidine-rich transcripts. The selective pressure to contain charged hydrophilic amino acids in soluble proteins may also contribute to the maintenance of the purine-rich signal in soluble protein transcripts as well as the abundance of hydrophobic A residues in the loop regions of ribosomal RNA. A better understanding of these mitochondrial selection pressures may be gained in the future by comparing the pyrimidine-purine transcript asymmetries with that of non-coding mtDNA.
Methods
Mitochondrial gene sequences, amino acid sequences and genomes were downloaded from the NCBI website. Java (JDK 1.50) programs were written to analyze the sequences. The programs or software details are available from the authors upon request. Other websites such as the OGRe database of mitochondrial genomes [31] also allow analysis and graphing of base composition at the codon positions as well. The calculations were performed on a 2.8 GHz desktop computer and typically took less than a few seconds to run. The gene sequences analyzed are the mRNA synonymous sequences as available in PubMed. The mitochondrial genome strand labels (+) and (-) follow PubMed convention.
Authors' contributions
PB generated the data, constructed Figures 3, 4, 5, 6 and drafted the manuscript. AR wrote the majority of the computer code for sequence analysis, generated the data for and constructed Figures 1 and 2. DC designed and supervised the study.
Supplementary Material
Additional File 1
Table S1 Base percentages in the entire mitochondrial genome and in just protein-coding or RNA-coding sections. Figure S1 The pyrimidine-purine walk of each codon position of eight soluble mitochondrial-encoded protein transcripts from Arabidopsis thaliana. Figure S2 The 2-dimensional A-G and T-C walks of human mtDNA-encoded transcripts. Figure S3 Pyrimidine-purine walks of unspliced mitochondrial-encoded transcripts from Arabidopsis thaliana and Marchantia polymorpha. The introns in these transcripts do not encode known proteins.
Click here for file
Acknowledgements
The authors thank the Virginia Bioinformatics Institute for financial support and Harsha Rajasimha for many helpful discussions.
==== Refs
Sicheritz-Ponten T Kurland CG Andersson SG A phylogenetic analysis of the cytochrome b and cytochrome c oxidase I genes supports an origin of mitochondria from within the Rickettsiaceae Biochim Biophys Acta 1998 1365 545 551 9711305
Gray MW Burger G Lang BF Mitochondrial Evolution Science 1999 283 1476 1481 10066161 10.1126/science.283.5407.1476
Peng CK Buldyrev SV Goldberger AL Havlin S Sciortino F Simons M Stanley HE Long-range correlations in nucleotide sequences Nature 1992 356 168 170 1301010 10.1038/356168a0
Oiwa NN Glazier JA Self-similar mitochondrial DNA Cell Biochem Biophys 2004 41 41 62 15371639 10.1385/CBB:41:1:041
Grigoriev A Analyzing genomes with cumulative skew diagrams Nucl Acids Res 1998 26 2286 2290 9580676 10.1093/nar/26.10.2286
Krishnan NM Seligmann H Stewart CB De Koning AP Pollock DD Ancestral sequence reconstruction in primate mitochondrial DNA: compositional bias and effect on functional inference Mol Biol Evol 2004 21 1871 1883 15229290 10.1093/molbev/msh198
Gibson A Gowri-Shankar V Higgs PG Rattray M A comprehensive analysis of mammalian mitochondrial genome base composition and improved phylogenetic methods Mol Biol Evol 2005 22 251 264 15483324 10.1093/molbev/msi012
Hixson JE Wong TW Clayton DA Both the conserved stem-loop and divergent 5'-flanking sequences are required for initiation at the human mitochondrial origin of light-strand DNA replication J Biol Chem 1986 261 2384 2390 3944140
Wong TW Clayton DA In vitro replication of human mitochondrial DNA: accurate initiation at the origin of light-strand synthesis Cell 1985 42 951 958 2996785 10.1016/0092-8674(85)90291-0
Biro JC Benyo B Sansom C Szlavecz A Fordos G Micsik T Benyo Z A common periodic table of codons and amino acids Biochem Biophys Res Commun 2003 306 408 415 12804578 10.1016/S0006-291X(03)00974-4
Naylor GJ Collins TM Brown WM Hydrophobicity and phylogeny Nature 1995 373 565 566 7854410 10.1038/373565b0
Chiusano ML Alvarez-Valin F Di Giulio M D'Onofrio G Ammirato G Colonna G Bernardi G Second codon positions of genes and the secondary structures of proteins. Relationships and implications for the origin of the genetic code Gene 2000 261 63 69 11164038 10.1016/S0378-1119(00)00521-7
Woese CR Dugre DH Saxinger WC Dugre SA The molecular basis for the genetic code Proc Natl Acad Sci U S A 1966 55 966 974 5219702
Lesnik T Reiss C A method to detect transmembrane helical segments at the nucleotide level Biochem Mol Biol Int 1996 38 937 955 9132163
Clayton DA Replication of animal mitochondrial DNA Cell 1982 28 693 705 6178513 10.1016/0092-8674(82)90049-6
Saccone C De Giorgi C Gissi C Pesole G Reyes A Evolutionary genomics in Metazoa: the mitochondrial DNA as a model system Gene 1999 238 195 209 10570997 10.1016/S0378-1119(99)00270-X
Ames BN Shigenaga MK Hagen TM Mitochondrial decay in aging Biochim Biophys Acta 1995 1271 165 170 7599204
Reyes A Gissi C Pesole G Saccone C Asymmetrical directional mutation pressure in the mitochondrial genome of mammals Mol Biol Evol 1998 15 957 966 9718723
Bernardi G Isochores and the evolutionary genomics of vertebrates Gene 2000 241 3 17 10607893 10.1016/S0378-1119(99)00485-0
Lao PJ Forsdyke DR Thermophilic bacteria strictly obey Szybalski's transcription direction rule and politely purine-load RNAs with both adenine and guanine Genome Research 2000 10 228 236 10673280 10.1101/gr.10.2.228
Zhang CT Chou KC Graphic analysis of codon usage strategy in 1490 human proteins J Protein Chem 1993 12 329 335 8397791 10.1007/BF01028195
Li W Statistical properties of open reading frames in complete genome sequences Comput Chem 1999 23 283 301 10404621 10.1016/S0097-8485(99)00014-5
O'Brien TW Properties of human mitochondrial ribosomes IUBMB Life 2003 55 505 513 14658756
Klein DJ Moore PB Steitz TA The roles of ribosomal proteins in the structure assembly, and evolution of the large ribosomal subunit J Mol Biol 2004 340 141 177 15184028 10.1016/j.jmb.2004.03.076
Wang H Hickey DA Evidence for strong selective constraint acting on the nucleotide composition of 16S ribosomal RNA genes Nucleic Acids Research 2002 30 2501 2507 12034839 10.1093/nar/30.11.2501
Guy L Roten CAH Genometric analyses of the organization of circular chromosomes: a universal pressure determines the direction of ribosomal RNA genes transcription relative to chromosome replication Gene 2004 340 45 52 15556293 10.1016/j.gene.2004.06.056
Gutell RR Weiser B Woese CR Noller HF Comparative anatomy of 16-S-like ribosomal RNA Prog Nucleic Acid Res Mol Biol 1985 32 155 216 3911275
Springer MS Douzery E Secondary structure and patterns of evolution among mammalian mitochondrial 12S rRNA molecules J Mol Evol 1996 43 357 373 8798341
Lacey JCJ Mullins DWJ Experimental studies related to the origin of the genetic code and the process of protein synthesis--a review Orig Life 1983 13 3 42 6350974 10.1007/BF00928761
Xia X Maximizing Transcription Efficiency Causes Codon Usage Bias Genetics 1996 144 1309 1320 8913770
OGRe Organellar Genome Retrieval System
|
16185363
|
PMC1262711
|
CC BY
|
2021-01-04 16:32:46
|
no
|
BMC Genomics. 2005 Sep 26; 6:136
|
utf-8
|
BMC Genomics
| 2,005 |
10.1186/1471-2164-6-136
|
oa_comm
|
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-1371619119210.1186/1471-2164-6-137Research ArticleA TNF-induced gene expression program under oscillatory NF-κB control Tian Bing [email protected] David E [email protected] Allan R [email protected] Department of Medicine, The University of Texas Medical Branch, 301 University Blvd., Galveston, Texas 77555-1060, USA2 Sealy Center for Molecular Sciences, The University of Texas Medical Branch, 301 University Blvd., Galveston, Texas 77555-1060, USA2005 28 9 2005 6 137 137 4 5 2005 28 9 2005 Copyright © 2005 Tian et al; licensee BioMed Central Ltd.2005Tian et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The cytokine tumor necrosis factor (TNF) initiates tissue inflammation, a process mediated by the NF-κB transcription factor. In response to TNF, latent cytoplasmic NF-κB is activated, enters the nucleus, and induces expression of inflammatory and anti-apoptotic gene expression programs. Recently it has been shown that NF-κB displays two distinct activation modes, monophasic and oscillatory, depending on stimulus duration. Characterization of temporal expression patterns for the NF-κB network and determination of those genes under monophasic- or oscillatory control has not been experimentally addressed.
Results
To identify the kinetics of NF-κB-dependent gene expression and determine whether these two types of NF-κB translocation modes control distinct gene programs, a detailed kinetic analysis of a validated microarray data set was performed on 74 unique NF-κB-dependent genes in response to TNF. Hierarchical clustering identified distinct expression profiles termed the "Early", "Middle", "Late" response groups, peaking 1, 3, and 6 h after stimulation, respectively. These expression patterns were validated by Quantitative Real Time PCR (Q-RT-PCR) and NF-κB binding was demonstrated by chromatin immunoprecipitation (ChIP) assays. Each response group was mapped to its molecular function; this analysis indicated that the Early group encodes cytokines or negative regulators of the IKK-NF-κB pathway, and the Late group encodes cell surface receptors, adhesion molecules and signal adapters. That similar coordinated sequential cascades of gene expression were also seen in response to stimulation by the cytokine IL-1, and expression patterns observed in MRC-5 fibroblasts indicated that the epithelial NF-κB program is relatively stimulus- and cell type-independent. Bioinformatic analysis of the Early and Late gene promoters indicates that although both groups contain similar patterns of NF-κB-binding sites, only the Early gene promoters contain NF-κB-binding sites located in phylogenetically conserved domains. Stimulation protocols designed to produce either monophasic or oscillatory NF-κB activation modes showed that the oscillatory mode is required only for expression of the Late genes.
Conclusion
This analysis provides important insights into the TNF-regulated genetic response program in epithelial cells, where NF-κB controls sequential expression patterns of functionally distinct genes that depend on its oscillatory activation mode.
==== Body
Background
Tumor necrosis factor (TNFα, TNF ligand superfamily 2 [TNFSF2]) is a prototypical inflammatory and immunomodulatory cytokine inducibly expressed by activated macrophages, monocytes, neutrophils, T-cells and NK-cells [1]. TNFα is a central mediator of the host inflammatory response by its ability to activate adhesion molecule expression, enhance leukocyte trafficking, and affect the expression of secondary cytokine cascades controlling leukocyte recruitment and activation [1,2]. TNF signaling is mediated by binding and aggregating single-pass type I transmembrane receptors (TNFR-I, ref. [3]) that then serve as an anchor to recruit signaling proteins binding to the death domains on the cytoplasmic receptor tails. Upon assembly of this submembranous complex, two major downstream signaling pathways are activated; these are the jun NH2 terminal kinase -activating protein-1- and the IκB Kinase (IKK)-Nuclear Factor-κB (NF-κB) pathways [4,5].
Nuclear Factor-κB (NF-κB) is a latent cytoplasmic transcription factor maintained in a cytoplasmic location by binding the IκB inhibitors, proteins that bind and specifically inactivate it by masking its nuclear localization sequence, thereby preventing its nuclear entry [6]. NF-κB is activated by TNF signaling pathway indirectly as a result of targeted IκB proteolysis (reviewed in ref. [7]). Signal-induced IκB proteolysis is mediated by activation of the multiprotein cytoplasmic IKK (a.k.a., the "signalsome"ref. [8]), a kinase that phosphorylates IκB specifically in its NH2-regulatory domain, making it a substrate for proteolysis through the 26S proteasome and calpain pathways [8,9].
As a result, liberated NF-κB rapidly enters the nucleus to activate target gene expression by formating a nucleoprotein complex with chromatin-remodeling proteins, kinases, and other transcription factors [10]. Recent single cell fluorescence imaging experiments have shown that TNF can induce two distinct modes of NF-κB activation patterns [11]. In the monophasic mode, the result of a brief TNF stimulation, NF-κB enters the nucleus and induces the expression of IκB inhibitory proteins whose resynthesis redistributes NF-κB back into the cytoplasm, restoring cellular homeostasis [12,13]. Conversely, oscillatory NF-κB activation, a result of tonic TNF stimulation produces prolonged IKK activation and continued IκB proteolysis, results in repeated rounds of NF-κB translocation and cytoplasmic recapture [11]. This latter activation profile is characterized by a series of asynchronous, damped oscillations of nuclear NF-κB [14]. These findings explain the biphasic pattern of nuclear NF-κB binding that has been observed in response to tonic TNF stimulation in a number of distinct cell types [15,16], where the initial oscillation is observed due to stimulus-induced synchrony in the cell population, but lost afterwards because subsequent oscillations are asynchronous and appear damped in the population [14]. Whether these two modes of NF-κB activation produce distinct genetic programs has not been systematically studied.
Although a body of isolated work has reported that NF-κB controls expression of acute-phase reactants [17], cytokines [18], anti-apoptotic proteins [19], and autoregulators of the IKK-NF-κB pathway [12,13,20], the full spectrum of NF-κB-dependent genes are only beginning to be systematically understood [21]. In this study, we analyzed and validated a microarray time series experiment of cells expressing a regulated NF-κB dominant-negative inhibitor in response to TNF. From this data set, we have previously systematically identified known and novel NF-κB-dependent genes [21]. Because these represent a time series experiment, the data may contain genes that are under direct or indirect NF-κB control. That these "NF-κB-dependent" genes were directly controlled by NF-κB was verified by satisfying a series of experimental validation experiments: 1. Ectopic expression of constitutively active NF-κB/Rel A transactivated the endogenous "NF-κB-dependent" genes in the absence of TNF stimulation; 2. TNF induced the "NF-κB-dependent" genes in the absence of new protein synthesis; 3. NF-κB sites were computationally identified and confirmed by EMSA; and, 4. Chromatin immunoprecipitation (ChIP) assays showed the endogenous "NF-κB-dependent" promoters bound NF-κB/Rel A in TNF stimulated cells [21]. Based on these observations, we concluded that this was a robust dataset containing genes directly under NF-κB control.
Here we performed a kinetic analysis of the time series data set where, strikingly, four distinct kinetic groups were identified by cluster analysis. Gene Ontology and Ingenuity pathway analysis show that these response groups encode distinct biological functions from one another, with the Early group being composed of families of secreted cytokine/chemokines and the Late group being composed of cell surface receptors and adhesion molecules. Stimulation experiments producing monophasic or oscillatory NF-κB activation modes show that the oscillatory mode is required for Late gene expression. These data provide major new insights into the coordinated NF-κB response program to inflammatory stimuli, where the cellular response is dictated by the mode in which NF-κB is activated.
Results
We have previously isolated and characterized HeLatTA/FLAG-IκBα Mut, a clonal cell line expressing a tetracycline-regulated NF-κB dominant-negative inhibitor [21,22]. When doxycycline (Dox) is present in tissue culture medium, tTA is inactivated, and FLAG-IκBα Mut levels are barely detectable by Western immunoblot, resulting in a wild type phenotype, with normal levels of activated NF-κB in the nucleus being produced after stimulation [21,22]. Conversely, upon Dox withdrawal, tTA is activated, and FLAG-IκBα Mut expression occurs at similar levels to endogenous IκBα [22]. These levels of FLAG-IκBα Mut expression are sufficient to completely inhibit NF-κB translocation and target gene expression [21,22]. HeLatTA/FLAG-IκBα Mut cells were plated in parallel cultures in the absence or presence of Dox (2 μg/ml), and each group stimulated tonically with TNFα to induce NF-κB activation in the oscillatory mode [11]. RNA was then subjected to high density oligonucleotide microarray analysis. Reanalysis of the raw data set was performed using less stringent statistical filters to more fully identify the spectrum of biological functions under NF-κB control, where we identified 74 probe sets (Figure 1). From these, the scaled signal intensities were subjected to hierarchical clustering to identify coregulated genes, identifying 5 expression groups based on their kinetics of expression (Figure 2a). As an indicator of reproducibility, we noted that redundant probe sets generally clustered with one another, where multiple NF-κB2 and IL-8 probesets group together (Figure 2a). These findings indicated that the clustering analysis is robust, grouping probe sets representing the same genes based on similar expression patterns. Further inspection of the hierarchical clustering results indicates that TNF induces expression of five distinct groups: 1. "Early" genes whose expression profiles peak at 1 h and less; 2. "Middle" genes whose expression profiles peak at 3 h, falling thereafter; 3. "Late" genes whose expression profiles begin to peak at 6 h and later; 4. "Biphasic", genes whose expression profiles peak rapidly at 1 h, fall at 3 h, and peak again at 6 h; and, 5. "Paradoxical" genes whose expression is not significantly altered by TNF in the wild type cells, but whose expression are induced by TNF in the cells lacking NF-κB signaling (Figure 2a). Further, NF-κB dependence for these probe sets is seen by comparing the heat map profiles for each probe set obtained in the presence of Dox vs the profile obtained in the absence of Dox (Figure 2a). For example, the strong induction of gene expression in "Middle" genes at 3 h in the presence of Dox is not seen in the absence of Dox. Similar findings are made for the probe sets in the "Early", "Late", and "Biphasic" genes.
Figure 1 Schematic diagram of microarray data analysis. HeLaTetO-FLAG-IκBα Mut cells were plated in parallel into cultures in the absence or presence of Dox (2 μg/ml). After four days, cells were stimulated without (0 h) or with rhTNFα (25 ng/ml) at 6 h, 3 h, and 1 h prior to simultaneous harvest for RNA extraction. Experiments were conducted four independent times. Data sets were scaled for comparison. NF-κB dependent genes were identified using 2-way ANOVA where Dox treatment and TNF treatment were considered independent variables. Those changed by Dox treatment at a p-values [Pr(F)< 0.01] were then filtered for 3-fold change at any point during the experiment (signal intensity with NF-κB vs signal intensity without NF-κB).
Figure 2 Temporal Cascades of NF-κB Regulated Genes. (a) The Signal Intensity values from 74 probe sets identified as being NF-κB dependent were Z-score normalized and subjected to hierarchical clustering. Red corresponds Z > +2.5, green indicates Z < 0, and black indicates Z > 0.5. Expression groups are indicated at right by vertical line. (b) Distinct Expression Profiles. The normalized SI measurements for each of the genes in Clusters I–IV are presented as a percentage of the maximum value for any point across the stimulation.
To rigorously compare the expression profiles of the major clusters, the scaled and normalized hybridization intensities were retrieved for each gene and plotted by group as a percentage of each gene's maximal expression value during the time course (because of the limited number of probe sets in the Biphasic group, these were excluded from subsequent analysis). As seen in Figure 2b, as a group, the Early genes had ~ 10 % maximal expression at time 0, and rapidly peaked at 100% maximal expression by 1 h, falling to ~ 50 % expression at 3 h. The Middle and Late groups tended to have higher basal expression relative to their maximal induction. The genes within the Late group were tightly synchronized showing maximal expression at 6 h. As expected, the Paradoxical group showed no significant induction by TNF in the presence of Dox, but their expression increased 2-fold in the cells stimulated in its absence. As a test whether the TNF-induced profile grouping was statistically significant, the paired two-tailed Z-Test statistic was calculated to determine whether these three expression groups came from different populations during the TNF response [see Additional file 1]. We found that the P values for the two-tailed test statistic for Early gene group is significantly different than the profiles of the Middle or Late genes at all times of TNF stimulation, indicating that they come from a distinct population than the members of the Middle or Late genes. Similarly, the expression profiles of the Middle genes are different from the Late genes also at 1, 3, and 6 h of stimulation. Together this analysis indicated that the Early, Middle and Late expression groups are separable populations with distinct gene expression patterns.
To validate the gene expression kinetics and confirm their NF-κB dependence, Q-RT-PCR assays were developed for representative members of the Early (IL-6, IL-8, TNFAIP3/A20) and Late (NF-κB-2, Naf-1 and TRAF-1) genes. We then used these assays to measure mRNA changes in tonic TNFα-stimulated in HeLatTA/FLAG-IκBα Mut cells cultured in the absence or presence of Dox. A rapid induction of mRNA transcripts was observed within 1 h for the selected Early genes (Figure 3a). Here, IL-6 mRNA abundance peaked at 48 -fold within 0.5 h, whereas IL-8 and TNFAIP3/A20 peaked at 900-fold and 125-fold at 1 h, respectively. All mRNA signals then rapidly fell to < 30% of the maximal signal at 3 h of stimulation (Figure 3a). In addition, in the HeLatTA/FLAG-IκBα Mut cells cultured in the absence of Dox, mRNA expression for all of these genes were significantly inhibited, with IL-6 being induced no more than 5- fold, and no detectable induction was seen for IL-8 and TNFAIP3/A20 (Figure 3a). In contrast, mRNA transcript abundance for the Late gene group peaked 6 h after TNFα stimulation, with the exception of Naf-1, whose mRNA abundance continued to increase until 9 h (Figure 3b). Like the Early genes, TNFα-induced expression of the Late genes was also significantly inhibited in cells cultured in the absence of Dox.
Figure 3 Validation of expression profiles and NF-κB dependence. (a) Early gene profiles. HeLatTA/FLAG-IκBα Mut cells were plated in parallel in the absence or presence of Dox (2 μg/ml) and stimulated with rhTNFα. Changes in mRNA abundance (normalized by 18S) determined by Q-RT- PCR from total RNA. For each of the indicated mRNA transcripts, values are expressed as fold change relative to unstimulated cells and plotted on a logarithmic scale. +/-Dox, data obtained from cells cultured with or without Dox. (b) Late gene profiles. Experiment and data analysis are as in Figure 3a. (c) ChIP for NF-κB subunit binding to Early Gene promoters. ChIP was performed on control or TNFα-stimulated (30 min, 20 ng/ml) HeLa cells using the antibodies indicated at left. Shown is an ethidium-bromide stained agarose gel of the PCR products performed under linear amplification conditions. The target gene is indicated at the bottom. NC, negative control reaction (no template is added to the PCR reaction); PC, positive control reaction (25 ng of genomic DNA is used as template in PCR). (d) ChIP for NF-κB subunit binding to Late Gene promoters. ChIP was performed on HeLa cells stimulated as in Figure 3c.
We next asked whether TNF-induced specific NF-κB subunit binding to endogenous target genes. For this purpose, Chromatin immunoprecipitation (ChIP) assays were performed on representative members of the Early (IL-8, IL-6) and Late (Naf-1, NF-κB2) genes. In this assay, control or TNF-stimulated cells are exposed to protein-DNA cross-linking reagents to covalently stabilize the chromatin. The soluble chromatin is then extracted, sheared, and the target protein specifically immuno-precipitated (along with its associated DNA). After elution of the DNA, the crosslinks are reversed, and the presence of specific genes detected by PCR. As seen in Figure 3c, TNFα treatment strongly induced NF-κB/RelA subunit binding the IL-8 gene. By contrast, TNF induced only weak c-Rel binding to IL-8. For the DNA binding subunits, although NF-κB1 binding is detectable on IL-8 in the absence of stimulation (compare signal in absence of TNF vs that produced by IgG, Figure 3c), its levels increase in response to TNF treatment. By contrast, no NF-κB2 binding is detectable either in the presence or absence of TNF stimulation. Very similar patterns of NF-κB subunit binding are seen by ChIP assay of the IL-6 gene (Figure 3c, right panel). In Figure 3d, basal- and TNF-induced binding of the same NF-κB subunits is shown for two representative members of the Late genes. Like the early genes, TNF induces RelA and to a lesser extent, c-Rel, and NF-κB1 binding to both the Naf-1 and NF-κB2 genes (Figure 3d). In contrast, constitutive NF-κB2 binding is seen for both promoters. Because of differences in PCR efficiencies, it is not possible to determine whether NF-κB2 subunits are binding more or less strongly to the Late gene promoters than the Early genes. These findings indicate that the selected Early and Late genes directly and inducibly bind NF-κB/RelA, c-Rel, and NF-κB1 DNA binding subunits in an apparently similar pattern. Taken together, these studies validate the microarray profiles, confirm the relative cascades of gene expression, demonstrate their absolute dependence on intact NF-κB signaling, and indicate the Early and Late members show similar binding affinities for the transactivating NF-κB family subunits.
To determine the functional activities of the various NF-κB dependent groups, the individual probe sets in each expression profile genes were mapped to their primary Gene Ontology Biological Process and Molecular Function. Statistical analysis was performed on functional categories over-represented in the groups relative to the functional representation within the human proteome (Table 1, ref. [23]). For example, Cytokine Activity was significantly enriched in the Early dataset, representing 32 % of the genes, whereas peptide transporter and protein binding activity was enriched in the Late dataset, representing 12- and 31 % of the genes, respectively (Table 1). To more clearly display this functional difference, the 74 NF-κB-dependent genes were annotated by primary biochemical function and kinetic grouping (Table 2). From this analysis, it is clear that the NF-κB-dependent genes control a variety of cellular processes, including anti-apoptosis, cytokine signaling, growth factors and secreted proteins, metabolism, receptors and cell-surface adhesion molecules, signaling molecules, transcription factors and those with currently unknown function. Here, our analysis reveals that groups of genes controlling distinct cellular functions are sequentially expressed during the evolution of the TNF response. For example, the Early gene group is predominantly composed of secreted cytokines, including IL-6, IL-8, CXCL-1 through- 3, TNF and CCL20/Exodus-1 (Table 2). Conversely, the Late gene group encodes cell-surface adhesion molecules (ICAM, KLRC2), signaling adapter molecules (TRAF1/3), and NF-κB2. The Middle group functionally overlaps with those of the Late genes in that they control expression of cell-surface receptors, signaling molecules, autoregulators of the IKK-NF-κB pathway and metabolic enzymes. To avoid the potential problem of bias inherent in expert classification, we employed Ingenuity Pathways Analysis (IPA). IPA compares groups of genes from each expression profile to an annotated database generated from published protein and genetic networks and displays a rank-ordered list of pathways whose function is most likely to be affected by that expression pattern. For each pathway, its members and their relationships (functional and physical) are displayed. Consistently, the highest scoring IPA pathway for Early gene group was an NF-κB-dependent pathway controlling production of extracellular cytokines (Figure 4a). Similarly, although the highest scoring pathway for Late gene group was also an NF-κB-dependent pathway, the major targets of this pathway are extracellular adhesion proteins (Figure 4b). Together, our data suggests that NF-κB controls waves of sequential expression of functionally distinct genes.
Table 1 GO Mapping of NF-κB-dependent genes. Affymetrix probe sets were mapped to Gene Ontology (GO) Biological Process and Molecular Function categories [DAVID, Ref [23]]. For each group, the top 5 ranked processes or functions are tabulated with the number of probe sets and the percentage of the dataset (%) that map to the given process or function, and the statistical significance for enrichment (p value).
Early Middle
Molecular Function Num % p Value Molecular Function Num % p Value
CYTOKINE ACTIVITY 6 31.6 1.44E-06 PROTEIN BINDING 5.000 29.400 0.048
RECEPTOR BINDING 7 36.8 8.1E-06 SUGAR BINDING 2.000 11.800 0.092
CHEMOKINE ACTIVITY 4 21.1 1.37E-05 CARBOHYDRATE BINDING 2.000 11.800 0.096
CHEMOKINE RECEPTOR BINDING 4 21.1 1.37E-05
G-PROTEIN-COUPLED RECEPTOR BINDING 4 21.1 1.52E-05
Late Paradoxical
Molecular Function Num % p Value Molecular Function Num % p Value
PEPTIDE TRANSPORTER ACTIVITY 2 12.5 0.005 CTD PHOSPHATASE ACTIVITY 2.000 12.500 0.044
PRIMARY ACTIVE TRANSPORTER ACTIVITY 3 18.8 0.059 Mg-DEPENDENT Ser/Thre PHOSPHATASE 2.000 12.500 0.044
PROTEIN BINDING 5 31.3 0.063 MYOSIN PHOSPHATASE ACTIVITY 2.000 12.500 0.044
PROTEIN PHOSPHATASE TYPE 2B ACTIVITY 2.000 12.500 0.044
PROTEIN PHOSPHATASE TYPE 2C ACTIVITY 2.000 12.500 0.044
Table 2 Functional classification of NF-κB-dependent genes.
Function Name GenBank Locus Pr(F) Cluster Function Name GenBank Locus Pr(F) Cluster
Anti-apoptosis Receptors TAP1 X57522 6p21.3 1E-08 Late
BID AF042083 22q11.1 1E-07 Middle TAPBP AF029750 6p21.3 1.2E-07 Late
BIRC2 U37547 11q22 0.00017 Middle NK4 AA631972 16p13.3 4E-08 Late
TNFAIP3 M59465 6q23 4.67E-11 Early KCNG1 AL050404 20q13 1.21E-10 Paradox
Cytokine ITGB5 X53002 3q21.2 0.003318 Paradox
IL8 M28130 4q13 4.02E-08 Early GPR49 AF062006 12q22 0.006668 Paradox
IL6 X04430 7p21 8.89E-08 Early CHRNB4 U48861 15q24 0.00637 Paradox
TNF X02910 6p21.3 0.002926 Early F2RL1 U67058 5q13 2.67E-05 Paradox
CXCL1/Gro-a X54489 4q21 6.04E-07 Early AQP3 N74607 9p13 0.008818 Biphasic
CXCL3/Gro-g M36821 4q21 4.62E-08 Early Signaling
CXCL2/Gro-b M36820 4q21 9.88E-15 Early IkBe U91616 6p21.1 1.10E-10 Middle
CCL20/Exodus-1 U64197 2q33 1.96E-07 Early BCL3 U05681 19q13.1 8.44E-05 Middle
Growth Factors/Secreted Proteins TRAF2 U12597 9q34 0.001067 Middle
TNFAIP2/B94 M92357 14q32 1.20E-09 Middle TRAF1 U19261 9q33-q34 5.87E-07 Late
Comp B L15702 6p21.3 1.2E-08 Middle TRAF3 U21092 14q32.33 0.000139 Late
EFNA1 M57730 1q21 8.34E-05 Early IkBa M69043 14q13 5.55E-15 Early
Follistatin M19481 5q11.2 2.26E-05 Paradox PTGS2 U04636 1q25.2 3.43E-08 Early
CTGF X78947 6q23.1 3.6E-06 Paradox PPP1R3C N36638 10q23-q24 0.000734 Paradox
SCGF AF020044 19q13.3 0.002939 Paradox DUSP4 U48807 8p12 8.4E-06 Paradox
Metabolic PTGES AF010316 9q34.3 0.002286 Late
SOD2 X07834 6q25.3 2.98E-07 Middle Transcription factor
GCH1 U19523 14q22.1 3.93E-07 Middle NF-kB1 M58603 4q24 1.73E-08 Middle
GFPT2 AB016789 5q34-q35 8.00E-10 Middle RELB M83221 19q13.32 4.00E-14 Middle
TIMP2 U44385 17q25 0.009703 Paradox NFKB2 X61498 10q24 3.62E-14 Late
HES1 L19314 3q28 0.000133 Paradox REL X75042 2p13 0.00011 Early
CYB5 L39945 18q23 0.005638 Biphasic IRF1 L05072 5q31.1 1.79E-06 Early
PSMB9 AA808961 6p21.3 1E-09 Biphasic TRIM16 AF096870 17p11.2 0.000221 Late
PSMB8 X87344 6p21.3 0.007828 Biphasic Unknown
Receptor/cell surface Unknown HG371-HT26388 - 7.53E-05 Late
KLRC3 AJ001685 12p13 0.000029 Middle TNIP1/Naf-1 AJ011896 5q32 1.00E-11 Late
SDC4 D79206 20q12 6.60E-08 Middle PLAU X02419 10q24 0.000399 Early
SLC7A2 D29990 8p22 0.004933 Middle OLFML2A AL050002 9q34.11 0.000027 Paradox
CD83 Z11697 6p23 1.50E-11 Middle 31 HBE1 AI349593 11p15.5 2E-07 Paradox
IFNGR2 U05875 21q22.11 4.18E-06 Middle chimeric Y15915 - 0.005238 Paradox
ECE1 Z35307 1p36.1 0.004074 Middle DLX2 U51003 2q32 0.008043 Paradox
KLRC2 AJ001684 12p13 0.000822 Late Transgelin D17409 11q23.2 0.00012 Paradox
ICAM1 M24283 19p13.3 0.000124 Late IFI35 U72882 17q21 0.004532 Biphasic
IL27RA AI263885 19p13.11 6E-09 Late MVP X79882 16p13.1 0.004212 Biphasic
The 74 NF-κB dependent probe sets were analyzed. Duplicate probe sets (e.g., those mapping to the same gene) were eliminated and unique genes tabulated. For each gene identified, the primary cellular function (Function), the common name (Name), the Genbank Accession number (GenBank), the chromosomal locus (Locus), the p-value indicating its significance that its expression is affected by NF-κB [Pr(F)], and its Cluster location (Cluster). Clusters are colored according to indicated expression pattern. Abbreviations used are: BID, BH3 domain interacting agonist; BIRC3, IAP homolog 3; TNFAIP3, TNF alpha induced protein3 (A20); IL, interleukin; CXCL, CXC motif ligand; CCL, CC motif ligand; Comp B, complement factor B; EFNA1, ephrin-A1: CTGF, connective tissue growth factor; SCGF, stem cell growth factor; SOD, superoxide dismutase; GCH1; GTP cyclohydrolase; GFPT2, glutamine-fructose-6-phosphate transaminase 2; PSMB, proteasome subunit; CYB5, cytochrome B5, HES, hairy enhancer of split; TIMP, tissue inhibitor of metalloprotein;, KLRC, SDC4, syndecan 4; SCLC7A2, human cationic transporter; IFNGR2, IFN gamma receptor 2; ECE, endothelial converting enzyme; KLRC2, natural lectin killer receptor 2; ICAM, intercellular adhesion molecule; IL27RA, interleukin 27 receptor alpha; TAP, transporter of antigen peptide; TAPBP, TAP binding protein; NK4, natural killer receptor 4; AQP3, aquaporin 3; KCNG1, potassium voltage-gated channel; ITGB5, integrin beta 5; GPR49, orphan G coupled receptor 49; CHRNB4 beta 4 nicotinic acetylcholine receptor; F2RL1, proteinase activated receptor -2; BCL-3, B cell lymphoma 3; TRAF, TNF receptor associated factor; PTGES, prostaglandin endoperoxide synthase; PTGS2, prostaglandin synthase 2; PPP1R3C, regulatory subunit of protein phosphatase 1; DUSP, dual specificity phosphatase; IRF, interferon response factor; TRIM 16, tripartite motif containing-16/estrogen responsive B-Box protein; TNIP/Naf-1, TNF inducible protein/Nef-associated factor-1; MVP, major vault protein; PLAU, urokinase plasminogen-activator gene; OLFML2A, olfactomedin-like 2A; HBE1, hemoglobin epsilon chain; DLX2, distal-less homeobox2.
Figure 4 Ingenuity Pathway Analysis of biological pathways controlled by Early and Late genes. (a) Early gene pathway. Shown is a graphical representation of the highest scoring pathway controlled by the genes in Cluster III. Shown are labeled nodes representing individual protein functions and their relationship represented by edges. Nodes are colored by changes in expression, with red indicating > 10 fold change; pink > 2-fold and < = 10-fold change; no color indicating < = 2-fold change or data is not present. Squares indicate cytokines, circles indicate chemokines, and ovals indicate transcription factor. For the edges, an arrow indicates "acts on". Horizontal lines indicate the most likely subcellular location for the protein encoded by each node. See Legend to Table II for the index of relevant abbreviations. (b) Late gene pathway. Graphical representation of the highest scoring pathway controlled by the genes in Cluster III. See Fig. 2A for explanation of figure and symbols.
To determine whether the NF-κB-dependent gene expression cascades produced by TNFα are observed with other NF-κB activating stimuli, we stimulated HeLatTA/FLAG-IκBα Mut with IL-1α. IL-1α shares the ability to rapidly activate the IKK- NF-κB pathway with indistinguishable kinetics [20]. The mRNA expression profiles displayed as a heat map shows four expression profiles (Figure 5). IL-8, TNFAIP3/A20 and IL-6 (Early genes in response to TNF) were also rapidly induced by IL-1, peaking 1 h after stimulation. The genes encoding Naf-1, PTGES and PSMB9 (Late genes in TNF response) peaked 9 h after IL-1 stimulation. NF-κB1, and NF-κB2 constituted a Middle expression group. Together we conclude that similar temporal expression programs and NF-κB-dependence are seen in response to IL-1 signaling for members of the Early- and Late NF-κB-dependent genes. To partially address whether these expression profiles could be observed in other cell types, a time course experiment of TNFα – stimulated MRC-5 fibroblasts was analyzed for changes in a representative member of the Early gene group (IL-8) and a member of the Late gene group (Naf-1) by Northern blot analysis. As seen in Figure 5b, IL-8 is induced with an apparent plateau 2 h after TNF stimulation. Conversely, Naf-1 expression is not detectably induced at 2 h, but rather begins to increase after 3 h of stimulation, apparently reaching a plateau 6 h and later after stimulation. These findings suggest that these waves of genomic NF-κB responses can be observed in other cell types.
Figure 5 (a) IL-1 induces sequential cascades of NF-κB dependent gene expression. HeLatTA/FLAG-IκBα Mut cells were plated in parallel in the absence or presence of Dox (2 μg/ml) and stimulated with IL-1α. Changes in mRNA abundance (normalized by 18S) was then determined by Q-RT- PCR from total RNA. Shown is a Z-score representation, where red corresponds to Z > +2.5, green indicates Z < 0, and black indicates Z > 0.5. The common name of each gene is indicated at right. (b) TNF sequential cascades of NF-κB dependent gene expression in MRC-5 fibroblasts. Human MRC-5 fibroblasts were stimulated for the times indicated at top with TNFα (20 ng/ml) and RNA extracted. Shown is a northern blot hybridization of 20 μg RNA using probes specific to IL-8 (top) and Naf-1 (bottom). Asterix indicates apparent plateau of gene expression.
The promoters of the Early and Late response groups were subjected to bioinformatics analysis, to determine whether the kinetics of NF-κB-inducible transcription was a function of the location or number of high-affinity NF-κB-binding sites [24]. Within 1000 bp of the transcription start site, between 1 to 6 high-affinity NF-κB-binding sites were found in both expression groups [see Additional files 2, 3]; when subjected to unsupervised hierarchical clustering, neither the location or number of NF-κB-binding sites was apparently predictive of target gene expression pattern (Figure 6a). For example, Naf-1, a Late gene, co-clustered with A20, an Early gene (Figure 6a). Both promoters had >5 putative high-affinity NF-κB DNA-binding sites in their promoters. To initially address the possibility that the combinatorial context in which the NF-κB site was located may determine its pattern of response, we examined whether AP-1 binding sites were enriched in the Early gene promoters. For example, previous promoter analyses from our lab have shown that the presence of AP-1 binding site affects the magnitude of IL-8 gene induction in response to TNFα [25]. The location of putative high-affinity AP-1 binding sites in relationship to the NF-κB binding sites are shown in Figure 6b. Although AP-1 binding sites are frequent in the promoters analyzed, with 31 sites found in the 26 promoters analyzed, the frequency of those containing AP-1 was not different between the expression groups. Here, 10 of 13 Early gene promoters contained at least one AP-1 site, and exactly the same number of Late gene promoters (10/13) contained them.
Figure 6 (a) Hierarchical clustering of high-affinity NF-κB DNA-binding sites. The probability over 100 bp intervals for finding a high-affinity NF-κB-binding site was used for hierarchical clustering (data from Table I) of the early and late NF-κB dependent gene promoters. Data is shown as a heat map, where green = 0, red = 1. The common name of each gene is shown at right. Note that there is no separation of early and late gene promoters based on the pattern or location of the NF-κB-binding sites. (b) Co-occurrence of high-affinity NF-κB- and AP-1 DNA-binding sites. Superimposed on the NF-κB binding site analysis is the presence and location of high-affinity AP-1 DNA-binding sites. The location of each AP-1 DNA-binding site is indicated in black.
However, when the location of NF-κB-binding sites located within phylogenetically conserved domains was considered, striking differences between the two groups emerged. For the Early genes, promoter alignments between human and mouse genes showed that the NF-κB-binding sites were highly conserved, where the A10, CXCL-1, CCL-20, IκBα, IL-6, IRF-1 and TNF genes contained NF-κB-binding sites, representing 7 of the 9 genes amenable for analysis (Figure 7a). Conversely, for the Late genes, only CYB5 and ICAM-1 had NF-κB-binding sites within phylogenetically conserved domains, representing only 2 of the 9 genes (Figure 7b). Together, these data indicate that the Early gene promoters may be under selective conservation pressure to contain NF-κB-binding sites, whereas the Late gene promoters may not be. To further explore the question of co-occurrence of AP-1-binding sites, the frequency of AP-1 sites in phylogenetically conserved domains was also examined. As seen by the green asterixes in Figure 7a, only two phylogenetically conserved domains in Early genes contained high-affinity AP-1-binding sites. More work will be required to understand the biological significance of these apparent differences in binding patterns between the two groups.
Figure 7 Phylogenetic analysis of NF-κB dependent promoters. (a) Early gene promoters. Promoters spanning from -1000 bp to the first nontranslated exon were aligned between human and mouse genes. Shown are the VISA identity curves [49]. For each curve, the percent sequence conservation is plotted over a sliding 20 base pair window (from 0–100% identity). Shaded regions indicated significant regions of sequence conservation. The location of NF-κB-binding sites within these conserved domains are displayed at top (location indicated by I). The presence of AP-1 sites is indicated by green asterix (*). (b) Late gene promoters. For each late gene promoter indicated, analysis as in 7a.
TNF can induce two distinct modes of NF-κB activation patterns- a single, synchronized "monophasic" NF-κB translocation vs a series of damped, desynchronized oscillations ("oscillatory") whose differential effects on cellular genetic response has not been explored [11,14]. Pulse TNF stimulation rapidly activates IKK briefly over 5–15 min, after which the kinase inactivates, thereby allowing newly resynthesized IκBα to recapture activated NF-κB and return it to its inactivated cytoplasmic form. Conversely, tonic TNF stimulation produces a low level of persistent IKK activity. This persistent IKK activity produces continuous IκBα proteolysis and NF-κB binding [14]. To illustrate, the DNA binding profiles of "pulse" TNFα stimulation (15 min, Figure 8a) were then compared with "tonic" TNFα stimulation. Over the first 1 h, NF-κB DNA binding activity in EMSA was indistinguishable between the "tonic" and "pulse" TNF stimulation (Figure 8b). However, after 3 h, NF-κB activated by pulse TNF stimulation is no longer detectable in the nucleus, being relocated into the cytoplasm, whereas tonic TNF stimulation produced a low level of NF-κB binding (Figure 8b; compare 3- and 6 h Tonic vs Pulse stimulated cells). Although high resolution single-cell fluorescence microscopy indicates this is due to a series of dampened oscillations, the oscillations have desynchronous cycle times and presents as an apparently tonic binding pattern in the homogenated cell population (Figure 8b). As expected, cytoplasmic IκBα is rapidly reduced within 30 min in cells subjected to either pulse or tonic stimulation, but only those subjected to tonic treatment show persistent IκB proteolysis (see Western blot in Figure 8c), producing an oscillatory NF-κB translocation profile (compare with Figure 8b). Using these two stimulation modes, we tested their effect on the Early and Late gene expression profiles by Q-RT-PCR. For the Early genes, we found that the expression patterns for IL-8 and TNFAIP3/A20 gene expression were quite similar (Figure 8d). Surprisingly, IL-6 response to pulse stimulation was much greater than that of identically cultured cells that were tonically TNF stimulated (Figure 8d). Cells in the pulse-treated plates are washed in PBS to remove the TNFα ligand after the 15 min exposure time. It may be possible that a secreted TNFα-inducible inhibitor of IL-6 expression (such as an arachidonic acid metabolite) is removed during this processing, accounting for the enhanced IL-6 expression. Nevertheless, and in marked contrast, Late gene expression patterns were significantly reduced in response to pulse stimulation. Tonic TNF stimulation produced a 12-fold induction of NF-κB2 and 120-fold induction of Naf-1 mRNAs, whereas the pulse stimulation produced less than 2-fold mRNA induction for either gene (Figure 8e). Also, the Early expression of TRAF-1 (15 fold at 1 h) was similar for both treatment conditions; however at later times, TRAF-1 expression returned to unstimulated values. Together these data indicate that expression of the Late genes are dependent on tonic stimulation producing continuous oscillatory NF-κB activity, and suggest that the Late genes are recruited into activated expression modes by time-dependent NF-κB exposure.
Figure 8 Late gene expression requires the NF-κB oscillatory mode. (a) Experimental Strategy. Schematic diagram of the tonic and pulse stimulation paradigm. Parallel plates of cells were stimulated with TNF continuously ("tonic" treatment), without removing the agonist. Pulse stimulated cells were exposed to TNF to activate the NF-κB pathway (activation is maximal within 15 min of stimulation), whereupon the agonist is removed from the medium. At identical times after application of the stimulus, cells are harvested for gel shift (Figure 8b) or Q-RT-PCR (Figures 8c, d). (b) NF-κB-binding in tonic- vs pulse-stimulated cells. Nuclear extracts from tonic- or pulse stimulated HeLa cells were prepared and NF-κB-binding measured. Shown is an autoradiogram of the bound NF-κB complexes by EMSA. The specific NF-κB/Rel A and NF-κB1 complexes previously identified by supershift analyses are indicated at left (see Ref [21] for further details). (c) IκB proteolysis and resynthesis in tonic- vs pulse-stimulated cells. Cytoplasmic extracts from tonic- or pulse stimulated HeLa cells were prepared and abundance of IκB determined by Western blot. IκB is rapidly proteolyzed, with both treatments, however, the steady state levels are reduced 3 and 6 h in tonic treated cells compared to those pulse-treated. (d) Early gene expression profiles. HeLa cells were treated as in Figure 8a, total RNA extracted and mRNA abundance (normalized by 18S) determined by Q-RT- PCR. For each of the indicated mRNA transcripts, values are expressed as fold change relative to unstimulated cells and plotted on a logarithmic scale. (e) Late gene expression profiles. Samples obtained as in Figure 8d. The mRNA transcript measured is indicated for each plot.
Discussion
TNFα is a potent inflammatory and immunomodulatory cytokine expressed by macrophages, monocytes, neutrophils, T-cells and natural killer (NK)-cells following stimulation by bacterial endotoxin. Upon binding to high-affinity cell surface receptors, TNFα activates the expression of secondary cytokine cascades and adhesion molecules that, in turn, play important roles in tissue inflammation by coordinating leukocyte activation, chemotaxis and cell death [1,2,26,27]. The intracellular signaling pathways in response to TNF are well understood. Ligation of TNFRI induces protein recruitment to its cytoplasmic death domains, assembling a submembranous signaling complex composed of TRADD, FADD, TRAF2 and other proteins. These, in turn, activate two divergent intracellular signals, the JNK-AP-1 and the IKK-NF-κB pathways responsible for producing homeostatic genomic responses. Although the IKK-NF-κB pathway is critical for inducing tissue inflammation and preventing TNF-induced programmed cell death, surprisingly little is known about its downstream gene targets and their kinetics of induction. In this study, we have systematically analyzed the kinetics of NF-κB-dependent gene expression. Our findings suggest that NF-κB controls distinct groups of target genes whose pattern of expression appear to be an orchestrated cascade of Early, Middle and Late target gene responses. These kinetically separable waves of NF-κB-dependent gene expression control distinct biochemical processes, with the Early gene group primarily encoding for cytokines that mediate TNF's ability to amplify local cytokine cascades in inflamed tissue. Moreover, we find the orchestration of distinct temporal gene expression cascades is a general feature of cytokine-induced NF-κB activation, being also observed in response to stimulation with IL-1. Undoubtedly cell type-specific influences may affect the precise timing of expression and composition of the kinetic groups that we have identified here for epithelial cells, we nevertheless find similar distinct temporal profiles of representative member of the Early and Late gene groups in unrelated human MRC-5 fibroblasts. Finally, our study identifies differences in gene expression depending on NF-κB activation modes that affect target genes within non-phylogenetically conserved regulatory domains. These findings shed important new insights into the genetic responses to cytokine action.
The Early genes are enriched in cytokines and regulatory components of the IKK-NF-κB pathway as analyzed by Gene Ontology, Ingenuity pathway analysis, and expert classification. An important biological property of TNF is to initiate the cytokine cascade in target cells, where the expression of secondary (downstream) cytokines are produced, each with their own distinct biological properties [1,27]. In this manner, TNF amplifies the inflammatory process. We find that a major part of the Early gene group is the CXC chemokine family. CXC chemokines are the numerically largest of the chemokine families, responsible for inducing migration of neutrophilic leukocytes, stimulating wound healing, initiating angiogenesis and promoting tumorigenesis [28]. In addition, a CC chemokine, CCL-20, is responsible for stimulating monocytes and dendritic cells [29]. Another Early gene, the cytokine IL-6, induces B cell differentiation and is a major mediator of the hepatic acute phase reaction. Therefore, TNF stimulation of epithelial cells rapidly induces secondary cytokine cascades that control leukocyte trafficking, wound healing, angiogenesis, and systemic inflammation. Our phylogenetic analysis shows that the Early gene promoters contain NF-κB-binding sites in evolutionarily conserved regions between human and mouse, perhaps suggesting existence of selection pressure for this rapid TNF response. Moreover, since chemokine activity is produced as a major portion for the most rapidly expressed genes, the primary responses of the TNF-stimulated epithelium appear to be the paracrine propogation of the inflammatory response, with the induction of homeostatic factors a secondary priority for the cell.
The other important members of the Early genes encode intracellular regulatory molecules involved in inhibition of the IKK-NF-κB pathway itself. The NF-κB pathway is tightly controlled by negative feedback loops at multiple steps in its signaling pathway [14,30]. One level of feedback inhibition involves the inactivation of nuclear NF-κB and return to its cytoplasmic localization, a process termed the NF-κB-IκB autoregulatory loop [13,15]. In this loop, activated NF-κB produces enhanced expression of IκBα mRNA. IκBα protein is then replenished to bind and inactivate NF-κB/Rel A, returning it back into to the cytoplasm to restore homeostasis. At a second level, activated NF-κB induces inhibitors of the activated IKK complex. This inhibition is mediated by the TNFAIP3/A20 protein, a ubiquitin ligase that associates with RIP and mediates its proteasomal degradation [31], resulting in inhibition of IKK signal [14,31-34]]. Together, these observations indicate that an additional effect of the Early NF-κB response is to terminate the TNFR-IKK-NF-κB signaling pathway at several levels whose effect is to restore cellular homeostasis.
Conversely, the Late gene group encodes adhesion molecules (ICAM, KLRC2), MHC I antigen processing/presentation (TAP, TAPBP). These molecules play important roles in cytotoxic T cell mediated cytolysis. The finding that tonic TNF stimulation is required for adhesion molecule expression and MHC I antigen presentation suggests that tonically TNF stimulated cells, such as those produced in the context of persistent infection, would be targeted for enhanced immune recognition and clearance. Also in this group, the TRAF signal adapter molecules couple TNF receptors to intracellular responses. TRAF1 is distinct from other TRAF isoforms in that it apparently serves to protect cells from apoptosis and plays a role in the negative feedback regulation of receptor signaling [35]. Similarly TRAF3 has inhibitory functions to those of TRAF 2/6 in TNF induced NF-κB activation [35]. In this regard, TRAF-1 and 3 dependence on tonic TNF stimulation suggests a mechanism how the cell attempts to restore homeostasis in the presence of a strong pro-apoptotic stimulus by additional down- regulation of the TNFR-IKK signaling pathway. Our new findings that Late gene expression is dependent on tonic TNF stimulation is mechanistically significant because it means that TNF may produce distinct phenotypic responses depending on the stimulus duration.
Although the mechanisms underlying the different patterns of Early and Late gene expression control were not the focus of this study, several findings merits further discussion. Our preliminary analyses indicate that expression of both classes of genes is absolutely dependent on NF-κB translocation, because expression of both groups is completely blocked by overexpression of the nondegradable IκBα inhibitor (Figures 2, 3). Bioinformatic analysis shows that both groups contain high-affinity NF-κB-binding sites (Figures 6, 7), many of which have been experimentally verified [21,36]. Moreover, we show here that members from both groups inducibly bind NF-κB within their native chromatin environment (Figures 3c,d). Additionally, we previously showed that TNFα robustly induces expression of three members of the Middle gene group (IκBε, NF-κB1, and RelB), and three members of the Late gene group (TRAF-1, NF-κB2, and Naf-1) in the absence of new protein synthesis [21]. Protein synthesis independence excludes paracrine factors mediating Late gene expression in epithelial cells, unlike those seen in other cell types [37]. Finally, we have previously shown that expression of a constitutively active NF-κB/RelA transactivator is sufficient to activate expression of representative Middle and Late genes, excluding a requirement for other TNF-induced signaling pathways in expression of these genes. Together, these data strongly argue that TNFα – induced NF-κB binding to high-affinity DNA-binding sites in the Late gene promoters is necessary and sufficient for their expression. NF-κB, therefore, is a direct regulator of Late gene expression.
Recent studies suggesting that NF-κB binding occurs in two distinct "waves" in LPS-stimulated macrophages [38] raises the question whether Late gene expression could be due to different rates of NF-κB recruitment. Unfortunately, our findings do not support this as a mechanism controlling Late gene activation by NF-κB in epithelial cells. For example, we have previously shown that the kinetics of the potent transactivating NF-κB/RelA subunit binding to the Late gene, Naf-1, is rapid and indistinguishable from that for the Early gene, IκBα [21]. It is still possible that other Late genes not yet tested are bound by RelA more slowly, but at least we can conclude that differences in rate of NF-κB binding cannot account for the late pattern of Naf-1 expression. Alternatively, expression differences could be due to different compositions of NF-κB subunit binding to the Early and Late gene promoters. Although the ChIP assays presented in Figures 3c and 3d show that the RelA, c-Rel and NF-κB1 subunits bind similarly to the Early and Late genes, NF-κB2 appeared to be binding more strongly to the late genes. It therefore is possible that exchange of various transactivating subunits for NF-κB2 may occur later in the time series, a possibility that will require further investigation.
Another possible explanation for the different rates of promoter activation may be through the environment in which the NF-κB binding sites are located in the Early and Late gene promoters. The rate of response of some NF-κB dependent genes has been suggested to be modified by adjacent transcription factor regulatory sites. In our previous studies of IL-8 gene expression, we found that the magnitude of its TNF-induced transcriptional response is partly dependent on an intact upstream AP-1 binding site [25]. Although our preliminary bioinformatic analysis does not indicate any differences in the frequency of co-regulatory AP-1 binding sites between the two groups (Figure 6b), other sites or combinations of sites may be important. For example, a rapid transcriptional response of the A20 gene has been suggested to be due to a "pre-assembled" pre-initiation complex that is nucleated by the SP-1 transcription factor [39]. In this way, the A20 promoter is poised for rapid transcriptional induction when NF-κB is activated. It will be interesting to compare whether the Early genes are pre-loaded with TFIID or RNA Polymerase by ChIP, and whether these patterns are different from the Late gene promoters. However, we note from our previous genomic footprinting studies have shown that TNF induces both NF-κB- and TFIID binding simultaneously to the IL-8 promoter in epithelial cells [18]. We therefore think SP-1 mediated promoter pre-loading is not likely to be a universal mechanism explaining Early gene expression.
Another possible explanation for the delay in Late gene expression is that this group undergoes an additional rate-limiting step necessary for promoter activation after NF-κB binding has occurred. This step is apparently dependent on a TNF stimulation protocol that induces oscillatory NF-κB binding behavior. In a previous mathematical treatment of this phenomenon at a single cell level, we demonstrated that the Late gene expression profiles can be simulated using a theoretical construct of two sequential activator binding steps, the first one being NF-κB [40]. This second activator, yet to be experimentally identified, could be chromatin modification, nucleosomal re-positioning, pre-initiation complex formation, or coactivator recruitment (reviewed in [41-43]]). These possibilities will require further experimentation.
Another conclusion from our study is that Early gene expression is being actively terminated. Comparing the microarray and Q-RT-PCR profiles with NF-κB binding profiles show that IL-8 expression is falling to control levels 3 h after TNF stimulation, even though NF-κB binding continues to be detectable at these times. Both EMSA (Figure 8) and ChIP assays show NF-kB binding is strongly at these times (See Figure 7F in [21]). To our interpretation, these findings indicate a repressive activity is being recruited to the Early gene promoters during the evolution of the TNF response, an activity or factor which has yet to be experimentally identified [see also discussion in [40]].
Finally, we have not addressed regulation of the Paradoxical genes. These genes are not affected by TNF stimulation in the presence of intact NF-κB signaling pathway (Figure 2), but are induced by its absence. One possibility is that they represent a group of genes whose expression is tonically inhibited by basal NF-κB nucleo-cytoplasmic shuttling. This could be through a competition for rate-limiting, shared, coactivators. In the absence of NF-κB translocation, these limiting coactivators are now able to bind and activate expression of the Paradoxical genes.
Conclusion
This study is the first systematic dissection of the NF-κB response profiles downstream of TNF. We have found evidence for temporal waves of NF-κB-dependent target expression encoding distinct molecular functions. The expression profiles are stimulus-independent, being induced in the same coordinated cascades in response to IL-1. Finally, we have identified a subnetwork of the NF-κB response program whose expression is dependent on its oscillatory mode of activation. This finding is significant in that it indicates distinct cellular phenotypes can be produced depending on the duration of TNF stimulation.
Methods
Cell culture and treatment
HeLatTA/FLAG-IκBα Mut, Tet-transactivator (tTA)-expressing HeLa cells stably transfected with a Tet Operator controlled non-degradable IκBα (IκBα Ser32Ala/Ser36Ala) plasmid, were cultured as described [22]. For pulse TNF stimulation, cells stimulated with 25 ng/ml recombinant TNFα for 15 min, and rapidly washed 3 times with PBS before returning to culture medium. For tonic stimulation, 25 ng/ml TNF was added to the culture medium and left for indicated times prior to harvest.
RNA analysis
Twenty micrograms acid guanidium-phenol extracted RNA was analyzed by Northern blot as previously described [21]. The washed membrane was exposed to a Molecular Dynamics PhosphorImager cassette for quantitation. Quantitative real-time reverse transcriptase-polymerase chain reaction (Q- RT-PCR) assays used commercially available primer and probe sequences (ABI, P/N 4331182). For TNFAIP3/A20, the probe sequence was 5'-CAATTGCCG TCACCGTTC-3'; the forward primer was 5'-AGCTTGTGGC GCTGAAAAC-3', and reverse primer was 5'-ACTGAGAAGTG GCATGCATGAG-3'. The cycling parameters for one-step RT-PCR were: reverse transcription 48°C for 30 min, AmpliTaq activation 95°C for 10 min, denaturation 95°C for 15 s, and annealing/extension 60°C for 1 min (repeat 40 times) on an ABI7000 thermocycler. Duplicate CT values were analysed using comparative CT(ΔΔ CT) method. The amount of target (2 -ΔΔCT) was obtained by normalizing to an endogenous reference (18S RNA) and relative to a calibrator (one experimental sample).
Oligonucleotide array data analysis
Four independent Hu95Av2 GeneChip (Affymetrix Inc, Santa Clara, CA) hybrdiziations were performed using RNA isolated from control (0 h), 1, 3 and 6 h TNF stimulated HeLatTA/FLAG-IκBα Mut cells in the presence or absence of Doxycyline (2 μg/ml). For comparison of the fluorescent intensity (Signal Intensity) values among multiple experiments, the Signal Intensity (SI) values for each "experimental" GeneChip were scaled to that of the "base" GeneChip and subjected to a 2 way analysis of variance with replications (ANOVA, Splus 6, Insightful Inc.). As seen in Figure 1 (supplementary information), 343 probe sets were changed by Dox treatment a p value [Pr(F)] of < 0.01. The probe sets were then filtered to identify any that showed a 3-fold difference in SI at any time during the TNF treatment (SI with NF-κB vs SI without NF-κB), identifying 74 probe sets being under NF-κB control. Agglomerative hierarchical clustering was performed using the Weighted Pair-Group Method with Arithmetic mean (WPGMA, Spotfire Array Explorer, v. 8, Spotfire Inc., Cambridge MA) using Euclidian Distance. The primary data has been deposited with GEO or can be found at our website [44].
Functional annotation mapping was performed using the NIAID DAVID database [23,45]). Pathways Analysis was performed using individual clusters as input into the Ingenuity Knowledge Base database [46]. NF-κB-dependent human promoters were obtained from the Human Genome Browser gateway using the Human May 2004 (hg17) assembly (UCSC Genome Bioinformatics Site, [47]. NF-κB-binding sequences were identified by TRANSFAC 4.2 filtering matrix scores by minimizing the sum the false positive and negative error rates. Human and mouse promoters were aligned using the VISTA genome browser 2.0 [48,49].
Protein extraction and analysis
Nuclear and cytoplasmic proteins were fractionated as previously described [21]. 15 μg of nuclear extracts (NE) were subjected to Electrophoretic Mobility Shift Assay (EMSA) using the high-affinity NF-κB-binding site [22]. The complexes were fractionated on 6 % native polyacrylamide gels, dried, and exposed to Kodak X-AR film at 70°C. For Western blot, equal amounts of cytoplasmic protein were fractionated by SDS-PAGE and transferred to PVDF membrane. The membranes were incubated with affinity purified rabbit polyclonal antibodies to IκBα (Santa Cruz Biotechnology). Washed membranes were then incubated with IRDye 800 labeled anti-rabbit IgG antibodies (Rockland Immunochemicals, Gilbertsville, PA), and immune complexes quantified using the Odyssey Infrared Imaging system (LICOR Biosciences, Lincoln, NE.).
Chromatin immunoprecipitation (ChIP) assay
The ChIP assay was as described [21]. On the day prior to experiment, 2–4 × 106 cells were plated in 0.5 % BSA containing growth medium. Cells were stimulated for indicated times, and sequentially crosslinked with disuccinimidyl glutarate and 1 % formaldehyde in serum-free medium for 15 min at 37°C. The cells were washed, transferred to Eppendorf tubes, and solubilized in 400 ml of SDS Lysis Buffer (1% SDS 10 mM Tris, ph 8.0, 1 mM EDTA) with protease inhibitor cocktail (Sigma Aldrich). The samples were sonicated 3 times, 15 sec at setting 2 until DNA fragments were 300–400 bp or less. Equal amounts of DNA were immunoprecipitated overnight at 4°C in ChIP dilution buffer (50 mM NaCl, 1 mM HEPES, pH 7.4, 1% IGEPAL-630, 10 % glycerol, 1 mM DTT) with 20 μg of indicated NF-κB subunit specific antibody (Santa Cruz Biotech) or IgG as indicated. Immunoprecipitates were collected with protein-A magnetic beads (Dynal, Inc), and washed sequentially with ChIP dilution buffer, high salt buffer, LiCl buffer and TE buffer (10 mM Tris, ph 8.0, 1 mM EDTA). DNA was eluted in 1 ml of Elution Buffer 1 % SDS in 0.1 M NaHCO3). Samples were de-crosslinked in 200 mM NaCl at 65°C, 1 h. DNA was phenol extracted, ethanol precipitated and used for PCR. PCR primers and conditions for semiquantiative PCR are in [21]. PCR products were fractionated by agarose gel chromatography and stained with ethidium bromide.
Additional data files
Additional data are available with the online version of this manuscript. File 1 is the data showing the Z-Test analysis of TNF-regulated genes. Files 2 and 3 are the NF-κB-binding site predictions and human-mouse promoter mapping for the Early genes and Late genes, respectively.
Supplementary Material
Additional File 1
Contains the table showing the Z-Test analysis of TNF-regulated genes.
Click here for file
Additional File 2
Contains the NF-κB-binding site predictions and human-mouse promoter mapping for the Early genes.
Click here for file
Additional File 3
Contains the NF-κB-binding site predictions and human-mouse promoter mapping for the Late genes.
Click here for file
Acknowledgements
This project was supported by NIAID grants R01 AI40218, P01 AI062885 (to A.R.B.) and J.W. McLaughlin Predoctoral Fellowship (to D.E.N.). The authors thank the UTMB Genomics Core Laboratory (T. Wood, Director) supported by NIEHS grant P30 ES06676 (to J. Halpert, UTMB) and the Sealy Center for Cancer Biology Real Time PCR Core (T. Ko, Director) for support.
==== Refs
Beutler B TNF, immunity and inflammatory disease: Lessons of the past decade J Invest Med 1995 43 227 235
Brasier AR Li J Mechanisms for inducible control of angiotensinogen gene transcription J Hypertens 1996 27(2) 465 475
Smith CA Farrah T Goodwin RG The TNF receptor superfamily of cellular and viral proteins: Activation, costimulation, and death Cell 1994 76 959 962 8137429 10.1016/0092-8674(94)90372-7
Raingeaud J Gupta S Rogers JS Dickens M Han J Ulevitch RJ Davis RJ Pro-inflammatory cytokine and environmental stress cause p38 mitogen-activated protein kinase activation by dual phosphorylation on tyrosine and threonine J Biol Chem 1995 270 7420 7426 7535770 10.1074/jbc.270.13.7420
Hsu H Shu HB Pan MG Goeddel DV TRADD-TRAF2 and TRADD-FADD interactions define two distinct TNF receptor 1 signal transduction pathways Cell 1996 84 299 308 8565075 10.1016/S0092-8674(00)80984-8
Beg AA Baldwin ASJ The I kappa B proteins: multifunctional regulators of Rel/NF- kappa B transcription factors Genes Dev 1993 7 2064 2070 8224838
Karin M The beginning of the end: IkB kinase (IKK) and NF-kB activation J Biol Chem 1999 274 27342 10.1074/jbc.274.39.27339
Maniatis T Catalysis by a multiprotein IkB kinase complex Science 1997 278 818 819 9381193 10.1126/science.278.5339.818
Han Y Weinman S Boldogh I Walker RK Brasier AR Tumor necrosis factor-alpha-inducible IkappaBalpha proteolysis mediated by cytosolic m-calpain. A mechanism parallel to the ubiquitin-proteasome pathway for nuclear factor-kappab activation J Biol Chem 1999 274 787 794 9873017 10.1074/jbc.274.2.787
Yie J Senger K Thanos D Mechanism by which the IFN-beta enhanceosome activates transcription Proc Natl Acad Sci USA 1999 96 13108 13113 10557281 10.1073/pnas.96.23.13108
Nelson DE Ihekwaba AEC Elliott M Johnson JR Gibney CA Foreman BE Nelson G See V Horton CA Spiller DG Edwards SW McDowell HP Unitt JF Sullivan E Grimley R Benson N Broomhead D Kell DB White MRH Oscillations in NF-{kappa}B Signaling Control the Dynamics of Gene Expression Science 2004 306 704 708 15499023 10.1126/science.1099962
Brasier AR Lu M Hai T Lu Y Boldogh I NF-kB inducible BCL-3 expression is an autoregulatory loop controlling nuclear p50/NF-kB1 residence J Biol Chem 2001 276 32080 32093 11387332 10.1074/jbc.M102949200
Sun SC Ganchi PA Ballard DW Greene WC NF-kappa B controls expression of inhibitor I kappa B alpha: evidence for an inducible autoregulatory pathway Science 1993 259 1912 1915 8096091
Lipniacki T Paszek P Brasier AR Luxon B Kimmel M Mathematical model of NF-kB regulatory module. J Theor Biol 2004 228 195 215 15094015 10.1016/j.jtbi.2004.01.001
Han Y Brasier AR Mechanism for biphasic Rel A:NF-kB1 nuclear translocation in tumor necrosis factor a-stimulated hepatocytes J Biol Chem 1997 272 9823 9830
Ladner KJ Caligiuri MA Guttridge DC Tumor Necrosis Factor-regulated Biphasic Activation of NF-kappa B Is Required for Cytokine-induced Loss of Skeletal Muscle Gene Products J Biol Chem 2003 278 2294 2303 12431991 10.1074/jbc.M207129200
Brasier AR Li J Wimbish KA Tumor necrosis factor activates angiotensinogen gene expression by the Rel A transactivator J Hypertens 1996 27 1009 1017
Brasier AR Jamaluddin M Casola A Duan W Shen Q Garofalo R A promoter recruitment mechanism for TNFa-induced IL-8 transcription in type II pulmonary epithelial cells: Dependence on nuclear abundance of Rel A, NF-kB1 and c-Rel transcription factors J Biol Chem 1998 273 3551 3561 9452482 10.1074/jbc.273.6.3551
Krikos A Laherty CD Dixit VM Transcriptional activation of the tumor necrosis factor alpha-inducible zinc finger protein, A20, is mediated by kappa B elements J Biol Chem 1992 267 17971 1381359
Han Y Meng T Murray NR Fields AP Brasier AR IL-1 Induced NF-kB-IkBa autoregulatory feedback loop in hepatocytes: A role for PKCa in post-transcriptional regulation of IkBa resynthesis J Biol Chem 1999 274 939 947 9873035 10.1074/jbc.274.2.939
Tian B Nowak DE Jamaluddin M Wang S Brasier AR Identification of direct genomic targets downstream of the NF-kappa B transcription factor mediating TNF signaling J Biol Chem 2005 280 17435 17448 15722553 10.1074/jbc.M500437200
Tian B Zhang Y Luxon BA Garofalo RP Casola A Sinha M Brasier AR Identification Of NF-kB Dependent Gene Networks In Respiratory Syncytial Virus-Infected Cells J Virol 2002 76 6800 6814 12050393 10.1128/JVI.76.13.6800-6814.2002
Dennis GJ Sherman BT Hosack DA Yang J Gao W Lane HC Lempicki RA DAVID: Database for Annotation, Visualization, and Integrated Discovery Genome Biol 2003 4 R60 10.1186/gb-2003-4-9-r60
Wingender E Chen X Hehl Karas H Liebich I Matys V Meinhardt T Pruss M Reuter I Schacherer F TRANSFAC: an integrated system for gene expression regulation Nucleic Acids Res 2000 28 316 319 10592259 10.1093/nar/28.1.316
Vlahopoulos S Boldogh I Brasier AR NF-kB dependent induction of interleukin-8 gene expression by tumor necrosis factor a: evidence for an antioxidant sensitive activating pathway distinct from nuclear translocation Blood 1999 94 1878 1889 10477716
Gabay C Kushner I Acute-phase proteins and other systemic responses to inflammation N Engl J Med 1999 340 448 454 9971870 10.1056/NEJM199902113400607
Standiford TJ Kunkel SL Basha MA Chensue SW Lynch JPIII Toews GB Westwick J Strieter RM Interleukin-8 gene expression by a pulmonary epithelial cell line: A model for cytokine networks in the lung J Clin Invest 1990 86 1945 1953 2254454
Baggiolini M Dewald B Moser B Interleukin-8 and related chemotactic cytokines -CXC and CC chemokines Adv Immunol 1994 55 97 179 8304236
Hromas R Gray PW Chantry D Godinska R Krathwohl M Fife K bell GI Takeda J Aronica S Gordon M Cooper S Broxmeye H Kelmsz MJ Cloning and characterization of exodus a novel b-chemokine Blood 1997 89 3315 3322 9129037
Hoffmann A Levchenko A Scott ML Baltimore D The Ikappa B-NF-kappa B Signaling Module: Temporal Control and Selective Gene Activation Science 2002 298 1241 1245 12424381 10.1126/science.1071914
Wertz I O'Rourke KM Zhou H Eby M Aravind L Seshagiri S Wu P Wiesmann C Baker R Boone D Ma A Koonin EV Dixit VM De-ubiquitination and ubiquitin ligase domains of A20 downregulate NF-kB signalling Nature 2004 430 694 699 15258597 10.1038/nature02794
Zhang SQ Kovalenko A Cantarella G Wallach D Recruitment of the IKK signalosome to the p55 TNF receptor: RIP and A20 bind to NEMO (IKKgamma) upon receptor stimulation Immunity 2000 12 301 311 10755617 10.1016/S1074-7613(00)80183-1
He KL Ting AT A20 inhibits tunor necrosis factor (TNF) alpha-induced apoptosis by disrupting recruitment of TRADD and RIP to the TNF receptor 1 complex in Jurkat T cells Mol Cell Biol 2002 22 6034 6045 12167698 10.1128/MCB.22.17.6034-6045.2002
Song HY Rothe M Goeddel DV The tumor necrosis factor-inducible zinc finger protein A20 interacts with TRAF1/TRAF2 and inhibits NF-kappa B activation Proc Natl Acad Sci USA 1996 93 6721 8692885 10.1073/pnas.93.13.6721
Chung JY Park YC Ye H Wu H All TRAFs are not created equal: common and distinct molecular mechanisms of TRAF-mediated signal transduction J Cell Sci 2002 115 679 688 11865024
Pahl H Activators and target genes of Rel/NF-kB transcription factors Oncogene 1999 18 6853 6866 10602461 10.1038/sj.onc.1203239
Pimentel-Muinos FX Mazana J Fresno M Biphasic control of nuclear factor-kappa B activation by the T cell receptor complex: role of tumor necrosis factor alpha Eur J Immunol 1995 25 179 186 7843230
Saccani S Pantano S Natoli G Two waves of nuclear factor kB recruitment to target promoters J Exp Med 2001 193 1351 1359 11413190 10.1084/jem.193.12.1351
Ainbinder E Revach M Wolstein O Moshonov S Diamant N Dikstein R Mechanism of Rapid Transcriptional Induction of Tumor Necrosis Factor Alpha-Responsive Genes by NF-{kappa}B Mol Cell Biol 2002 22 6354 6362 12192035 10.1128/MCB.22.18.6354-6362.2002
Paszek P Lipniacki T Brasier AR Nowak D Tian B Kimmel M Stochastic Effects of Multiple Regulators on Expression Profiles in Eukaryotes J Theor Biol 2004 233 423 433 15652150 10.1016/j.jtbi.2004.10.023
Wolffe AP Transcription: in tune with the histones Cell 1994 77 13 16 8156588 10.1016/0092-8674(94)90229-1
Smith JL Freebern WJ Collins I De Siervi A Montano I Haggerty CM McNutt MC Butscher WG Dzekunova I Petersen DW Kawasaki E Merchant JL Gardner K Kinetic profiles of p300 occupancy in vivo predict common features of promoter structure and coactivator recruitment Proc Natl Acad Sci USA 2004 101 11554 11559 15286281 10.1073/pnas.0402156101
Beato M Chromatin structure and the regulation of gene expression: remodeling at the MMTV promoter J Mol Med 1996 74 711 724 8974015 10.1007/s001090050076
Lab B 2005
NIAID DAVID
Ingenuity Pathways
UCSC Genome Browser
VISTA Genome Browser
Loots GG Ovcharenko I Pachter L Dubchak I Rubin E rVista for comparative sequence-based discovery of functional transcription factor binding sites Genome Res 2002 12 832 839 11997350 10.1101/gr.225502. Article published online before print in April 2002
|
16191192
|
PMC1262712
|
CC BY
|
2021-01-04 16:32:46
|
no
|
BMC Genomics. 2005 Sep 28; 6:137
|
utf-8
|
BMC Genomics
| 2,005 |
10.1186/1471-2164-6-137
|
oa_comm
|
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-1381619755210.1186/1471-2164-6-138Methodology ArticleIdentification of disease causing loci using an array-based genotyping approach on pooled DNA Craig David W [email protected] Matthew J [email protected] Diane [email protected] Victoria L [email protected] Michael C [email protected] Anne M [email protected] Erik G [email protected] John M [email protected] Dietrich A [email protected] Neurogenomics Division, Translational Genomics Research Institute (TGen) Phoenix, Arizona 85004, USA2 Clinic for Special Children, Strasburg, PA 17579, USA2005 30 9 2005 6 138 138 13 5 2005 30 9 2005 Copyright © 2005 Craig et al; licensee BioMed Central Ltd.2005Craig et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Pooling genomic DNA samples within clinical classes of disease followed by genotyping on whole-genome SNP microarrays, allows for rapid and inexpensive genome-wide association studies. Key to the success of these studies is the accuracy of the allelic frequency calculations, the ability to identify false-positives arising from assay variability and the ability to better resolve association signals through analysis of neighbouring SNPs.
Results
We report the accuracy of allelic frequency measurements on pooled genomic DNA samples by comparing these measurements to the known allelic frequencies as determined by individual genotyping. We describe modifications to the calculation of k-correction factors from relative allele signal (RAS) values that remove biases and result in more accurate allelic frequency predictions. Our results show that the least accurate SNPs, those most likely to give false-positives in an association study, are identifiable by comparing their frequencies to both those from a known database of individual genotypes and those of the pooled replicates. In a disease with a previously identified genetic mutation, we demonstrate that one can identify the disease locus through the comparison of the predicted allelic frequencies in case and control pools. Furthermore, we demonstrate improved resolution of association signals using the mean of individual test-statistics for consecutive SNPs windowed across the genome. A database of k-correction factors for predicting allelic frequencies for each SNP, derived from several thousand individually genotyped samples, is provided. Lastly, a Perl script for calculating RAS values for the Affymetrix platform is provided.
Conclusion
Our results illustrate that pooling of DNA samples is an effective initial strategy to identify a genetic locus. However, it is important to eliminate inaccurate SNPs prior to analysis by comparing them to a database of individually genotyped samples as well as by comparing them to replicates of the pool. Lastly, detection of association signals can be improved by incorporating data from neighbouring SNPs.
==== Body
Introduction
The ability to genotype hundreds of thousands of single nucleotide polymorphisms (SNPs) across the genome and to perform association analysis between cases and controls provides, for the first time, a discovery-based approach for determining the underpinnings of complex human genetic disorders. Technologies from Affymetrix (microarray-based GeneChip® Mapping arrays), Illumina (BeadArray™), and Sequenom (MassARRAY™) are now available with sufficient density to detect linkage disequilibrium between informative SNPs and nearby disease-causing nucleotide variants through non-hypothesis based whole-genome association scans for certain populations [1-3].
Several practical issues make whole-genome association studies by utilizing individual genotyping difficult to implement [4]. Power estimates predict that somewhere on the order of a thousand cases and control subjects must be genotyped to detect allelic differences of <5% between the cohorts, as well as to detect rare alleles which may be causative in only a subset of the cohort [5]. Additionally, population stratification and allelic imbalance may identify SNPs that have statistically significant allelic frequency differences yet have no relation to the disease [4,6,7]. Whole-genome association studies are now technologically possible, though the cost is several million dollars if samples are individually genotyped. Here we describe the validation of pooling genomic DNA samples as a rapid pre-screening to detect disease-causing loci for a few thousand dollars on SNP genotyping microarrays.
It is possible to identify SNPs that have significant differences in allelic frequencies between two populations while saving a significant amount in resources by pooling genomic DNA and then SNP genotyping on a single microarray, or preferably on a series of replicated arrays. Indeed, a limited number of studies have been conducted that demonstrate the possibility to predict accurately the allelic frequencies of a SNP from a pooled sample on a microarray, and, in fact identify quantitative trait loci [8-12]. Typically, these studies have validated the pooled allelic frequencies by later individually genotyping between ten to a few hundred SNPs. The most elegant validations of pooling have used indirect approaches, such as spiking a single individual of known genotype into a pooled group with unkown genotypes [9].
One cannot realistically expect all probes on a microarray to function equally, especially considering that the objective of these platforms is to identify allelic differences of 0%, 50%, and 100%. Indeed, as platforms move to 100,000+ SNPs, the ability to select preferentially the best performing SNPs, such as was done in the design of the Affymetrix 10K GeneChip®, will likely be compromised. As a result, our prediction is that many SNPs will be unreliable for pooling, and thus may be more likely to lead to false positives. In a pooling study, limiting false positives that are a result of the assay, rather than the underlying population, will be a major factor in being able to realistically identify SNPs that can predict disease status. In this study, we investigated the reliability of SNP allelic frequency measurements as determined from pooling genomic DNA samples on SNP mapping arrays. We further demonstrate our ability to identify poorly predictive SNPs prior to analysis.
Results
We compared the predicted allelic frequencies from pools of genomic DNA to the known allelic frequencies determined by individual genotyping in order to establish the accuracy of pooling. The goal was to compare allelic frequencies for all the SNPs on a microarray, since not all SNPs will be equally accurate for the prediction of frequencies. Inaccurate SNPs are expected to be problematic as microarrays progress to probe hundreds of thousands of SNPs, whereby SNPs are chosen primarily for their physical position in the genome and not for their reproducibility. Indeed, in order to identify 11,500 SNPs for the Affymetrix 10K GeneChip® Mapping Array nearly 500,000 SNPs were screened by Affymetrix for reliability in the assay [13].
Individual genotyping of SNPs for 107 samples
Allelic frequencies for 10,205 SNPs on 107 samples were determined by individually genotyping on the 10K GeneChip®. These samples were genotyped over a one-year period; therefore, some samples were genotyped on version 1.0 of this platform and others on version 2.0. Only SNPs genotyped on both platforms were utilized for this study. Accuracy of SNP calls was approximately 99.8%, as determined by inheritance errors in family pedigrees, in line with the accuracy reported by Affymetrix (99.9%) [13]. We found no significant decrease in accuracy between the two versions of the 10K GeneChip®. The average percentage of SNPs called across all 107 samples was 90.9%. Only individuals with a call rate above 80% were included in the present study. For example, 2,783 SNPs were called for all individuals and 3,525 SNPs were called for >98% of individuals. In our experience using this platform on over 4,000 samples we determined that the call rate is highly dependent on DNA quality and that high quality genomic DNA yields a call rate of 95–98%. The samples used in this study have been collected over several years with variable DNA quality. It is to be expected expect that large-scale whole-genome association studies will also be forced to utilize DNA of less than optimal quality since hundreds to thousands of individuals are needed. Thus the genomic DNA used in this study will likely be representative of what could be expected in a whole-genome association study on a disease where several-thousand individuals are needed.
Construction of pools
Pools were created in triplicate from the individually genotyped samples. The individuals were from the Old Order Amish and Old Order Mennonite populations of southeastern Pennsylvania [14]. Pool 1 consisted of 52 individuals, Pool 2 consisted of a different 52 individuals, and Pool 3 consisted of 3 patients who died of a form of sudden infant death syndrome known as SIDDT and had a known region of identity-by-descent (shared a pre-defined allele on all six chromosomes across the case cohort) [14]. This region was defined on the 10K microarray by 13 consecutive autozygous SNPs, 6 of which were informative. All DNA was quantitated using PicoGreen reagent (Molecular Probes, Eugene, Oregon) to ensure equal amounts were contributed to the pool from each individual. These three pools were then genotyped in replicates of three on the 10K GeneChip®. In all, 9 microarrays were used for the pooled genotyping compared to 107 microarrays for the individual genotyping.
Calculation of allelic frequencies from pooled samples
The predicted allelic frequencies from pooled genotype samples were calculated for each SNP using a k-correction factor based on their derivation from over 3,000 individuals genotyped on the 10K GeneChip®. The training set consisted of 3,000 individuals that were genotyped in our lab within the past year. All had call rates above 80% with an average call rate of 95%. None of the 3,000 individuals used for calculation of k-correction factors were included in the pooled genomic DNA.
The purpose of the k-correction factor is to allow for calculation of a predicted allelic frequency from peak heights, or in this case fluorescence signal, whereby p = A/(A+kB) [15]. K-correction factors have recently become well established and have been used successfully in primer extension assays whereby measurements in SNP allelic frequencies on pooled genomic DNA have been taken by HPLC, mass spec, and by fluorescence in TAQMAN assays [15-19]. For the 10K Mapping array assay, p is the predicted allelic frequency of the A allele, A is the fluorescent signal intensity measure of the A allele, and B is the fluorescent signal intensity measure of the B allele. The k-correction factor can be calculated for a given SNP using a heterozygote who is AB, effectively 50% A and 50% B. Conveniently, output of the Affymetrix GeneChip software for the Affymetrix 10K Mapping Array includes Relative Allele Signal (RAS) values which have been previously used to determine k-correction values (see Figure 1) [11]. Generally, RAS = A/(A+B). Here, A refers to the median match/mismatched differences of the major allele and B for the minor allele (Affymetrix Technical Manual). There are two RAS values, RAS1 (sense) and RAS2 (antisense) since both sense and antisense directions are probed.
Figure 1 Example of RAS statistics for three SNPs based on genotyping of 100 individuals with an average call rate of all SNPs greater than 98%. These example SNPs illustrate how SNP call reliability can vary both between SNPs and within the same SNP, as measured by RAS1 and RAS2 values. Blue spheres are BB individuals, orange triangles are AA individuals, and green squares are AB individuals, grey stars are "Not Called".
Whereas k-correction factors based on the Affymetrix 10K GeneChip® Mapping Array have been previously calculated directly using only heterozygous RAS values [11], we suggest that this can be improved upon since the RAS values are generally not 0 or 1 for homozygotes (See Figure 1). Indeed, we observed significant deviation for many SNPs, which could potentially add significant bias (see discussion) [11]. Thus for each SNP, we normalized RAS values, referred to as nRAS, using the individuals from the training set that were AA (normalized to 1) and BB (normalized to zero). Without this normalization, predicted frequencies will be systematically biased as the pooled samples approach homozygosity. Thus nRASx = (RASx-AAave)/(BBave) where AAave is the average RASx score for individuals AA in the training set, and BBave is the RASx score for individuals BB in the training set. The value of X refers to whether the calculation is for RAS1 or RAS2, and nRAS values are calculated for both RAS1 and RAS2. Thus, two predictions of allelic frequency are obtained: one from RAS1 and one from RAS2. Each RAS variable has distinct variability, and as shown in Figure 1(b), RAS1 may be very precise with low variance, while RAS2 may exhibit high variance, and vice versa. Averaging the two RAS values will mask the RAS value with lower variance. Because of this independent variability, we do not recommend averaging RAS1 and RAS2 for all SNPs as was suggested in other pooling studies [8,10,12]. Rather, we recommend treating the two RAS values as separate experiments, and preferably removing RAS values with the greatest variance prior to analysis.
Values making up each of the RAS1 and RAS2 mean values are provided for homozygous AA, homozygous BB, and heterozygous individuals on a website based on our 3,000 person database which is being made available to the public as part of this publication. These k-correction factors derived from RAS1 and RAS2 values using this training dataset are available at [20] and as supplementary material.
Comparison of allelic frequencies: Pooling vs. individual genotyping
For the 10,205 SNPs on the Affymetrix 10K GeneChip® we found a median difference in allelic frequency between individually genotyped samples and pooled samples of 5.1%, a mean difference of 6.3%, and a standard deviation of 4.9%. Figure 2 shows a histogram for all the SNPs and their difference between the predicted frequency from the pools and the individual genotypes data. Other studies have reported slightly lower differences between pooled and individually genotyped methods for determining allelic frequencies (3–5%) [8,9,12]. Many reasons are likely for this difference: We used DNA that was collected over ten years and was of varying quality; we also compared all the SNPs on the microarray rather than selecting only a few SNPs for comparison. Realistically, the greater difference seen in this study may be more representative of large association studies, in which thousands of genomic samples of varying quality are pooled.
Figure 2 (A) Allele frequency differences between individual and pooled genotypes. Histogram representing the total number of SNPs at each allele frequency difference between individual and pooled samples. (B) Accuracy of predicted SNP frequencies increases for those SNPs that perform well on Mapping 10K individual assays and decreases for poorly performing SNPs. The mean and median absolute difference between the predicted allelic frequency and individually genotyped allelic frequencies are shown vs. the binned performance of SNPs on individual assays. Performance is ranked by the frequency of calls in a set of 3,000 individually genotyped samples.
Identification of assay false positives
While the differences in frequencies between pooled and individual genotyped samples show that calculating frequencies from pooled samples is highly accurate, it is perhaps of greater importance that we are able to predict those SNPs that are unreliable and largely inaccurate. Assays genotyping 500,000 SNPs will likely not have the ability to be as selective and thus are likely to provide a large number of SNPs that do not reliably quantitate allelic frequencies from pooled genomic samples. As shown in Table 1, we found that the 100 SNPs most likely to give a "NoCall" in individual genotyped samples more often gave unreliable predictions of allelic frequencies in pooled samples. Furthermore, as shown in Figure 2b, those SNPs that are the worst 10% (in terms of % called for individual genotypes) also gave rise to higher allelic frequency differences.
Table 1 Inaccurate SNPs with the largest difference between SNP allele frequencies when genotyped individually vs. calculated from pooled DNA can be partially predicted. Nearly 40% of the SNPs found to be the 100 most inaccurate SNPs were also either (a) one the 500 worst performing SNPs in individual genotyping or (b) had the largest variability between replicates in the pool.
All SNPs (a) 500 worst performing SNPs Criteria: NoCalls on 3000 person database (b) 500 worst performing SNPs Criteria: Pool variability from 3 replicates SNPs found in (a) or (b) Inaccuracy (predicted vs. genotyped)
100 most inaccurate SNPs (individual genotyped vs. pooled) 24.2% 27.3% 38.6% 27.2%
500 most inaccurate SNPs (individual genotyped vs. pooled) 12.5% 14.5% 22.1% 20.2%
Remaining 9605 SNPs 4.4% 5.5% 8.6% 5.0%
We found that rarely called SNPs are also likely to be called inaccurately (Table 1 and Figure 3c). In this case, constructing k-correction factors and predicting allelic frequencies will be unreliable for these SNPs, even if pooled replicates show low variability. SNPs with the highest variance in pool replicates were also unreliable. As a practical measure, we found that applying both filters for too many "NoCalls" in a training set and having a high variance in pooled replicates was more effective than either measure alone. We could identify 1/3rd of the worst performing SNPs (greater than 12% difference), by removing the worst performing 5% of SNPs based on variance in pool replicates and removing the worst performing 5% of SNPs based on excessive "No Calls" when individually genotyped. Consequentially, the removal of those SNPs that are either poorly called in a training set of individually genotyped samples or highly variable across pooled replicates significantly decreases the number of false positives. The number of SNPs removed should maintain a balance between retaining dense SNP coverage and excluding those SNPs more likely to give false positives. Ultimately, removing potential false positives will be a compromise between the coverage of the SNP microarray and the genetic diversity of the population.
Figure 3 Identification of the SIDDT locus from pooled genomic DNA by calculating the mean test-statistic for a rolling window of consecutive SNPs. The moving window was determined across the genome and the p-value was calculated from a distribution of 400 bootstraps of the original dataset. Mean window sizes of 1, 3, 5, 10, 15, and 20 are shown and the SIDDT locus is highlighted in yellow. The SIDDT disease locus is the top region for window sizes of 1, 5, 10, 15, and 20.
It is of interest to note that allelic frequencies calculations were more accurate as SNPs approached homozygosity. For example, for those SNPs with allelic frequencies from 0% to 20% and from 80% to 100% the mean difference was 5.2% vs. a mean difference of 7.1% for SNPs with allelic frequencies between 20% and 80%. This finding may be due to inaccuracies in the assays as SNPs approach 50%, since the variance for heterozygotes is higher than the variance measured for homozygotes.
Identification of a disease locus from genotyping of pooled samples
In order to assess whether it is feasible to use pooled genotyping to identify the genetic locus for a disease, we created case and control pools for the disease sudden infant death with dysgenesis of the testes (SIDDT) and a pool of Amish control individuals.
A test for proportions was employed to detect statistical differences between cases and controls. This test-statistic is more often used in pooled studies since frequency data are generally not whole-integers [19]. Shown in equation 1 is the calculation for the test statistic (T) where fcase is the allele frequency for the case group and fcontrol is the allele frequency for the control group.
The distribution follows an approximate χ2 distribution with one degree of freedom. The SNP with the highest significance, rs949748, had a p-value of 0.00016 and was in the SIDDT locus at chromosome 6q21. However, it is expected that the SNP with the lowest p-value will not always be at the correct disease locus. Even strong single SNP association signals will likely be obscured in the noise when 500,000+ SNPs are probed. Thus we employed a moving window whereby the mean test-statistic of several consecutive SNPs was calculated at each SNP position across the genome. The objective of the moving window was to leverage the fact that neighbouring SNPs will likely be in linkage disequilibrium, whereby one SNP is at least partially predictive of the neighbouring SNP. The number of SNPs contained in the moving window was varied between 1 and 25. Shown in table 2 is the rank of the 6q21 region for varying window sizes. It is of interest to assess sensitivity of this windowing approach to SNPs within the region. Thus, we consecutively removed the top three SNPs contributing to the overall association signal. Removing the first two SNPs has little effect on detecting the association signal. The 6q21 region remains the most significant for window sizes of four and greater even when these top two SNPs are removed. In comparison, all three top SNPs has a marginal effect, lowering the rank of the region from highest to within the top ten.
Table 2 Identification of disease locus using a moving window. SNPs were ranked by test statistics and sorted by physical position. The average was calculated for a moving window of consecutive SNPs across the genome. The region 6q22.1 was already known to contain the mutation leading to the SIDDT. The rank of region 6q22.1 for a various window sizes in shown in the second column. In the 3rd, 4th, and 5th columns, the top 1, 2, and 3 SNPs were removed from the 6q22.1 regions to probe sensitivity of window size.
# SNPs Averaged in Moving Window 6q22.1 Rank Region (All SNPs) 6q22.1 Region Rank (Exclude Top 1 SNP in Region) 6q22.1 Region Rank (Exclude Top 2 SNPs in Region) 6q22.1 Region Rank (Exclude Top 3 SNPs in Region)
1 1 22 24 60
2 11 11 19 11
3 6 6 6 14
4 1 1 1 2
5 1 1 1 8
6 2 2 2 3
7 1 1 1 13
8 1 1 1 3
9 1 1 1 9
10 1 1 1 3
To compute the statistical significance of averaged test-statistics, we used a permutation test. With this approach the consecutive order of SNPs was randomized in four hundred separate bootstrapped datasets. P-value statistics were calculated from the distribution of these datasets. Shown in Table 2, the SIDDT locus (6q22.1-q22.31) was generally revealed as the most significant region of association for window sizes between 4 and 20 SNPs.
It will not always be the case that SNPs are in linkage disequilibrium and a windowing-based approach will be effective. The permutation statistics can be used to test this scenario in order to see if the frequency of a given mean window test-statistic is indeed significant. The Old Order Amish and Mennonite populations used in this study arise from a population founded in approximately the sixteenth century with expectedly larger regions of identity by descent. The Amish and Mennonites are not one large isolated population. It is more accurate to say that both these populations derive from the Swiss Anabaptists (circa 1525). These groups are socially and genetically unique even though both came from the same geographical region. Thus undoubtedly some stratification exists between our two cohorts and it is encouraging that the correct region was easily identifiable despite any stratification [21].
Based on previous research in this population, the 10K Mapping Array was anticipated to be of sufficient density whereby many of the SNPs would be in relative linkage disequilibrium throughout this regional population. Indeed, the permutation statistics of moving windows support this notion as the 6q21 region shows a p-value of <1e-6 for window sizes of 10 SNPs and greater, far lower than would be expected with ~10,000 SNP measurements. Other methods have been developed that reduce noise using haplotype data from SNPs in linkage disequilibrium [22]. In the absence of this haplotype data, which may often be the case, it is encouraging that the very straightforward statistical approach described here is effective at identifying the correct locus.
Discussion
Our results show that (1) pooling genomic samples is highly accurate; (2) unreliable SNPs most likely to give false-positives can be largely identified and removed prior to association analysis; and (3) a moving window of averaged test-statistics can be used to detect association signals. Additionally, we have described modifications as to how allelic frequencies are calculated from RAS values of pooled samples that remove systematic biases.
Pioneering work on pooling studies by other research groups has shown that the average Relative Allele Signal (RASave) can be effectively used to derive k-correction factors by k = RASAVE/(1-(RASAVE), and as such, can be used to accurately predict allelic frequencies [8,11,12]. Pooling studies are intended to be screening approaches. RAS values are highly convenient since they are generated by the Affymetrix GDAS software on the 10K platform and fairly intuitive to understand. We suggest significant improvements to this innovative approach that will remove biases; allow for continued use of RAS values; and result in more accurate predictions. These improvements focus on lowering the number of false positives due to added variance or systematic biases, since the utility of pooling-based approaches will be based on how one can detect association signals given a high number of false positives.
First, RAS1 and RAS2 should not be averaged since they are separate probe sets with distinct variances. One may unnecessarily propagate unwanted variance by averaging. For example in Figure 1b, it is clearly visible that RAS2 is highly predictable of the particular SNP allele whereas RAS1 is highly inaccurate. In this case, averaging RAS2 and RAS1 will produce a RASAVE value that is less accurate than RAS2 alone. We suggest instead that these values be treated as separate measures, each with their distinct variance. In the case of RAS values with a large variance, these values should not be used due to the increased chance of a false positive.
Second, we highly recommend that RAS values for each SNP be normalized prior to calculation of allelic frequencies. When these values are not normalized prior to calculating a predicted allelic frequency a significant bias is introduced since the RAS values, as produced by the Affymetrix GDAS software, generally are not 0.0 or 1.0 for homozygous BB and AA respective alleles. Indeed, on a training set of 1000 individuals we found that 34% of SNPs who were called AA had a RAS value less than 0.9 and 35% of SNPs called BB had a RAS value greater than 0.1. This bias can be seen in an example calculation using k-correction factors derived from a typical RAS value directly obtained from the GDAS software. For example, the average RAS1 for a given SNP of an AA individual may be 0.9, the average RAS1 for a heterozygous individual may be 0.5, and the average RAS2 for a BB individual may be 0.1. When one uses the approach outlined by Butcher, et al, the k-correction factor is 1.0, whereby the RAS value of the average heterozygote is divided by one minus this value [10]. In a pooled sample, the same SNP is expected to have a RAS value of 0.9 if it is completely homozygous for AA. However, using the k-correction approach on non-normalized RAS values, one would predict an allelic frequency of 90%, whereas the actual frequency is 100%, a bias of 10%. These biases would be most pronounced as pools approach dominance by one allele type, as would often be the case for a SNP highly associated to a disease.
While RAS values are readily obtainable from the Affymetrix software for the 10K GeneChip® arrays, they are not provided for the 100K or 500K. This is partly due to the fact that RAS values are no longer used to make a SNP call. We have developed a simple Perl script which generates RAS values, still useful in pooling, for the 100K and 500K Affymetrix GeneChip® platform from CHP files. This tool is available on our website [23]. While one may use these RAS values to find obvious differences in cases and controls, for many SNPs allelic frequencies are not linearly dependent on the RAS values; thus, one should calculate allelic frequencies when possible to reduce uneven biases between different SNPs.
Additionally, we are making public on the same site both normalized and non-normalized k-correction factors derived from over 3,000 genotyped individuals for the 10K version 2.0 SNP genotyping platform. Other research groups have created central repositories for k-corrections using non-normalized RAS values and we will work with these teams to contribute these values to this valuable centralized resource [11].
Conclusion
Prior to the investment of large resources into individual genotyping thousands of individuals, one may first consider pooling samples at a low cost to rapidly ascertain gross population stratification concerns and potentially identify the regions of the genome with the strongest association to the trait. The sheer number of SNPs interrogated will lead to a high number of false positives, due to both actual variation in genotype frequencies of the underlying groups and to technical variance. We demonstrate that technical variance can be detected a priori for each SNP using training sets from large numbers of individual microarrays or by replicates of pooled samples. We further show that despite the issues of population stratification, admixture, and subgroups that are difficult to detect when pooling, the cost savings make pooling a first step that we suggest should logically precede the investment of millions of dollars. We describe here a method by which 100K and 500K Affymetrix SNP array data can be parsed into RAS scores and pooled inbalances accurately assessed in an outbred population.
Methods
10K GeneChip® Mapping Array Genotyping
10K SNP genotyping was performed as detailed by Affymetrix on the 10K GeneChip® Mapping 1.0 and 2.0 Arrays [5]. In short, 250 ng of genomic DNA was digested with 10 units of Xba I (New England Biolabs, Beverly, MA) for 2 hours at 37°C. Adaptor Xba (P/N 900410, Affymetrix, Santa Clara, CA) was then ligated onto the digested ends with T4 DNA Ligase for 2 hours at 16°C. After dilution with water, samples were subjected to PCR using primers specific to the adaptor sequence (P/N 900409, Affymetrix) with the following amplification parameters: 95°C for 3 minutes initial denaturation, 95°C 20 seconds, 59°C 15 seconds, 72°C 15 seconds for a total of 35 cycles, followed by 72°C for 7 minutes final extension. PCR products were then purified and fragmented using 0.24 units of DNase I at 37°C for 30 minutes. The fragmented DNA was then end-labeled with biotin using 100 units of terminal deoxynucleotidyl transferase at 37°C for 2 hours. Labeled DNA was then hybridized onto the 10K Mapping Array at 48°C for 16–18 hours at 60 rpm. The hybridized array was washed, stained, and scanned according to the manufacturer's instructions.
The chp_2_ras.pl script processes one or more CHP text files from Affymetrix 10K and 100K SNP chips, calculates RAS1 and RAS2 scores and outputs them in an Excel spreadsheet. Testing shows that for 10K chips, chp_2_ras.pl produces the same scores as those produced by Affymetrix' GDAS software. chp_2_ras.pl is distributed as part of TGen-Array, a collection of Perl scripts and modules that provide parsing and object-oriented interfaces to common microarray files. The script can be downloaded at the TGen bioinformatics website [23].
Authors' contributions
DWC and MJH performed SNP genotyping, participated in the concept of the paper, and drafted the manuscript. DHL, VLZ, MJH, and AML conducted pooling and SNP genotyping. DWC and JMP performed statistical analysis of the SNP data. DAS participated in study design, coordination, and manuscript drafting. All of the authors have read and approved the final manuscript.
Supplementary Material
Additional file 1
Calculated k-correction factors for pooling on Affymetrix 10K GeneChip Mapping Array based on 3,000 person database.
Click here for file
Additional file 2
The chp_2_ras.pl script processes one or more CHP text files from Affymetrix 10K, 100K, and 500K EA SNP chips, calculates RAS1 and RAS2 scores and outputs them in an Excel spreadsheet. Testing shows that for 10K chips, chp_2_ras.pl produces the same scores as those produced by Affymetrix' GDAS software. GDAS does not calculate RAS values for 100K chips. It should be noted that SNPs on 100K chips do not necessarily contain even numbers of sense and antisense probes and in fact only about 40% have 5 sense and 5 antisense probes. The remaining SNPs have a 6–4 or 7–3 probe bias towards either sense or antisense. This is important because part of the RAS calculation involves taking the median of the "successful" probes and median may not be the best approach if only 3 probes exist in one direction and some may have failed and been discarded. chp_2_ras.pl is distributed as part of TGen-Array, a collection of Perl scripts and modules that provide parsing and object-oriented interfaces to common microarray files. The TGen-Array site contains online documentation for all modules and scripts in the distribution including pages that show the source code so the code and algorithms may be inspected.
Click here for file
Acknowledgements
We thank the Old Order Amish families who participated in the research and the Old Order Amish community for their willingness to participate in research studies.
==== Refs
Matsuzaki H Dong S Loi H Di X Liu G Hubbell E Law J Berntsen T Chadha M Hui H Yang G Kennedy GC Webster TA Cawley S Walsh E Jones KW Fodor SP Mei R Genotyping over 100,000 SNPs on a pair of olignucleotide arrays Nature Methods 2004 1 109 111 15782172 10.1038/nmeth718
Fan JB Oliphant A Shen R Kermani BG Garcia F Gunderson KL Hansen M Steemers F Butler SL Deloukas P Galver L Hunt S McBride C Bibikova M Rubano T Chen J Wickham E Doucet D Chang W Campbell D Zhang B Kruglyak S Bentley D Haas J Rigault P Zhou L Stuelpnagel J Chee MS Highly parallel SNP genotyping Cold Spring Harb Symp Quant Biol 2003 68 69 78 15338605 10.1101/sqb.2003.68.69
Marnellos G High-throughput SNP analysis for genetic association studies Curr Opin Drug Discov Devel 2003 6 317 321 12833663
Cardon LR Bell JI Association study designs for complex diseases Nat Rev Genet 2001 2 91 99 11253062 10.1038/35052543
Risch N Teng J The relative power of family-based and case-control designs for linkage disequilibrium studies of complex human diseases I. DNA pooling Genome Res 1998 8 1273 1288 9872982
Hirschhorn JN Daly MJ Genome-wide association studies for common diseases and complex traits Nat Rev Genet 2005 6 95 108 15716906 10.1038/nrg1521
Carlson CS Eberle MA Kruglyak L Nickerson DA Mapping complex disease loci in whole-genome association studies Nature 2004 429 446 452 15164069 10.1038/nature02623
Butcher LM Meaburn E Knight J Sham PC Schalkwyk LC Craig IW Plomin R SNPs, microarrays and pooled DNA: identification of four loci associated with mild mental impairment in a sample of 6000 children Hum Mol Genet 2005 14 1315 1325 15800012 10.1093/hmg/ddi142
Meaburn E Butcher LM Liu L Fernandes C Hansen V Al-Chalabi A Plomin R Craig I Schalkwyk LC Genotyping DNA pools on microarrays: tackling the QTL problem of large samples and large numbers of SNPs BMC Genomics 2005 6 52 15811185 10.1186/1471-2164-6-52
Butcher LM Meaburn E Liu L Fernandes C Hill L Al-Chalabi A Plomin R Schalkwyk L Craig IW Genotyping pooled DNA on microarrays: a systematic genome screen of thousands of SNPs in large samples to detect QTLs for complex traits Behav Genet 2004 34 549 555 15319578 10.1023/B:BEGE.0000038493.26202.d3
Simpson CL Knight J Butcher LM Hansen VK Meaburn E Schalkwyk LC Craig IW Powell JF Sham PC Al-Chalabi A A central resource for accurate allele frequency estimation from pooled DNA genotyped on DNA microarrays Nucleic Acids Res 2005 33 e25 15701753 10.1093/nar/gni028
Butcher LM Meaburn E Dale PS Sham P Schalkwyk LC Craig IW Plomin R Association analysis of mild mental impairment using DNA pooling to screen 432 brain-expressed single-nucleotide polymorphisms Mol Psychiatry 2005 10 384 92 15452586 10.1038/sj.mp.4001589
Matsuzaki H Loi H Dong S Tsai YY Fang J Law J Di X Liu WM Yang G Liu G Huang J Kennedy GC Ryder TB Marcus GA Walsh PS Shriver MD Puck JM Jones KW Mei R Parallel genotyping of over 10,000 SNPs using a one-primer assay on a high-density oligonucleotide array Genome Res 2004 14 414 425 14993208 10.1101/gr.2014904
Puffenberger EG Hu-Lince D Parod JM Craig DW Dobrin SE Conway AR Donarum EA Strauss KA Dunckley T Cardenas JF Melmed KR Wright CA Liang W Stafford P Flynn CR Morton DH Stephan DA Mapping of sudden infant death with dysgenesis of the testes syndrome (SIDDT) by a SNP genome scan and identification of TSPYL loss of function Proc Natl Acad Sci U S A 2004 101 11689 11694 15273283 10.1073/pnas.0401194101
Hoogendoorn B Norton N Kirov G Williams N Hamshere ML Spurlock G Austin J Stephens MK Buckland PR Owen MJ O'Donovan MC Cheap, accurate and rapid allele frequency estimation of single nucleotide polymorphisms by primer extension and DHPLC in DNA pools Hum Genet 2000 107 488 493 11140947 10.1007/s004390000397
Le Hellard S Ballereau SJ Visscher PM Torrance HS Pinson J Morris SW Thomson ML Semple CA Muir WJ Blackwood DH Porteous DJ Evans KL SNP genotyping on pooled DNAs: comparison of genotyping technologies and a semi automated method for data storage and analysis Nucleic Acids Res 2002 30 e74 12140336 10.1093/nar/gnf070
Norton N Williams NM Williams HJ Spurlock G Kirov G Morris DW Hoogendoorn B Owen MJ O'Donovan MC Universal, robust, highly quantitative SNP allele frequency measurement in DNA pools Hum Genet 2002 110 471 478 12073018 10.1007/s00439-002-0706-6
Giordano M Mellai M Hoogendoorn B Momigliano-Richiardi P Determination of SNP allele frequencies in pooled DNAs by primer extension genotyping and denaturing high-performance liquid chromatography J Biochem Biophys Methods 2001 47 101 110 11179766 10.1016/S0165-022X(00)00156-1
Moskvina V Norton N Williams N Holmans P Owen M O'Donovan M Streamlined analysis of pooled genotype data in SNP-based association studies Genet Epidemiol 2005 28 273 282 15700279 10.1002/gepi.20062
Website
Puffenberger EG Genetic heritage of the Old Order Mennonites of southeastern Pennsylvania Am J Med Genet C Semin Med Genet 2003 121 18 31 12888983 10.1002/ajmg.c.20003
Hinds DA Seymour AB Durham LK Banerjee P Ballinger DG Milos PM Cox DR Thompson JF Frazer KA Application of pooled genotyping to scan candidate regions for association with HDL cholesterol levels Hum Genomics 2004 1 421 434 15606997
Website2
|
16197552
|
PMC1262713
|
CC BY
|
2021-01-04 16:32:46
|
no
|
BMC Genomics. 2005 Sep 30; 6:138
|
utf-8
|
BMC Genomics
| 2,005 |
10.1186/1471-2164-6-138
|
oa_comm
|
==== Front
BMC GenomicsBMC Genomics1471-2164BioMed Central London 1471-2164-6-1411620971910.1186/1471-2164-6-141Research ArticleThe stress response against denatured proteins in the deletion of cytosolic chaperones SSA1/2 is different from heat-shock response in Saccharomyces cerevisiae Matsumoto Rena [email protected] Kuniko [email protected] Randeep [email protected] Hitoshi [email protected] Graduate School of Science and Technology, Chiba University, 1-33 Yayoi-cho, Inage, Chiba, Chiba 263-8522, Japan2 International Patent Organism Depositary (IPOD), National Institute of Advanced Industrial Science and Technology (AIST), Central 6, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8566, Japan3 Human Stress Signal Research Center (HSSRC), AIST, Central 6, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8566, Japan2005 7 10 2005 6 141 141 25 7 2005 7 10 2005 Copyright © 2005 Matsumoto et al; licensee BioMed Central Ltd.2005Matsumoto et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
A yeast strain lacking the two genes SSA1 and SSA2, which encode cytosolic molecular chaperones, acquires thermotolerance as well as the mild heat-shocked wild-type yeast strain. We investigated the genomic response at the level of mRNA expression to the deletion of SSA1/2 in comparison with the mild heat-shocked wild-type using cDNA microarray.
Results
Yeast cDNA microarray analysis revealed that genes involved in the stress response, including molecular chaperones, were up-regulated in a similar manner in both the ssa1/2 deletion mutant and the mild heat-shocked wild-type. Genes involved in protein synthesis were up-regulated in the ssa1/2 deletion mutant, but were markedly suppressed in the mild heat-shocked wild-type. The genes involved in ubiquitin-proteasome protein degradation were also up-regulated in the ssa1/2 deletion mutant, whereas the unfolded protein response (UPR) genes were highly expressed in the mild heat-shocked wild-type. RT-PCR confirmed that the genes regulating protein synthesis and cytosolic protein degradation were up-regulated in the ssa1/2 deletion mutant. At the translational level, more ubiquitinated proteins and proteasomes were detected in the ssa1/2 deletion mutant, than in the wild-type, confirming that ubiquitin-proteasome protein degradation was up-regulated by the deletion of SSA1/2.
Conclusion
These results suggest that the mechanism for rescue of denatured proteins in the ssa1/2 deletion mutant is different from that in the mild heat-shocked wild-type: Activated protein synthesis in the ssa1/2 deletion mutant supplies a deficiency of proteins by their degradation, whereas mild heat-shock induces UPR.
==== Body
Background
Exposure to certain kinds of environmental stress factors, such as chemical, heat, osmotic, etc., induces living organisms to express stress proteins, thereby enabling the organism to acquire stress tolerance. This phenomenon is called the "stress response". Especially, the heat-inducible proteins termed "heat-shock proteins (Hsps)" constitute an important part of the stress-responsive proteins [1]. HSP70s (70 kDa HSPs) were discovered in Drosophila melanogaster, and their homologs have been found in various organisms including yeast [2,3]. HSP70s also function as molecular chaperones [2,3]. In the Saccharomyces cerevisiae genome, there are ca. 14 HSP70-like genes. The SSA, SSB and SSE families are cytosolic HSP70 [4-6], whereas the SSC1 is localized to the mitochondria [7,8]. In addition, KAR2 (BiP) is localized to the endoplasmic reticulum [9-12]. The SSA family contains 4 genes, SSA1, SSA2, SSA3 and SSA4 [13]. Not only are the SSA1 and SSA2 genes constitutively expressed, they are also 96% identical at the nucleotide level [2]. Moreover, there is no change in the phenotype of deletion in either of the SSA1 and SSA2 genes compared with the wild-type. In addition, they do not show thermotolerance without pre-heat treatment at 37°C [14]. However, the ssa1/2 double deletion mutant acquires thermotolerance even at 23°C, and shows a slow growth rate [14]. A suppressor, EXA3-1 which is an allele of HSF1 encoding a heat shock factor [15,16] recovers its growth rate. This phenomenon in the ssa1/2 deletion mutant is speculated to result from the overexpression of certain Hsps [17]. HSP104 and SSA4 are found to be highly expressed in the ssa1/2 deletion mutant [4,18].
SSA1 is involved in protein transport and the rescue of denatured proteins [19-22], and possesses ATPase activity [23]. Sti1p activates ATPase activity of Ssa1p [24]. In addition, Hsp70 is a co-chaperone with Hsp104 and Hsp40 in both S. cerevisiae and E.coli [25,26]. The relationship between these chaperones and human misfolding disease has been shown [27,28]. On the other hand, SSA2 is involved in protein transport into the vacuole [29,30]. Thus, SSA1 is multi-functional, and the ssa1/2 double deletion mutant shows drastic changes needed to acquire thermotolerance, which is similar to the mild heat-shocked wild-type. As Ssa1p and Ssa2p are cytosolic molecular chaperones, it is hypothesized that unfolded proteins appear by the double deletion of SSA1/2.
However, genome-wide expression analysis of the ssa1/2 deletion mutant using cDNA microarray has not been carried out. We believe that gene expression profiling of the ssa1/2 deletion mutant is necessary not only to describe the genomic response developed by yeast to the deletions, but also to reveal the mechanism of the response to denatured proteins. To support the cDNA microarray data, we also performed RT-PCR, and immunoblot analysis of several yeast proteins separated by two-dimensional gel electrophoresis (2-DGE). We demonstrate that the deletion of SSA1/2 genes induces up-regulation of the genes involved in both protein degradation and synthesis, whereas mild heat shock induces UPR.
Results
Comparison of the mRNA expression profiles between the ssa1/2 deletion mutant and the mild heat-shocked wild-type
To investigate the mechanism of the response to denatured proteins comprehensively, the mRNA expression profiling of the ssa1/2 deletion mutant was carried out using yeast cDNA microarray, in comparison with the mild heat-shocked wild-type. The number of up-regulated genes in the ssa1/2 deletion mutant was 144, while that in the mild heat-shocked wild-type by exposure for 30 and 60 min at 43°C was 274 and 400, respectively. The functionally categorized up-regulated genes are shown in Figs. 1A and 2A. The most highly up-regulated genes were categorized into "Cell rescue, defense, and virulence (stress-inducible proteins)" in both ssa1/2 and mild heat-shocked wild-type, of which the rates were 8% and 10%, respectively. On the other hand, the number of down-regulated genes in ssa1/2 was 94, while that in the mild heat-shocked wild-type by exposure for 30 and 60 min at 43°C was 610 and 643, respectively. The functionally categorized down-regulated genes are shown in Figs. 1B and 2B.
Figure 1 The overview of expressed genes in the ssa1/2 deletion mutant. S. cerevisiae JN14 (ssa1/2) and JN54 (wild-type) cells were incubated at 30°C to a logarithmic phase (OD660 = 1). The up-regulated genes (over 2- fold expressed) and down-regulated genes (over 2-fold suppressed) in the ssa1/2 deletion mutant were determined by twice induction of three individual experiments. These genes were functionally categorized using Comprehensive Yeast Genome Database (CYGD) in Munich International Center of Protein Sequence (MIPS) [52]. A, up-regulated genes; B, down-regulated genes.
Figure 2 The overview of expressed genes in the mild heat-shocked wild-type. S. cerevisiae JN54 (wild-type) cells were incubated at 30°C to a logarithmic phase (OD660 = 1), and were then treated with mild heat-shock at 43°C for 30 or 60 min. The up-regulated genes (over 2- fold expressed) and down-regulated genes (over 2-fold suppressed) in mild heat-shocked wild-type were determined by twice induction of three individual experiments. These genes were functionally categorized as in Figure 1. A, up-regulated genes. B, down-regulated genes.
In the ssa1/2 deletion mutant, the percentages of up-regulated genes categorized in "Cell rescue defense and virulence", "Transport facilitation" and "Protein fate" was approximately 2–6 times larger than those of the down-regulated genes (Fig. 3A), and the opposite results were found in "Cellular communication/signal transduction mechanism" category (Fig. 3B). In the mild heat-shocked wild-type, there were no categories in which the percentages of up-regulated genes were over 2-times larger than those of the down-regulated genes (Fig. 4A). However, the percentage of down-regulated genes in "Protein synthesis" was particularly larger (ca. 170-times) than that of the up-regulated genes (Fig. 4B). Thus, the number of up-regulated genes in "Protein synthesis" was remarkably smaller than that of the down-regulated genes in the heat-shocked wild type. Conversely, in the ssa1/2 deletion mutant, the number of up-regulated genes in "Protein synthesis" was larger than that of the down-regulated genes. Therefore, we focused on protein synthesis and correlated protein fate as well as "Cell rescue, defense and virulence" in the ssa1/2 deletion mutant.
Figure 3 The comparison between up-regulated genes and down-regulated genes in the ssa1/2 deletion mutant. Functional categories were the same as in Figure 1. The ratios of up/down-regulated genes or down/up-regulated genes were calculated using the percentages of each category in Fig. 1. A, up/down-regulated genes; B, down/up-regulated genes.
Figure 4 The comparison between up-regulated genes and down-regulated genes in the mild heat-shocked wild-type. Functional categories are as given in Figure 1. The ratios of up/down-regulated genes or down/up-regulated genes were calculated using the percentages of each category in Fig. 2. A, up/down-regulated genes; B, down/up-regulated genes.
Figure 5 shows a detailed comparison of these categorized genes up-regulated in the ssa1/2 deletion mutant and in the mild heat-shocked wild-type. Figure 5A shows the comparison of all the genes up-regulated in the ssa1/2 deletion mutant and the mild heat-shocked wild-type. In the categories of "Cell rescue, defense and virulence", several Hsps, including molecular chaperones, were commonly up-regulated in the ssa1/2 deletion mutant and the mild heat-shocked wild-type (Fig. 5B). Although genes related to protein synthesis were greatly suppressed in the mild heat-shocked wild-type, ribosomal protein genes were found to be up-regulated only in the ssa1/2 deletion mutant (Fig. 5C). Table 1 shows the expression level of these genes: RPL37A, RPL25, MRP8, RPS15, MRPL10, RSM18, RPL8B and RSM10.
Figure 5 The comparison of up-regulated genes in the ssa1/2 deletion mutant with those in mild heat-shocked wild-type. Venn Diagrams were constructed by the "GeneSpring" software (Silicon Genetics). Functional subcategories are according to Figure 1.
Table 1 The genes involved in ribosomal proteins up-regulated in the ssa1/2 deletion mutant.
Gene name Expression level* Description
RPL37A 3.2 60S ribosomal protein L37A (YL35)
RPL25 2.8 Ribosomal protein L25 (rpl6L)
MRP8 2.4 Mitochondrial ribosomal protein
RPS15 2.2 40S ribosomal protein S15 (S21) (rp52) (RIG protein)
MRPL10 2.0 Mitochondrial ribosomal protein MRPL10 (YmL10)
RSM18 1.9 Protein of the small subunit of the mitochondrial ribosome
RPL8B 1.8 Ribosomal protein L8B (L4B) (YL5)
RSM10 1.7 Protein of the small subunit of the mitochondrial ribosome
*The expression levels are the average value of three independent experiments.
On the other hand, in the category of "Protein fate", the PRE1, RPN4, RPN12 and SCL1 genes that encode for cytosolic proteasome subunits, were found to be up-regulated in ssa1/2 deletion mutant (Fig. 5D). In addition, the UBC4 (ubiquitin conjugating enzyme) gene was also up-regulated. Table 2 shows the expression level of the genes involved in protein degradation. These results suggest that the ubiquitin-proteasome protein degradation pathway is activated in the ssa1/2 deletion mutant. Although a few ubiquitin-proteasome genes (UBI4, UBC5 and UBP9) were up-regulated in the mild heat-shocked wild-type, they were not in common with those up-regulated in the ssa1/2 deletion mutant (Fig. 5D). The proteasome genes up-regulated in the mild heat-shocked wild-type included vacuolar protein genes (AUT7, LAD4 and APG17), and unfolded protein response (UPR) genes (DER1, PDI1 and ERO1) (Fig. 5D).
Table 2 The genes involved in proteolytic degradation up-regulated in the ssa1/2 deletion mutant.
Gene name Expression level* Description
CDC23 3.7 Cell division cycle protein
PRE1 3.0 22.6 kDa proteasome subunit (20S proteasome subunit C11 (beta4))
UBC4 2.8 Ubiquitin-conjugating enzyme
RPN4 2.6 Ubiquitin-mediated 26S proteasome subunit
MET30 2.5 Met30p contains 5 copies of WD40 motif and interacts with and regulates Met4p
SCL1 2.4 20S proteasome subunit YC7alpha/Y8 (protease yscE subunit 7)
PBI2 2.3 Proteinase inhibitor that inhibits protease Prb1p (yscB)
RPN12 2.0 26S proteasome regulatory subunit
*The expression levels are the average value of three independent experiments.
Confirmatory RT-PCR for proteolytic degradation- and ribosomal biogenesis-related genes
To verify that both protein synthesis and degradation are activated in the ssa1/2 deletion mutant, RT-PCR analysis of several proteolytic degradation genes and cytosolic ribosomal protein genes was carried out. Proteasome subunit genes (PRE1, RPN4, RPN12, and SCL1), an ubiquitin conjugating enzyme gene (UBC4), and cytosolic ribosomal protein genes were found to be up-regulated in the ssa1/2 deletion mutant compared with the wild-type (Fig. 6A and Fig. 6B). This result supports the cDNA microarray data showing that both ubiquitin-proteasome protein degradation and protein synthesis were activated by deletion of the SSA1/2 genes. Only KAR2 was highly expressed among the UPR genes (Fig. 6A).
Figure 6 RT-PCR analysis of ribosomal protein and proteolytic degradation genes in the ssa1/2 deletion mutant and wild-type. RT-PCR was carried out as described in "Materials and Methods", and primers, product size and numbers of PCR cycle are described on Table 3. The RT-PCR products were run on a 4% Nu-Sieve 3:1 Plus agarose gel. A, genes for proteolytic degradation. B, ribosomal protein genes.
Immunoblot analysis of proteolytic degradation-related gene products
To confirm that ubiquitin-proteasome protein degradation is activated at the translational level in the ssa1/2 deletion mutant, immunoblot analysis was performed. Pre1p (20 S proteasome subunit) and Rpn4p (Ubiquitin-mediated 26 S proteasome subunit) increased in the ssa1/2 deletion mutant compared with the wild-type (Fig. 7). An anti-multi ubiquitin mouse monoclonal antibody (FK2) [31,32] detects only ubiquitin that is covalently bound to substrate proteins, i.e. ubiquitinated proteins, and not free ubiquitin [31,32]. Ubiquitinated proteins especially with molecular weights less than 30 kDa increased in the ssa1/2 deletion mutant in comparison with the wild-type (Fig. 8).
Figure 7 Immunoblot analysis of proteasome subunit genes in the ssa1/2 deletion mutant and wild-type. 2-DGE was performed as described in "Material and Methods", followed by immunoblot analysis. A, Pre1p (20 S proteasome subunit). B, Rpn4p (Ubiquitin-mediated 26 S proteasome subunit).
Figure 8 Ubiquitinated proteins in the ssa1/2 deletion mutant and wild-type. 2-DGE was performed followed by immunoblot using an anti-multi ubiquitin mouse monoclonal antibody (FK2).
Discussion
In the present study, we reveal global differences in gene expression between yeast cells lacking two cytosolic HSP70s, SSA1 and SSA2, and the mild-heat-shocked wild-type using cDNA microarray technologoly.
Results from cDNA microarray analysis reveal that the stress-inducible protein genes, including molecular chaperones, were up-regulated in the ssa1/2 deletion mutant in a similar fashion as seen in the mild heat-shocked wild-type (Figs. 1A, 2A, and 5B). It is clear that thermotolerance is due to expression of these stress-inducible proteins. In the ssa1/2 deletion mutant, HSF1 suppressing growth rate of the ssa1/2 [15,16] was expressed normally and its expression level was unchanged (data not shown). Several genes involved in the ubiquitin-proteasome protein degradation pathway were up-regulated in the ssa1/2 deletion mutant (Fig. 5D and Table 2). UBC4 [33,34] was also up-regulated in the ssa1/2 deletion mutant, which is consistent with a previous report [35]. UBC4/5 is necessary for binding between the substrates and Lys48 of ubiquitin that is a target of the 26 S proteasome [34], and for binding between the substrates and Lys63 of ubiquitin, that is not a target of the 26 S proteasome. In addition to UBC4, we found up-regulation of several proteasome genes (PRE1, RPN4, RPN12 and SCL1) in the ssa1/2 deletion mutant. PRE1 and SCL1 encode 20 S proteasome, and RPN4 and RPN12 encode 26 S proteasome [36]. RPN4 (SON1) is a factor involved in ERAD (endoplasmic reticulum associated degradation) [37]. All these genes are essential for degradation of the ubiquitinated proteins [36,38]. RT-PCR data support the up-regulation of these proteasome genes by the deletion of SSA1/2 (Fig. 6A). Moreover, we confirmed that Pre1p and Rpn4p were up-regulated in the ssa1/2 deletion mutant at the translational level by immunoblotting (Fig. 7). This result provided further evidence that proteolytic degradation by proteasomes was stimulated by the deletion of SSA1/2. As shown in Fig. 8, more ubiquitinated proteins, especially with molecular weights less than 30 kDa, were detected in the ssa1/2 deletion mutant than in the wild-type. The deletion of UBP3 in ssa1/2 has been reported to lead to a significant increase in the number of the ubiquitinated proteins, mainly with molecular weights of more than 30 kDa [35]. The expression level of UBP3 did not change in the ssa1/2 deletion mutant compared with the wild-type (data not shown). Therefore, the increase of ubiquitinated proteins in ssa1/2 is not caused by the deletion of UBP3. There are two possibilities for the increase of ubiquitinated proteins in the ssa1/2 deletion mutant. First, insufficiency of the UPR in the ssa1/2 deletion mutant may lead to activation of the ubiquitin-proteasome protein degradation system. Thus, ubiquitination of the target proteins increases and the expression of proteasome genes is induced. Second, some defect of deubiquitination occurs in the ssa1/2 deletion mutant, consequently leading to the accumulation of ubiquitinated proteins in the cell followed by cell death. Furthermore, proteolytic degradation by proteasomes is facilitated. On the other hand, we found that several genes encoding ribosomal proteins were up-regulated in the ssa1/2 deletion mutant (Figs. 5C, 6A, and Table 1), implying that protein synthesis is activated by the deletion of SSA1/2.
In case of the mild heat-shocked wild-type, genes involved in protein synthesis were significantly suppressed (Figs. 2B and 4B), and the proteasome genes up-regulated in the ssa1/2 deletion mutant did not show any change in their expression levels (Fig. 5D). Instead, some UPR genes (PDI1, DER1, ERO1 and KAR2) were up-regulated (Fig. 5D), implying that UPR occurs during mild heat-shock. The mechanism of UPR is known to induce the up-regulation of ER chaperones for refolding when unfolded proteins accumulate in the ER [39]. In the ssa1/2 deletion mutant, the expression of three UPR genes (PDI1, DER1 and ERO1) remained unchanged (data not shown), and only KAR2 was up-regulated (Fig. 3).
Gasch et al. has reported the genome-wide expression analysis of yeast cells exposed to environmental changes [40]. We compared our data on the mild heat shocked wild-type yeast cells with their data on the wild-type cells shifted to 37°C from 25°C. The stress-inducible protein genes up-regulated in the mild-heat shocked wild type (SSA3, SSA4, SSE2, CTT1, HSP26, HSP78, and HSP104) (Fig. 5B) are in common with the results obtained by Gasch et al. [40]. In the category of "Protein fate", more than 70% of the up-regulated genes in our experiments are also in common with their results [40], even though there is a time lag with their experiments. However, DER1, one of the UPR genes, was not found to be up-regulated in their study during the entire the heat-shock treatment period [40]. In contrast, DER1 was up-regulated in our experiments (Fig. 5D). This may be due to the difference in the temperature and time of heat-shock treatment. On the other hand, in the category of "Protein synthesis", ribosomal protein genes are significantly suppressed in their experiments [40], which is consistent with our data. From these comparisons, it can be said that our data on the mild heat-shocked wild-type is similar to that reported by Gasch et al. [40] in the categories of "Cell rescue, defense, and virulence", "Protein fate" and "Protein synthesis".
It is reasonable that UPR is activated and protein synthesis is suppressed in the mild heat-shocked wild-type. We speculate on the reasons as to why the genes involved in both protein degradation and protein synthesis are up-regulated in ssa1/2 deletion mutant. In the normal state, proteins are synthesized on the ribosome, followed by post-translational modifications in the ER or the Golgi apparatus to finally become mature and functional entitles. Schubert et al [41] showed that 30% of the de novo synthesized proteins are degraded before coming to maturity. Therefore, it can be reasoned that post-translational protein denaturation occurs moderately even under normal conditions. However, organisms have developed several mechanisms in their response to the denatured proteins. UPR is one of the ER quality control mechanisms [39]. In addition, the refolding of denatured proteins is carried out by cytosolic chaperones [25,42,43], including SSA1/2 [20,21]. It can be hypothesized that the deletion of SSA1/2 leads to the suppression of refolding, which is then followed by an accumulation of the denatured proteins in cells. The genes involved in proteolytic degradation may be up-regulated to remove such denatured proteins. However, if the ubiquitin-proteasome system keeps on degrading proteins, the depletion of the proteins essential for growth and development will occur. It is suggested that protein synthesis is activated to supply the proteins deleted by proteolytic degradation in the ssa1/2 deletion mutant. In the ssa1/2 deletion mutant, several hexose transporter genes (HXT2, HXT4, HXT6, HXT7), and the genes that belong to early part of glycolysis (GLK1, HXK1) were up-regulated (data not shown). The expression of these genes, involved in energy generation, may be required for sustaining the increased protein synthesis in the ssa1/2 deletion mutant. HXT genes up-regulated in the ssa1/2 deletion mutant are low-glucose dependent [44-46]. It is possible that the uptake of glucose is activated to generate energy, because energy is consumed by protein synthesis that is induced by the deletion of SSA1/2.
These results indicate that different mechanisms of the response to denatured proteins are employed between the ssa1/2 deletion mutant and the mild heat-shocked wild-type even though several up-regulated Hsps (molecular chaperones) are common between the ssa1/2 deletion mutant and the mild heat-shocked wild-type (Fig. 5B). When Hsp104p, Ydj1p (yeast Hsp40p), and Ssa1p exist together, their chaperone activities increase significantly [25]. From this, it is suggested that the deletion of SSA1/2 induces the suppression of their chaperone activities. Recently, the cooperation of Hsp26p wih Hsp104p/Hsp70p/Hsp40p chaperone system on protein disaggregation in yeast was reported [47,48]. Hsp26p co-aggregated with substrate is suggested to be a target of the Hsp104p/Hsp70p/Hsp40p chaperone system [47,48]. Although Ssa1p is able to disaggregate the early Hsp26p-substrate complex (small soluble aggregates), Hsp104p is essential in refolding the late Hsp26p-substrate complex (big insoluble aggregates) [47,48]. Moreover, excess or stoichiometric Hsp26p against denatured substrates is essential for effective refolding [47]. In the ssa1/2 deletion mutant, an increase in the mRNA expression levels of HSP104 and HSP26 was seen (Fig. 5B). Although the refolding of denatured proteins is sure to succeed if HSP104/Hsp104p and HSP26/Hsp26p are highly expressed, it is a fact that the ubiquitin-proteasome degradation system is facilitated in the ssa1/2 deletion mutant. It can be speculated that as constitutive protein denaturation occurs, the ubiquitin-proteasome degradation system is required in addition to the chaperone refolding system in the ssa1/2 deletion mutant. Furthermore, there is a possibility that protein refolding by molecular chaperones and ubiquitin-proteasome protein degradation are related. In mammalian cells, the following model has been reported; denatured proteins are refolded by Hsp70-HSP40 chaperone-mediated maturation pathway under the treatment of Hsp90 inhibitor, and then denatured proteins are degraded by ubiquitin-proteasome [49]. It is interesting to note that the ubiquitin-proteasome protein degradation system in yeast is induced when the chaperone function is inhibited by the deletion of SSA1/2. However, at present, our data are not sufficient to propose a similar model in yeast, and this remains a topic for future study.
Conclusion
The protein synthesis and ubiquitin-proteasome degradation system were up-regulated in the ssa1/2 deletion mutant, whereas UPR genes were up-regulated but protein synthesis was strongly suppressed in the mild heat-shocked wild-type. These results suggest that the mechanism for rescue of denatured proteins in the ssa1/2 deletion mutant differs from that in the mild heat-shocked wild-type, although the phenomena on acquisition of thermotolerance are similar.
Methods
Strains and growth condition
S. cerevisiae JN14 is the ssa1/2 deletion mutant strain (MATa his3-11, 3-15 leu2-3, 2-112 ura3-52 trp1-?1 lys2? Ssa1-3::HIS3 ssa2-2::URA3) [15]. S. cerevisiae JN54 (MATa his3-11, 3-15 leu2-3, 2-112 ura3-52 trp1-?1 lys2?) is the parent strain (wild-type) of the ssa1/2 double mutant [15]. Yeast cells were incubated in 100 ml of YPD (1% yeast extract, 2% polypeptone and 2% glucose) medium at 30°C to a logarithmic phase (OD660 = 1) using 500 ml Erlenmeyer flasks, and collected by centrifugation (2,800 × g). Cells were washed with distilled water three times, and stocked in a -80°C deep freezer until used for total RNA extraction.
Heat-shock treatment
S. cerevisiae JN54 (wild-type) cells were incubated in YPD medium at 30°C to a logarithmic phase (OD660 = 1), followed by treatment with mild heat-shock at 43°C for 30 or 60 min in pre-warmed (43°C) 100 ml of YPD medium using 500 ml Erlenmeyer flasks. Heat-shocked cells were collected by centrifugation. Cells were washed with distilled water three times, and stocked in a -80°C deep freezer until used for total RNA extraction.
RNA extraction and hybridization to a cDNA microarray
Total RNA was extracted by the hot-phenol method [50]. The extraction of mRNA and reverse-transcription to cDNA was done according to Momose and Iwahashi [51]. Poly (A) +RNA was purified from total RNA with an Oligotex-dT30 mRNA purification kit (TaKaRa, Otsu, Shiga, Japan). Fluorescence-labeled cDNA was synthesized with a Cyscribe cDNA Labeling Kit (Amersham Biosciences, Little Chalfont, Buckinghamshire, UK) and 0.5 mM Cy3-UTP (Amersham Biosciences) or 0.5 mM Cy5-UTP. Cy3-UTP was used for the wild-type (as control), and Cy5-UTP was used for the ssa1/2 deletion mutant (as sample). For the heat-shock experiment, Cy3-UTP was used for control cells (30°C), and Cy5-UTP was used for mild heat-shocked cells (43°C). Synthesized cDNA were hybridized to a Kuhara DNA chip (DNA Chip Research, Yokohama, Kanagawa, Japan) at 65°C for 48 h.
cDNA microarray analysis
Hybridized cDNA microarray were washed, dried, and scanned using Scanarray 4000 (GSI Lumonics, Billerica, MA, USA). Quantification of gene expression was done using the Genepix ver. 4.0 quantitative microarray analysis application program (Axon Instruments, Union City, CA, USA). The ratio of intensity Cy5/Cy3 was calculated and normalized with negative control spots. All the calculations and normalizations were done using "Chip Cleanser" program [52]. The functional categorization of genes was performed using GeneSpring (Silicon Genetics, Redwood City, CA, USA), and Comprehensive Yeast Genome Database (CYGD) at the Munich International Center of Protein Sequence (MIPS) database [52]. The over 2- fold expressed genes by the deletion of SSA1/2 or the mild heat-shocked treatment in the wild-type were selected as up-regulated genes and determined by at least twice- induced, out of three individual experiments. Similarly, the over 2- fold suppressed genes by the deletion of SSA1/2 or the mild heat-shocked treatment in the wild-type were selected as down-regulated genes, and determined by twice suppressed, out of three individual experiments [51,53]. The values for up- or down-regulated genes were the average ratio from three independent experiments. The data obtained in this experiment are available with the accession numbers GSE3315 (ssa1/2 deletion mutant) and GSE3316 (mild heat-shocked wild-type) in the Gene Expression Omnibus Database (GEO) [54].
RT-PCR analysis
Total RNA extraction was carried out as described above. RT-PCR was performed using the One Step RNA PCR Kit (AMV) (TaKaRa), according to the instructions provided by the manufacturer. The primers used for RT-PCR are described in Table 3, and 0.1 μg of total RNA were used for RT-PCR. After reverse transcription, samples were subjected to a cycling regime of 20–25 cycles (details are mentioned in Table 3). Five μl of RT-PCR products were loaded into the wells of a 4% Nu-Sieve 3:1 Plus agarose (Cambrex Bio Science Rockland, Inc. Rockland, ME, USA) gel, and electrophoresis was carried out for 50 min at 100 V. The gels were stained using 10 μg/ml ethidium bromide followed by visualization of the stained bands with an UV-transilluminator (ATTO, Tokyo, Japan).
Table 3 List of primers for RT-PCR
Gene name Forward primer (5'-3') Reverse primer (5'-3') Product size (bp) No. of cycles
PRE1 TGACTTCCAGGCACAGTGAA TCTCACTCTGCCAACAAAAA 187 25
RPN4 CGAAGCATGAAGATTTGTCG AAGAACATTCCTGAATGCAGAT 202 25
RPN12 CCAATCAAAGGAGAAAGCTGA CTCCGGGAGAGAAAAAGTTG 178 22
SCL1 AGTCGGTGTCGCTACAAAGG CGACAAAAGGGCTTGAAAAG 229 20
UBC4 CAGCCAGAGAATGGACAAAGA AGGTTCCCCTGTACTGTTGC 220 20
KAR2 GTTCTGGTGCCGCTGATTAT CGAAAATTGTATGAAGCTCGAA 205 20
RPS15 AGAGCCGGTGCTACTACTTCC CGTGTACAACCCCCATTCAC 200 22
RPL25 CGTTACCAAGAAGGCTTACG CGTGCACTCTGCCACTACAC 203 22
RPL37A CAAACCGGCTCTGCTTCTAA TTCCCGTAAGCACTCAAAGG 194 25
ACT1 CCTTCCAACAAATGTGGATCT CAGTGCTTAAACACGTCTTTTCC 200 25
Antibodies
Anti-Pre1p peptide (Res. No., 17-38) and anti-Rpn4p peptide (Res. No., 499-509) rabbit polyclonal antibodies were ordered to Sigma Genosys (Tokyo, Japan). The anti-multi ubiquitin mouse monoclonal antibody (FK2, Cat. No. SPA-205) was purchased from Stressgen Bioreagents Ltd. Partnership (Victoria, B.C., Canada) [31,32].
Two-dimensional gel electrophoresis (2-DGE) and Immunoblot analysis
Yeast cells were washed with distilled water three times. Total protein was extracted from cells homogenized with lysis buffer [(7 M urea (ICN Biomedicals, Aurora, OH, USA), 2 M thiourea (Sigma, St. Louis, MO, USA), 4% CHAPS (Sigma), 1% carrier ampholyte (pH 3.5-10, Amersham Biosciences), 18 mM Tris-HCl, pH 7.5, 14 mM Trizma base (Sigma), EDTA-free Proteinase Inhibitor (Roche Diagnostics, Manheim, Germany), 0.2 % Triton X-100, reduced (Sigma), 14.4 mM DTT (Sigma)]. Resuspended cells were broken with glass beads at 4°C for 10 min, and centrifuged at 20,000 × g for 10 min. Cell lysate was centrifuged again at 20,000 × g for 7 min. Equal amounts (350 μg) of protein were subjected to 2-DGE, following O'Farrell's method [55]. Electrophoresis [IEF, carried out in a glass capillary tube of 13 cm length and 3 mm diameter (Nihon Eido, Tokyo, Japan) and SDS-PAGE (12.5% or 15% polyacrylamide gel, 5% stacking and 12.5% or 15% separation gel; using standard glass gels plates obtained from Nihon Eido) in the second dimension] was carried out at a constant current of 35 mA for 2-1/2 h or until the dye (250 μL BPB; 0.1% (w/v) in 10% (v/v) glycerol in MQ) reached the bottom of the gel [56]. Ten μL of the commercially available "ready-to-use" molecular mass standards (Precision Plus Protein Standards, Dual Color, Bio-Rad, Hercules, CA, USA) were loaded next to the acidic end of the IEF tube gel. Reproducibility of 2-DGE protein profiles was confirmed by running at least 3 independent protein samples extracted from the cells of wild-type and the ssa1/2 deletion mutant. Electrotransfer of proteins on gel to a PVDF (NT-31, Nihon Eido) membrane was carried out at 1 mA/cm2 with a semi-dry blotter (Nihon Eido) as described previously [57], followed by immunostaining using antibodies (described above). The anti-Pre1p and anti-Rpn4p rabbit polyclonal antibodies were diluted to 1:50,000, and anti-multi ubiquitin mouse monoclonal antibody (FK2) was diluted to 1:60,000. The ECL plus Western Blotting Detection System protocol for blocking, primary and secondary antibody (anti-Rabbit IgG, Horseradish peroxidase linked whole antibody; from donkey) incubation was followed exactly as described (Amersham Biosciences). Immunoassayed proteins were visualized on an X-ray film (X-OMAT AR, Kodak, Tokyo, Japan) using an enhanced chemiluminescence protocol according to the manufacturer's directions (Amersham Biosciences).
List of abbreviations
SSA: stress seventy family A
RT-PCR: reverse transcription polymerase chain reaction
HSP: heat-shock protein
2-DGE: two-dimensional gel electrophoresis
CHAPS: 3- [(3-cholamidopropyl) dimethylaminol]-1-propanesulfonate
EDTA: ethylenediaminetetraacetic acid
DTT: dithiothreitol
IEF: isoelectric focusing
SDS-PAGE: sodium dodecyl sulfate-polyacrylamide gel electrophoresis
PVDF: polyvinylidene difluoride
CYGD: Comprehensive Yeast Genome Database
MIPS: Munich International Center of Protein Sequence
GEO: Gene Expression Omnibus Database
UPR: unfolded protein response
ERAD: endoplasmic reticulum associated degradation
Authors' contributions
RM planned and designed the study, performed the experiments and the data analysis, wrote the main draft of the paper, and generated the figures. KA organized all the research, and provided advice for preparing the manuscript. RR designed the protein experiments and RT-PCR analysis, and contributed in figure making and in editing the manuscript. HI planned all the research and designed the experiments, and suggested the draft of the paper. All authors read and approved the final manuscript.
Acknowledgements
The authors appreciate the helpful comments from the anonymous referees which helped improve the manuscript. We thank Dr Elizabeth A Craig for providing the JN14 and the JN54 strains. We also thank Dr. Yuko Momose, Ms. Emiko Kitagawa, Dr. Yoshinori Murata, Mr. Mine Odani and Dr. Satomi Murata-Mizukami for advice relating to cDNA microarray analysis, and Ms. Junko Shibato for supporting the protein experiments.
==== Refs
Nover L Nover L Inducers of HSP synthesis: heat shock and chemical stressors Heat Shock Response 1991 Florida: CRC Press 5 40
Craig EA The heat shock response Crit Rev Biochem 1985 18 239 280 2412760
Lindquist S The heat shock response Annu Rev Biochem 1986 55 1151 1191 2427013 10.1146/annurev.bi.55.070186.005443
Werner-Washburne M Stone DE Craig EA Complex interactions among members of an essential subfamily of hsp70 genes in Saccharomyces cerevisiae Mol Cell Biol 1987 7 2568 2577 3302682
Lopez-Buesa P Pfund C Craig EA The biochemical properties of the ATPase activity of a 70-kDa heat shock protein (Hsp70) are governed by the C-terminal domains Proc Natl Acad Sci USA 1998 95 15253 15258 9860955 10.1073/pnas.95.26.15253
Shaner L Trott A Goeckeler JL Brodsky JL Morano KA The function of the yeast molecular chaperone Sse1 is mechanistically distinct from the closely related hsp70 family J Biol Chem 2004 279 21992 22001 15028727 10.1074/jbc.M313739200
Craig EA Kramer J Shilling J Werner-Washburne M Holmes S Kosic-Smithers J Nicolet CM SSC1, an essential member of the yeast HSP70 multigene family, encodes a mitochondrial protein Mol Cell Biol 1989 9 3000 3008 2674677
Liu Q Krzewska K Liberek K Craig EA Mitochondrial Hsp70 Ssc1: role in protein folding J Biol Chem 2001 276 6112 6118 11096111 10.1074/jbc.M009519200
Rose MD Misra LM Vogel JP KAR2, a karyogamy gene, is the yeast homolog of the mammalian BiP/GRP78 gene Cell 1989 57 1211 1221 2661018 10.1016/0092-8674(89)90058-5
Tokunaga M Kawamura A Kohno K Purification and characterization of BiP/Kar2 protein from Saccharomyces cerevisiae J Biol Chem 1992 267 17553 17559 1325440
Okamura K Kimata Y Higashio H Tsuru A Kohno K Dissociation of Kar2p/BiP from an ER sensory molecule, Ire1p, triggers the unfolded protein response in yeast Biophys Biochem Res Commun 2000 279 445 450 10.1006/bbrc.2000.3987
Young BP Craven RA Reid PJ Willer M Stirling CJ Sec63p and Kar2p are required for the translocation of SRP-dependent precursors into the yeast endoplasmic reticulum in vivo EMBO J 2001 20 262 271 11226176 10.1093/emboj/20.1.262
Ingolia TD Slater MR Craig EA Saccharomyces cerevisiae contains a complex multigene family related to the major heat shock-inducible gene of Drosophila Mol Cell Biol 1982 2 1388 1398 6761581
Craig EA Jacobsen K Mutations of the heat inducible 70 kilodalton genes of yeast confer temperature sensitive growth Cell 1984 38 841 849 6386178 10.1016/0092-8674(84)90279-4
Nelson RJ Heschl MF Craig EA Isolation and characterization of extragenic suppressors of mutations in the SSA hsp70 genes of Saccharomyces cerevisiae Genetics 1992 131 277 285 1644272
Halladay JT Craig EA A heat shock transcription factor with reduced activity suprpresses a yeast HSP70 mutant Mol Cell Biol 1995 15 4890 4897 7651408
Werner-Washburne M Becker J Kosic-Smithers J Craig EA Yeast Hsp70 RNA levels vary in response to the physiological status of the cell J Bacteriol 1989 171 2680 2688 2651414
Sanchez Y Parsell DA Taulien J Vogel JL Craig EA Lindquist S Genetic evidence for a functional relationship between Hsp104 and Hsp70 J Bacteriol 1993 175 6484 6491 8407824
Becker J Walter W Yan W Craig EA Functional interaction of cytosolic hsp70 and a DnaJ-related protein, Ydj1p, in protein translocation in vivo Mol Cell Biol 1996 16 4378 4386 8754838
Bush GL Meyer DI The refolding activity of the yeast heat shock proteins Ssa1 and Ssa2 defines their role in protein translocation J Cell Biol 1996 135 1229 1237 8947547 10.1083/jcb.135.5.1229
Kim S Schilke B Craig EA Horwich AL Folding in vivo of a newly translated yeast cytosolic enzyme is mediated by the SSA class of cytosolic yeast Hsp70 proteins Proc Natl Acad Sci USA 1998 95 12860 12865 9789005 10.1073/pnas.95.22.12860
Shulga N James P Craig EA Goldfarb DS A nuclear export signal prevents Saccharomyces cerevisiae Hsp70 Ssb1p from stimulating nuclear localization signal-directed nuclear transport J Biol Chem 1999 274 16501 16507 10347213 10.1074/jbc.274.23.16501
Ziegelhoffer T Lopez-Buesa P Craig EA The dissociation of ATP from hsp70 of Saccharomyces cerevisiae is stimulated by both Ydj1p and peptide substrates J Biol Chem 1995 270 10412 10419 7737974 10.1074/jbc.270.18.10412
Wegele H Haslbeck M Reinstein J Buchner J Sti1 is a novel activator of the Ssa proteins J Biol Chem 2003 278 25970 25976 12716905 10.1074/jbc.M301548200
Glover JR Lindquist S Hsp104, Hsp70, and Hsp40: a novel chaperone system that rescues previously aggregated proteins Cell 1998 94 73 82 9674429 10.1016/S0092-8674(00)81223-4
Zolkiewski M ClpB cooperates with DnaK, DnaJ and GrpE in suppressing protein aggregation. A novel multi-chaperone system from Escherichia coli J Biol Chem 1999 274 28083 28086 10497158 10.1074/jbc.274.40.28083
Krobitsch S Lindquist S Aggregation of Huntingtin in yeast varies with the length of polyglutamine expansion and the expression of chaperone proteins Proc Natl Acad Sci USA 2000 97 1589 1594 10677504 10.1073/pnas.97.4.1589
Meriin AB Zhang X He X Newnam GP Chernoff YO Sherman MY Huntington toxicity in yeast model depends on polyglutamine aggregation mediated by a prion-like protein Rnq1 J Cell Biol 2002 157 997 1004 12058016 10.1083/jcb.200112104
Brown CR McCann JA Chiang HL The heat shock protein Ssa2p is required for import of Fructose-1,6-Bisphosphatase into Vid vesicles J Cell Biol 2000 150 65 76 10893257 10.1083/jcb.150.1.65
Satyanarayana C Schröder-Köhne S Craig EA Schu PV Horst M Cytosolic Hsp70s are involved in the transport of aminopeptidase 1 from the cytoplasm into the vacuole FEBS Lett 2000 470 232 238 10745074 10.1016/S0014-5793(00)01324-7
Fujimuro M Sawada H Yokosawa H Production and characterization of monoclonal antibodies specific to multi-ubiquitin chains of polyubiquitinated proteins FEBS Lett 1994 349 173 180 7519568 10.1016/0014-5793(94)00647-4
Takada K Hirakawa T Yokosawa H Okawa Y Taguchi H Ohkawa K Isolation of ubiquitin-E2 (ubiquitin conjugating enzyme) complexes from erythroloukaemia cells using immunoaffinity techniques Biochem J 2001 356 199 206 11336652 10.1042/0264-6021:3560199
Seufert W Jentsch S Ubiquitin-conjugating enzymes UBC4 and UBC5 mediate selective degradation of short-lived and abnormal proteins EMBO J 1990 9 543 550 2154373
Arnason T Ellison MJ Stress resistance in Saccharomyces cerevisiae is strongly correlated with assembly of a novel type of multiubiquitin chain Mol Cell Biol 1994 14 7876 7883 7969127
Baxter BK Craig EA Isolation of UBP3, encoding a de-ubiquitinating enzyme, as a multicopy suppressor of a heat-shock mutant strain of Saccharomyces cerevisiae Curr Genet 1998 33 412 419 9644204 10.1007/s002940050354
Hochstrasser M Ubiquitin-dependent protein degradation Annu Rev Genet 1996 30 405 439 8982460 10.1146/annurev.genet.30.1.405
Ng DT Spear ED Walter P Unfolded protein response regulates multiple aspects of secretory and membrane protein biogenesis and endoplasmic reticulum quality control J Cell Biol 2000 150 77 88 10893258 10.1083/jcb.150.1.77
Johnson ES Ma PC Ota IM Varshavsky A A proteolytic pathway that recognizes ubiquitin as a degradation signal J Biol Chem 1995 270 17442 17456 7615550 10.1074/jbc.270.29.17442
Patil C Walter P Intracellular signaling from the endoplasmic reticulum to the nucleus: the unfolded protein response in yeast and mammals Curr Opin Cell Biol 2001 13 349 356 11343907 10.1016/S0955-0674(00)00219-2
Gasch AP Spellman PT Kao CM Carmel-Harel O Eisen MB Storz G Botstein D Brown PO Genomic Expression programs in the response of yeast cells to environmental changes Mol Biol Cell 2000 11 4241 4257 11102521
Schubert U Antón LC Gibbs J Norbury CC Yewdell JW Bennink JR Rapid degradation of a large fraction of newly synthesized proteins by proteasomes Nature 2000 404 770 774 10783891 10.1038/35004754
Parsell DA Kowal AS Singer MA Lindquist S Protein disaggregation mediated by heat-shock protein Hsp104 Nature 1994 372 475 478 7984243 10.1038/372475a0
Kawai R Fujita K Iwahashi H Komatsu Y Direct evidence for the intracellular localization of HSP104 in Saccharomyces cerevisiae by immunoelectron microscopy Cell Stress Chaperones 1999 4 46 53 10467108
Özcan S Johnston M Three different regulatory mechanisms enable yeast hexose transporter (HXT) genes to be induced by different levels of glucose Mol Cell Biol 1995 15 1564 1572 7862149
Özcan S Johnston M Function and regulation of yeast hexose transporters Microbiol Mol Biol Rev 1999 63 554 569 10477308
Wieczorke R Krampe S Weierstall T Freidel K Hollenberg CP Boles E Concurrent knock-out of at least 20 transporter genes is required to block uptake of hexose in Saccharomyces cerevisiae FEBS Lett 1999 464 123 128 10618490 10.1016/S0014-5793(99)01698-1
Haslbeck M Meiss A Stromer T Walter S Buchner J Disassembling protein aggregates in the yeast cytosol. The cooperation of Hsp26 with Ssa1 and Hsp104 J Biol Chem 2005 280 23861 23868 15843375 10.1074/jbc.M502697200
Cashikar AG Duennwald M Lindquist SL A chaperone pathway in protein disaggregation. Hsp26 alters the nature of protein aggregates to facilitate reactivation by Hsp104 J Biol Chem 2005 280 23869 23875 15845535 10.1074/jbc.M502854200
Xu W Marcu M Yuan X Mimnaugh E Patterson C Neckers L Chaperone-depenedent E3 ubiquitin ligase CHIP mediates a degradative pathway for c-ErbB2/Neu Proc Natl Acad Sci USA 2002 99 12847 12852 12239347 10.1073/pnas.202365899
Kohrer K Domdey H Guthrie C, Funk GR Guide to Yeast Genetics and Molecular Biology Methods Enzymol 1990 194 San Diego: Academic Press 398 401 1706459
Momose Y Iwahashi H Bioassay of cadmium using a DNA microarray: genome-wide expression patterns of Saccharomyces cerevisiae response to cadmium Environ Toxicol Chem 2001 20 2353 2360 11596770 10.1897/1551-5028(2001)020<2353:BOCUAD>2.0.CO;2
MIPS: Munich International Center of Protein Sequence http://mips.gsf.de/
Kitagawa E Momose Y Iwahashi H Correlation of the structures of agricultural fungicides to gene expression in Saccharomyces cerevisiae upon exposure to toxic doses Environ Sci Technol 2003 37 2788 2793 12854720 10.1021/es026156b
GEO: Gene Expression Omnibus Database http://www.ncbi.nlm.nih.gov/geo/
O'Farrell PH High resolution two-dimensional electrophoresis of proteins J Biol Chem 1975 250 4007 4021 236308
Hajduch M Rakwal R Agrawal GK Yonekura M Pretova A High-resolution two-dimentional electrophoresis separation of proteins from metal-stressed rice (Oryza sativa L.) leaves: drastic reductions/ fragmentations of ribulose-1, 5-bisphosphate carbosylate/ oxygenase and induction of stress-related proteins Electrophoresis 2001 22 2824 2831 11545414 10.1002/1522-2683(200108)22:13<2824::AID-ELPS2824>3.0.CO;2-C
Rakwal R Agrawal GK Kubo A Yonekura M Tamogami S Saji H Iwahashi H Defense/ stress responses elicited in rice seedlings exposed to the gaseous air pollutant sulfur dioxide Environ Exp Bot 2003 49 222 235 10.1016/S0098-8472(02)00072-2
|
16209719
|
PMC1262714
|
CC BY
|
2021-01-04 16:32:48
|
no
|
BMC Genomics. 2005 Oct 7; 6:141
|
utf-8
|
BMC Genomics
| 2,005 |
10.1186/1471-2164-6-141
|
oa_comm
|
==== Front
BMC Health Serv ResBMC Health Services Research1472-6963BioMed Central London 1472-6963-5-641620212510.1186/1472-6963-5-64Research ArticleDevelopment of abbreviated measures to assess patient trust in a physician, a health insurer, and the medical profession Dugan Elizabeth [email protected] Felicia [email protected] Mark A [email protected] The Division of Geriatric Medicine, the University of Massachusetts Medical School, Worcester, USA2 The Department of Public Health Sciences, Wake Forest University School of Medicine, Winston-Salem, USA3 The New England Research Institutes, Watertown, USA2005 3 10 2005 5 64 64 16 3 2005 3 10 2005 Copyright © 2005 Dugan et al; licensee BioMed Central Ltd.2005Dugan et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Despite the recent proliferation in research on patient trust, it is seldom a primary outcome, and is often a peripheral area of interest. The length of our original scales to measure trust may limit their use because of the practical needs to minimize both respondent burden and research cost. The objective of this study was to develop three abbreviated scales to measure trust in: (1) a physician, (2) a health insurer, and (3) the medical profession.
Methods
Data from two samples were used. The first was a telephone survey of English-speaking adults in the United States (N = 1117) and the second was a telephone survey of English-speaking adults residing in North Carolina who were members of a health maintenance organization (N = 1024). Data were analyzed to examine data completeness, scaling assumptions, internal consistency properties, and factor structure.
Results
Abbreviated measures (5-items) were developed for each of the three scales. Cronbach's alpha was 0.87 for trust in a physician (test-retest reliability = 0.71), 0.84 for trust in a health insurer (test-retest reliability = 0.73), and 0.77 for trust in the medical profession.
Conclusion
Assessment of data completeness, scale score dispersion characteristics, reliability and validity test results all provide evidence for the soundness of the abbreviated 5-item scales.
==== Body
Background
Trust is a key element of therapeutic relationships. Patient trust may influence health status through continuity of care, adherence to treatment regimens, the willingness to seek care [1,2], and perhaps via the mind~body pathway, which is not yet well understood. Biomedical researchers have paid increasing attention to trust as theoretical and measurement developments have occurred [3-45], driven in part by a concern about the potential negative influence of managed care on the doctor-patient relationship. Additional causes for concern about the doctor-patient relationship include the near daily release of conflicting health information about diet, lifestyle, or medications, and well-publicized, yet rarely occurring, outrageous examples of malpractice and medical errors.
The relationship between doctors and their patients has received philosophical, legal, and literary attention since Hippocrates, and is the subject of more than 8,000 articles monographs, chapters, and books in the modern medical literature [3]. At the conclusion of an extended, competitive, and expensive period of education and training required to enter the profession of medicine, physicians take a vow to do no harm to their patients. For more than a century the American Medical Association has had a code of ethics that states that the chief aim of medicine is to render service to humanity [26].
Mechanic has described trust as the "glue" that holds communities together and allows us to pursue our affairs without excessive suspicion, policing, and regulation [14]. We define patient trust as the optimistic acceptance of a vulnerable situation in which the patient believes the healthcare provider will take care of the patient's interests [5]. This recognizes that the patient-provider relationship involves vulnerability that stems from the experience of illness, the profound imbalance of knowledge and power, and the importance of what is at stake: one's health and well-being [24,25]. Put simply, if there is no vulnerability, there is no need for trust. The object of trust may be a healthcare provider, a hospital or clinic, a health insurance provider, or the medical system as a whole [5].
Our previous research reported on the development and validation of three instruments to measure trust in: 1) a doctor (or other healthcare provider) [6], 2) a health insurer [7], and 3) in doctors in general (e.g., the medical profession) [8]. Interested readers are directed to a detailed description of the conceptual framework of trust that guided this work [5]. Briefly, the framework posits that patient trust involves patients' vulnerability and their resulting reliance on and confidence in their physicians' competence, motivation, honesty, and confidentiality.
Despite the recent proliferation in research on patient trust, it is seldom a primary outcome, and is often just one of several peripheral areas of interest. Thus, the length of the original scales (10 and 11 items) limits the likelihood the scales will be widely used because of the frequent practical needs to minimize both respondent burden and research cost. Because of these concerns the present paper reports on the feasibility, factor structure, reliability and validity of abbreviated versions of the three instruments.
Methods
Samples
Sample 1: National sample. The first sample was selected by random-digit dialing. Inclusion criteria were: age ≥ 21; had health insurance (n = 151 excluded); had visited a healthcare provider at least twice in the past two years (n = 248 excluded); able to speak and understand English; and, able to complete a telephone survey. Contacts with the 2172 potentially eligible adults resulted in the following dispositions: 1117 (51%) provided verbal informed consent and were interviewed; 571 (26%) refused; 484 (22%) were unable to participate (e.g., no answer after 15 callback attempts, too ill, or not able to speak and understand English). Complete data were obtained from 1064 adults and were used in analyses.
Sample 2: North Carolina health maintenance organization (HMO) sample. The second dataset was a random sample of enrollees in a managed care plan who resided in North Carolina [9]. This sample included English-speaking adults aged 21 or older, who had been with the HMO for at least 2 years, and had made at least 2 visits to a primary care provider. Telephone contact was made with 1,908 (94.4%) resulting in the following dispositions: 319 (17%) were ineligible, 378 (20%) refused, and 1,211 (76%) provided verbal informed consent and agreed to participate. Complete data were obtained from 1,045 adults and were used in analyses. Two months later, a random subsample of 306 of these participants was resurveyed to assess test-retest reliability.
The telephone interviews averaged 35 minutes and were conducted by trained interviewers at the Survey Research Center of the University of South Carolina using computer assisted telephone interviewing. Verbal informed consent was obtained at the start of the telephone interviews. To ensure the adequate protection of human subjects in research, the study protocols were reviewed and approved by the Wake Forest University Medical Center Institutional Review Board.
Measurement
The interviews collected information about patient and physician demographic characteristics; the name and type of health insurance; numerous items relating to trust in the subject's personal physician, including Kao's scale to measure physician trust in managed care members [20,21]; patient satisfaction with care [35]; single items to assess satisfaction in the doctor, insurer of interest, and doctors in general ("Overall, you are extremely satisfied with [doctor; health insurer; doctors in general]" coded (1) strongly agree to (5) strongly disagree). Also ascertained were self-reported adherence to doctor recommendations ("You always follow doctors' recommendations about treatment" responses: (1) strongly agree to (5) strongly disagree); whether one would recommend the doctor or insurer to family and friends (responses: (1) strongly agree to (5) strongly disagree); ever been upset or had a serious dispute with doctor [or, insurer] (yes, no); whether one had enough choice of doctor and insurer (yes, no); desire to switch doctor or insurer (yes, no); length of relationship with doctor and insurer of interest (number of years); self-reported physical and mental health (excellent, very good, good, fair or poor).
To reduce respondent burden the national sample was randomly divided; half were asked a battery of questions about health insurance trust, and the other half were asked a battery of questions about trust in doctors in general. Complete data were obtained from 410 adults on the health insurance trust items, and 502 adults on the medical system trust items. There were no statistically significant differences between the two samples on age, race, gender, or health status. All of the above measures were collected in the National sample, while the North Carolina sample did not collect information about trust in doctors in general, overall satisfaction, the willingness to recommend to family or friends, or the Kao trust scale.
Statistical analyses
The abbreviated scales were developed using data from the national sample and then validated with data from both samples. The items were drawn from the original 10 or 11 item scales, which where constructed using psychometric analyses focused on feasibility, factor structure, validity, and reliability described in detail elsewhere [6-8]. This same approach was used to develop the abbreviated scales. Feasibility analyses examined data completeness, floor and ceiling effects, and the dispersion of scores. The item response distributions were examined. Items were deemed not feasible and dropped from the scale if there was a high rate of missing data or responses were concentrated in one or two categories indicating a lack of power to discriminate.
The objective was to develop a 5-item or shorter scale. Items were selected so that the abbreviated form: reflected the content of the conceptual model (competence, honesty, fidelity, and global trust); and contained both positively and negatively worded items. When two questions were otherwise equivalent, the one with the higher factor loading was chosen. Exploratory iterated principal components factor analysis with squared multiple correlations as initial communality estimates was performed to examine dimensionality. Items with absolute factor loadings of <0.60 were identified, and subsequent items were dropped until 100% of the variance was explained.
Correlations between the 5-item scale and the original 10 or 11 item scale were examined, as well as correlations between the 5-item scale and key theoretically determined concepts (e.g., Kao's trust scale, general satisfaction with care, number of years with doctor or insurer, with desire to switch doctor or insurer, number of visits, satisfaction with physician or insurer, willingness to recommend to friends, and whether doctors' recommendations are always followed). A two-sample t-test was used for those variables with a binary response format (e.g., prior dispute with doctor or health insurer, having changed doctors, any or enough choice in selecting doctor or health insurance, having sought a second opinion, and membership in managed care).
Internal consistency was determined by Cronbach's alpha. Test-retest reliability could only be calculated using data from the North Carolina HMO sample for the physician trust and health insurance trust 5-item scales.
Results
A description of the two samples is reported in Table 1. The samples were similar in demographic characteristics; however the mean level of trust in insurer and physician in the North Carolina HMO sample was higher than that of the National sample.
Table 1 Demographic characteristics
National sample (N = 1064) HMO sample (N = 1045)
Age (mean) 49.75 years 46.56 years
Female 68% 55%
Hispanic ethnicity 5% 2%
Race
African American 10% 12%
American Indian or Alaska Native 1% 1%
Asian or Pacific Islander 2% .5%
White 84% 86%
Other 4% 0%
Education
< High School 8% 6%
High School Graduate 28% 28%
>High School 64% 66%
Physical Health
Excellent or Very Good 57% 58%
Good 28% 33%
Fair or Poor 14% 10%
Mental Health
Excellent or Very Good 72% 72%
Good 24% 25%
Fair or Poor 5% 3%
*Percentages may not equal 100% due to rounding.
Patient trust in a physician
Validity
Construct and concurrent validity were examined by correlation analyses and two-sample t-tests for items with binary responses. Table 2 reports the correlations for the 5-item scale in the National sample and the North Carolina HMO sample. Trust in a physician was correlated with: satisfaction with the physician; would recommend to friends and family; general satisfaction with care; no desire to switch to another doctor; number of years under physician's care; number of visits to physician. All correlations were significant at the p < 0.001 level. Binary validation analyses showed that trust was associated with having enough choice in the selection of the physician, not having had a dispute with the physician, and not having sought a second opinion due to concerns about care. Trust generally decreased with poorer physical health (Wilcoxon tests, p = 0.004). Trust also generally decreased with poorer mental health (Wilcoxon tests, p = 0.012). Trust did not vary by education level or income.
Table 2 The association of patient trust in a physician and key variables.
5-item scale national sample 5-item scale HMO sample
Satisfaction with the physician 0.729 0.778
Would recommend physician 0.726 Na
General satisfaction with care 0.478 Na
Desire to change physicians -0.660 -0.686
Number of years under dr.'s care 0.120 0.093
Number of visits to physician 0.127 0.150
Pearson correlation coefficients for continuous variables, Spearman correlation coefficients for categorical variables. All correlations significant at p < 0.001.
Reliability
The 5-item scale had a Cronbach's alpha of 0.87 in the National sample, and in the North Carolina HMO sample it was 0.86. As would be expected any time a scale is reduced in length, the reliability declined, albeit modestly. The 5-item scale had a lower internal consistency than either the original 10-item Wake Forest Scale (0.92) or the Kao scale (0.93).
Summary
There is strong evidence that a 5-item scale can be used to assess a patient's trust in her/his doctor. The 5-item scale is one-dimensional. Responses are summed and scores are on a 5–25 scale, with higher values indicating more trust. Ceiling and floor effects were acceptable. Flesch-Kincaid reading level of the scale was grade 4.3. The mean of the scale was 20.43, with a standard deviation of 3.13. The skewness was -1.05, and the reported kurtosis was 2.52.
Trust in the medical profession
Validity
Construct and concurrent validity were examined by correlation analyses and two-sample t-tests (Table 3) . Trust in the medical profession was correlated with the Kao scale (r = 0.313), general satisfaction with care (r = 0.482), and following doctor's recommendations (r = 0.440).
The binary response validations showed that lower trust was related to having had a dispute with a physician, having changed doctors, and having sought a second opinion. Reported trust was lower for those with poorer mental health (Wilcoxon tests, p = 0.012). Trust did not vary by education level or income.
Reliability
Cronbach's alpha for the 5-item scale was 0.77. The 5-item scale has a lower internal consistency than the original 10-item scale, but is acceptable. No test-retest reliability data were available because the questions about trust in the medical profession were not included in the North Carolina HMO survey.
Summary
There is adequate evidence that the 5-item scale can be used to assess a patient's trust in the medical profession. The 5-item scale is one-dimensional. Responses are summed and scores are on a 5–25 scale, with higher values indicating more trust. Flesch-Kincaid reading grade level is 5.5. The mean of the 5-item scale was 14.97, with a standard deviation of 3.38. The skewness was -1.149, and the reported kurtosis was -0.330.
Trust in a health insurer
Validity
Construct and concurrent validity were examined by correlation analyses and two-sample t-tests (Table 4) . Trust was correlated with Kao's trust scale (r = 0.279), general satisfaction with care (r = 0.465), satisfaction with health insurer (r = 0.646), and desire to find another health insurance provider (r = -0.753). Binary response validations showed that trust was related to having any choice in selecting health insurer, having enough choice in selecting health insurer, having a past dispute with the health insurer, and being in managed care. Adults with poorer mental health had significantly lower trust in their health insurance provider than adults in better mental health.
Reliability
The 5-item scale had a Cronbach's alpha of 0.84 in the National sample, and 0.83 in the North Carolina HMO sample. Test-retest reliability of the trust in health insurance provider 5-item scale was 0.729 in the general population.
Summary
There is evidence that the 5-item scale can be used to assess a patient's trust in a health insurer. Responses are summed and scores are on a 5–25 scale, with higher values indicating more trust. The mean score was 16.57 with a 3.94 standard deviation. The skewness was -0.729, kurtosis 0.339. Flesch-Kincaid reading grade level is 7.7. The scale is one-dimensional in the general population and explains 100% of the variance.
Conclusion
Trust in a medical provider, a health insurer, and the medical profession may be influenced by many factors. As financing pressures continue to force the rapid evolution of the healthcare environment, particularly the patient-provider relationship, research to understand the consequences of such changes will only grow in importance. The 5-item scales developed in this study provide tools to facilitate such research.
Development of the 5-item scales was informed by our theoretical model and data driven. We sought to develop scales that provide sufficient measurement precision and breadth, yet minimize burden and cost. The scales are brief, comprehensive and empirically validated tools. The scales require reading levels of grades 4.3 (physician trust), 5.5 (medical profession trust), and 7.5 (health insurer trust). Each 5-item scale had acceptable psychometric properties.
Several limitations of the current research should be noted. First, the results reported here are for telephone administration of the scales. The performance of the scales in other settings is yet unknown. Second, the interviews were only conducted with English-speaking adults, although subsequent research is currently in press reporting on Spanish translations of some of these items. Further research on older adults, the most frequent users of healthcare, is urgently needed. Research to determine the effectiveness of interventions to enhance the doctor-patient relationship, and whether such enhancements, by extension, will improve important patient outcomes, is also needed.
Appendix
Patient trust in a physician
*1. Sometimes Dr._ [INSERT NAME OF DR.]__ cares more about what is convenient for (him/her) than about your medical needs.
2. Dr. _ [INSERT NAME OF DR.]_ is extremely thorough and careful.
3. You completely trust Dr._ [INSERT NAME OF DR.]'s decisions about which medical treatments are best for you.
4. Dr._ [INSERT NAME OF DR.]__ is totally honest in telling you about all of the different treatment options available for your condition.
5. All in all, you have complete trust in Dr._ [INSERT NAME OF DR.]_.
Response choices (coding) are: Strongly Agree (5), Agree (4), Neutral (3), Disagree (2), Strongly Disagree (1). Responses are summed (range 5–25) with higher scores indicating more trust. *Negatively worded item is reverse coded.
Patient trust in the medical profession
*1. Sometimes doctors care more about what is convenient for them than about their patients' medical needs.
2. Doctors are extremely thorough and careful.
3. You completely trust doctors' decisions about which medical treatments are best.
4. A doctor would never mislead you about anything.
5. All in all, you trust doctors completely.
Response choices (coding) are: Strongly Agree (5), Agree (4), Neutral (3), Disagree (2), Strongly Disagree (1). Responses are summed (range 5–25) with higher scores indicating more trust. *Negatively worded item is reverse coded.
Patient trust in a health insurer
*1. [INSERT NAME OF HEALTH INSURER] Cares more about saving money than about getting you the treatment you need.
*2. You feel like you need to double check everything [INSERT NAME OF HEALTH INSURER ] does.
3. You believe [INSERT NAME OF HEALTH INSURER] will pay for everything it is supposed to, even really expensive treatments.
4. If you have a question, you think [INSERT NAME OF HEALTH INSURER] will give you a straight answer.
5. All in all, you have complete trust in [INSERT INSURER'S NAME].
Response choices (coding) are: Strongly Agree (5), Agree (4), Neutral (3), Disagree (2), Strongly Disagree (1). Responses are summed (range 5–25) with higher scores indicating more trust. *Negatively worded item is reverse coded.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Obtained research funding: MH, ED.
Research idea: ED.
Data collection: MH, ED
Statistical Analysis: FT.
Writing, revising, and final approval of manuscript: ED, MH, FT.
Table 3 The association of trust in the medical profession and key variables.
Original 11-item scale national sample 5-item scale national sample
Kao's trust scale 0.306 0.313
General satisfaction 0.498 0.482
Follow doctor's recommendations 0.449 0.440
Original WFU 11 item scale 0.957
Pearson correlation coefficients for continuous variables, Spearman correlation coefficients for categorical variables. All correlations significant at p < 0.001.
Table 4 The association of trust in health insurer and key variables.
5-item scale national sample 5-item scale HMO sample
Kao's trust scale 0.279
General satisfaction 0.465
Satisfaction with insurer 0.646
Desire to switch insurers -0.753 -0.589
Original WFU 10 item scale 0.952 0.948
Pearson correlation coefficients for continuous variables, Spearman correlation coefficients for dichotomous variables. All correlations significant at p < 0.001.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Research supported by the Robert Wood Johnson Foundation and the University of Massachusetts Medical School Division of Geriatric Medicine.
==== Refs
Safran DG Montgomery JE Chang H Murphy J Rogers WH Switching doctors: voluntary disenrollment from a primary physicians practice J Fam Pract 2001 50 130 136 11219560
Safran DG Taira DA Rogers WH Kosinkski M Ware JE Tarlov AR Linking primary care performance to outcomes of care J Fam Pract 1998 47 213 220 9752374
Safran DG Kosinski M Tarlov AR Rogers WH Taira DH Lieberman N Ware JE The Primary Care Assessment Survey: tests of data quality and measurement performance Med Care 1998 36 728 739 9596063 10.1097/00005650-199805000-00012
Murphy J Chang H Montgomery JE Rogers WH Safran DG The quality of physician-patient relationships J Fam Pract 2001 50 123 129 11219559
Hall M Dugan E Zheng B Mishra A Trust in physicians and medical institutions: what is it, can it be measured, and does it matter? Milbank Q 2001 79 613 639 11789119 10.1111/1468-0009.00223
Hall M Zheng B Dugan E Camacho F Kidd K Mishra A Balkrishnan R Measuring patients' trust in their primary care physician Med Care Res Rev 2002 59 293 318 12205830 10.1177/1077558702059003004
Zheng B Hall MA Dugan E Kidd KE Levine D Development of a scale to measure patients' trust in health insurers Health Services Research 2002 37 187 202 11949920
Hall MA Camacho F Dugan E Balkrishnan R Trust in the medical profession: conceptual and measurement issues Health Serv Res 2002 37 1419 1439 12479504 10.1111/1475-6773.01070
Hall M Dugan E Balkrishnan R Bradley D How disclosing HMO physician incentives affects trust Health Aff 2002 21 197 206 10.1377/hlthaff.21.2.197
Balkrishnan R Dugan E Camacho FT Hall MA Trust and satisfaction with physicians, insurers, and the medical profession Med Care 2003 41 1058 1064 12972845 10.1097/01.MLR.0000083743.15238.9F
Bonds DE Foley KL Dugan E Hall MA Extrom P An exploration of patients' trust in physicians in training J Health Care Poor Underserved 2004 15 294 306 15253380
Trachtenberg F Dugan E Hall MA How patients' trust relates to their involvement in medical care J Fam Pract 2005 54 344 352 15833226
Mechanic D Changing medical organization and the erosion of trust Milbank Q 1996 74 171 189 8632733
Mechanic D The functions and limitations of trust in the provision of medical care J Health Polit Policy Law 1998 23 661 686 9718518
Mechanic D Meyer S Concepts of trust among patients with serious illness Soc Sci Med 2000 51 657 668 10975226 10.1016/S0277-9536(00)00014-9
Thom DH Ribisl KM Stewart AL Luke DA Further validation and reliability testing of the Trust in Physician Scale Med Care 1999 37 510 517 10335753 10.1097/00005650-199905000-00010
Thom DH the Stanford Trust Study Physicians Physician behaviors that predict trust J Fam Pract 2001 50 323 328 11300984
Thom DH Training physicians to increase patient trust J Eval Clin Pract 2000 6 245 253 11083035 10.1046/j.1365-2753.2000.00249.x
Thom DH Hall MA Pawlson LG Measuring patients' trust in physicians when assessing quality of care Health Aff 2004 23 124 132 10.1377/hlthaff.23.4.124
Kao AC Green DC Zaslavsky AM Koplan JP Cleary PD The relationship between method of physician payment and patient trust JAMA 1998 280 1708 1714 9832007 10.1001/jama.280.19.1708
Kao AC Green DC Davis NA Koplan JP Cleary PD Patients' trust in their physicians: effects of choice, continuity, and payment method J Gen Intern Med 1998 13 681 686 9798815 10.1046/j.1525-1497.1998.00204.x
Anderson LA Dedrick RF Development of the Trust in Physician scale: a measure to assess interpersonal trust in patient-physician relationships Psychol Rep 1990 67 1091 1100 2084735
Freburger JK Callahan LF Currey SS Anderson LA Use of the Trust in Physician Scale in patients with rheumatic disease: psychometric properties and correlates of trust in the rheumatologist Arthritis Rheum 2003 49 51 58 12579593 10.1002/art.10925
Goold SD Trust, distrust and trustworthiness J Gen Intern Med 2002 17 79 81 11903779 10.1046/j.1525-1497.2002.11132.x
Goold SD Klipp G Managed care members talk about trust Soc Sci Med 2002 54 879 888 11996022 10.1016/S0277-9536(01)00070-3
Goold SD Lipkin M The doctor-patient relationship: Challenges, opportunities, and strategies J of Gen Intern Med 1999 14 S26 33 9933492 10.1046/j.1525-1497.1999.00267.x
Shore DA The trust prescription for healthcare: building your reputation with consumers 2005 Health Administration Press, Chicago, IL.
Gray BH Trust and trustworthy care in the managed care era Health Aff 1997 16 34 49 10.1377/hlthaff.16.1.34
Pearson SD Raeke LH Patients' trust in physicians: many theories, few measures, and little data J Gen Intern Med 2000 15 509 513 10940139 10.1046/j.1525-1497.2000.11002.x
Freburger JK Callahan LF Currey SS Anderson LA Use of the Trust in Physician Scale in patients with rheumatic disease: psychometric properties and correlates of trust in the rheumatologist Arthritis Rheum 2003 49 51 58 12579593 10.1002/art.10925
Caterinicchio RP Testing plausible path models of interpersonal trust in patient-physician treatment relationships Soc Sci Med 1979 13A 81 99
Boulware LE Cooper LA Ratner LE LaVeist TA Powe NR Race and trust in the health care system Public Health Reports 2003 118 358 365 12815085
Doescher MP Saver BG Franks P Fiscella K Racial and ethnic disparities in perceptions of physician style and trust Arch Fam Med 2000 9 1156 1163 11115223 10.1001/archfami.9.10.1156
Ramsay J The General Practice Assessment Survey (GPAS): tests of data quality and measurement properties Fam Pract 2000 7 372 9 11021894 10.1093/fampra/17.5.372
Hall JA Feldstein M Fretwell D Rowe JW Epstein AM Older patients' health status and satisfaction with medical care in an HMO population Med Care 1990 28 261 70 2314135
Greenfield S Kaplan SH Ware JE Yano EM Frank HJL Patients' participation in medical care: Effects on blood sugar control and quality of life in diabetes J Gen Intern Med 1988 3 448 57 3049968
Greenfield S Kaplan S Ware JE Expanding patient involvement in care: Effects on patient outcomes Ann Intern Med 1985 102 520 528 3977198
LaVeist TA Nickerson KJ Bowie JV Attitudes about racism, medical mistrust, and satisfaction with care among African American and White cardiac patients Med Care Res Rev 2000 57 146 161 11092161 10.1177/107755800773743637
Hunt KA Gaba A Lavizzo-Mourey R Racial and ethnic disparities and perceptions of health care: does health plan type matter? Health Serv Res 2005 40 551 576 15762907 10.1111/j.1475-6773.2005.0z373.x
Corbie-Smith G Thomas SB Williams MW Moody-Ayers S Attitudes and beliefs of African Americans toward participation in medical research J Gen Intern Med 1999 14 537 546 10491242 10.1046/j.1525-1497.1999.07048.x
Corbie-Smith G Thomas SB St. George DMM Distrust, race, and research Arch Intern Med 2002 162 2458 2463 12437405 10.1001/archinte.162.21.2458
Bouleware LE Cooper LA Ratner LE LaVeist TA Powe NR Race and trust in the health care system Public Health Rep 2003 188 366 367
Corbie-Smith G Ammerman AS Katz ML St. George DMM Blumenthal C Washington C Weathers B Keyserling TC Switzer B Trust, benefit, satisfaction, and burden: A randomized controlled trial to reduce cancer through African-American Churches J Gen Intern Med 2003 18 531 541 12848836 10.1046/j.1525-1497.2003.21061.x
Doescher MP Saver BG Franks P Fiscella K Racial and ethnic disparities in perceptions of physician style and trust Arch Fam Med 2000 9 1156 1163 11115223 10.1001/archfami.9.10.1156
Ahern MM Hendryx MS Social capital and trust in providers Soc Sci Med 2003 57 1195 203 12899904 10.1016/S0277-9536(02)00494-X
|
16202125
|
PMC1262715
|
CC BY
|
2021-01-04 16:31:50
|
no
|
BMC Health Serv Res. 2005 Oct 3; 5:64
|
utf-8
|
BMC Health Serv Res
| 2,005 |
10.1186/1472-6963-5-64
|
oa_comm
|
==== Front
BMC Health Serv ResBMC Health Services Research1472-6963BioMed Central London 1472-6963-5-661621909910.1186/1472-6963-5-66Study ProtocolDoes a pre-hospital emergency pathway improve early diagnosis and referral in suspected stroke patients? – Study protocol of a cluster randomised trial [ISRCTN41456865] Ferri Marica [email protected] Luca Assunta [email protected] Paolo Giorgi [email protected] Giuliano [email protected] Gabriella [email protected] Agenzia di Sanità Pubblica della Regione Lazio, Via di Santa Costanza, 53, 00198 Roma, Italy2005 11 10 2005 5 66 66 5 7 2005 11 10 2005 Copyright © 2005 Ferri et al; licensee BioMed Central Ltd.2005Ferri et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Early interventions proved to be able to improve prognosis in acute stroke patients. Prompt identification of symptoms, organised timely and efficient transportation towards appropriate facilities, become essential part of effective treatment. The implementation of an evidence based pre-hospital stroke care pathway may be a method for achieving the organizational standards required to grant appropriate care. We performed a systematic search for studies evaluating the effect of pre-hospital and emergency interventions for suspected stroke patients and we found that there seems to be only a few studies on the emergency field and none about implementation of clinical pathways.
We will test the hypothesis that the adoption of emergency clinical pathway improves early diagnosis and referral in suspected stroke patients. We designed a cluster randomised controlled trial (C-RCT), the most powerful study design to assess the impact of complex interventions. The study was registered in the Current Controlled Trials Register: ISRCTN41456865 – Implementation of pre-hospital emergency pathway for stroke – a cluster randomised trial.
Methods/design
Two-arm cluster-randomised trial (C-RCT). 16 emergency services and 14 emergency rooms were randomised either to arm 1 (comprising a training module and administration of the guideline), or to arm 2 (no intervention, current practice). Arm 1 participants (152 physicians, 280 nurses, 50 drivers) attended an interactive two sessions course with continuous medical education CME credits on the contents of the clinical pathway. We estimated that around 750 patients will be met by the services in the 6 months of observation. This duration allows recruiting a sample of patients sufficient to observe a 30% improvement in the proportion of appropriate diagnoses.
Data collection will be performed using current information systems. Process outcomes will be measured at the cluster level six months after the intervention. We will assess the guideline recommendations for emergency and pre-hospital stroke management relative to: 1) promptness of interventions for hyperacute ischaemic stroke; 2) promptness of interventions for hyperacute haemorrhagic stroke 3) appropriate diagnosis. Outcomes will be expressed as proportions of patients with a positive CT for ischaemic stroke and symptoms onset <= 6 hour admitted to the stroke unit.
Discussion
The fields in which this trial will play are usually neglected by Randomised Controlled Trial (RCT). We have chosen the Cluster-randomised Controlled Trial (C-RCT) to address the issues of contamination, adherence to real practice, and community dimension of the intervention, with a complex definition of clusters and an extensive use of routine data to collect the outcomes.
==== Body
Background
Stroke is the third most common cause of death in developed countries [1]. In 80% of cases stroke is ischaemic (caused by thrombotic or embolic occlusion of cerebral artery) [2] the remainder are caused by intracerebral or subarachnoid haemorrages.
Around 10% of all people with acute ischaemic stroke will die within 30 days of stroke onset, while 50% of the survivors will experience some level of disability after 6 months [3].
As early effective interventions proved to be able to improve prognosis [1], implementation of an evidence based pre-hospital stroke care pathway may be a method for achieving early identification of symptoms, and organised timely and efficient transportation towards appropriate care. Clinical pathways have been indicated with different names: care pathways, critical pathway, critical path method, and Care Maps(tm) [4]. We adopted "clinical pathway" to indicate a document targeted to all emergency service personnel (dispatchers, physicians, nurses and any other professionals, such as ambulances drivers) involved in the management of a suspect acute stroke from pre-hospital to emergency phases.
A systematic review on the adoption of in-hospital clinical pathways suggested that the currently available evidence is insufficient to support routine implementation of care pathways for the hospital management of acute stroke or stroke rehabilitation [5]. Therefore there is still uncertainty about the effectiveness of clinical pathways to achieve earlier interventions for stroke patients in pre-hospital setting. We are performing a systematic review on the effectiveness of clinical pathways in pre-hospital field and we also decided to design the present study. The results of the study will be the basis to decide about the widespread implementation of the clinical pathway in the whole region.
Within Lazio, the region with 5,302,302 inhabitants in which the capital city of Rome is located, the emergency medical services (ES) are accessible by dialling "118" with automatic connection to the nearest dispatch facility. The emergency network is organised in three levels of complexity. The basic level involves emergency rooms for first aid; the first level is equipped for mild-severe diagnosis and the second level for the most severe urgency. These services are in a number and a geographical disposition to grant the required assistance in emergency (see figure 1). Calls for interventions are answered by ES dispatchers who activate the nearest ambulance unit. The Basic Life Service (BLS) facilities are superior in number than the Advanced Life Services (ALS) and are commonly forwarded for the suspected stroke patients. Nowadays, BLS are compelled by law to deliver patients to the nearest emergency hospital. There, medical doctors provide the assessment of patients conditions and diagnosis and decide about referral to appropriate facility. The provision of Computerised Tomography (CT) to confirm the diagnosis of stroke, and the transportation of patients to their final destination, may require some times. With the actual organization of system, the process from the emergency call to the admission to the stroke unit may take from several hours up to days.
Figure 1 Territorial distribution of the entities enclosed in the study.
The availability of effective early interventions for stroke, calls for reengineering these procedures.
The need for prompt diagnosis and referral requires a that more responsibility is placed on nurses (active in the BLS) who should be trained to recognize early symptoms and should be empowered to contact ES in order to identify the appropriate level of care required, and to transport patients directly to stroke units.
An evidence-based clinical pathway was developed with the methods adopted by the most qualified guidelines development agencies (such as Scottish Intercollegiate Guidelines Network, National Institute for Clinical Excellence and the Italian National Plan for Guidelines Development-PNLG) with the involvement of the emergency health workers in the expert panel, and we decided to evaluate its effectiveness in standardising and improving the procedures in pre-hospital (ES) and emergency (ER) setting, before considering its widespread adoption in our region.
The cluster randomised trial design is indicated in assessing the effectiveness of complex interventions [6] and it allows comparisons of alternative strategies such as guidelines versus conventional educational methods of influencing doctors' management of a particular problem. In our study randomization at cluster level is not only indicated but also necessary to avoid hindrance to the normal activity of the services. We, therefore, adopted as unit of randomization the whole teams of ER or ES stations to avoid contamination.
We present the protocol of study that compares the adoption of an evidence-based prehospital pathway versus current practice.
It is our intention to describe the methodology of the study in the present publication in order to ensure independence of the results and to stimulate criticism and suggestions from the journal readers. The protocol has been presented in several international conferences to share our initiative with the scientific community and to collect their comments and ideas.
As we are aware of the many limits and peculiarity of this study design we believe that lessons on how to deal with complex intervention in the difficult environment will be the added value of the present study.
Methods/design
Participants
We identified all the emergency services referring to the two stroke units presently available in Rome (figure 1). There were 47 entities comprising 18 emergency rooms and 29 emergency service stations (52 ambulances) from which we created 20 clusters. The criteria for cluster creation was grouping together the services sharing personnel and/or referring patients each other (figure 2-table 1).
Figure 2 Flow chart of randomization procedure. Overall we had 20 units of randomization (cluster) of which 10 are couples composed of one ES+1ER on the basis of closeness (usually the ES is situated in the same place of the ER, or systematically refers patients in the same closest ER); 6 are ES and 4 are ER which could not be matched.
Table 1 Availability of ambulances and organization of structures in the study area, as reported April 2005.
Place/Area Emergency Medical Services (ES) Emergency room (ER) Notes
Viterbo
Viterbo Call Center with Medical Doctors ER level I
Montalto di Castro Medical doctors (h12 closed for holidays) -
Tarquinia Only nurses ER first aid Staff sharing ES/ER
Tuscania Only nurses (h12) ER first aid
Vetralla Medical doctors -
Ronciglione Only nurses - Staff sharing ES/ER
Civita Castellana Only nurses ER first aid Staff sharing ES/ER
Orte Medical doctors -
Monte Fiascone Only nurses ER first aid Staff sharing ES/ER
Acquapendente Medical doctors ER first aid Staff sharing ES/ER
ASL FR
Anagni 1 full staffed ambulance + 1 BLS ER first aid
Alatri 1 full staffed ambulance + 1 BLS ER first aid
Fiuggi 1 ALS – ambulances with medical doctors (4) -
Ferentino 1 ALS – ambulances with medical doctors (8) -
Frosinone Call Center with medical doctors(8) ER level I
Ceccano 2 BLS ER first aid
Veroli 1 BLS -
Ceprano 2 BLS -
Sora 1 full staffed ambulance + 1 BLS ER first aid
Isola liri 2 BLS -
Atina 1 ALS – ambulance with medical doctor (5) -
Pontecorvo 1 First level (staffed for most severe conditions)+ 1BLS ER first aid
Cassino 2BLS (only inter-hospital transfer) ER level I
ASL RMG
Tivoli 2 BLS ER level I A new ER will be opened (during the study period)
Monterotondo 1 BLS ER first aid ES and ER randomized independently
Palombara 1 BLS -
Palestrina 1 BLS ER first aid ES and ER randomized independently
Colleferro 1 full staffed ambulance(h12) + 2 BLS ER first aid ES and ER randomized independently
Valmontone 1 BLS -
Subiaco 1 BLS ER first aid
Olevano Romano 1 BLS -
Lunghezza 1 ALS – ambulances with medical doctors (4) -
Arsoli 1 BLS -
Montelanico 1 BLS -
ASL RMA
S. Giacomo full staffed ambulance – ambulances with medical doctors ER level I ES and ER randomized independently
Nomentano Red Cross (as ALS) -
Arno-Treviso 3 BLS + 1 full staffed ambulance – ambulances with medical doctors -
Addolorata 3 BLS + 1 ALS – ambulances with medical doctors -
Marcigliana 1 BLS -
Participants are all the workers belonging to the services enclosed in the study (enclosing the ambulances' drivers).
About 152 physicians, 280 nurses, 50 drivers will be trained on the contents of the clinical pathway. We estimated that around 750 patients will be met by the services in the 6 months of observation.
Interventions
Participants in the intervention arm will be trained on the content of the clinical pathway and will be given the clinical pathway itself for consultation and discussion in groups.
The clinical pathway is based on available evidence based pre-hospital and emergency interventions for suspected stroke patients.
It consists of the following main points:
- ES dispatcher uses a short form of Cincinnati pre-Hospital Stroke Scale (CHSS) to identify suspected stroke patients during the telephone call;
- ES health workers confirm diagnosis by CHSS on the scene;
- patients are provided CT and referred to appropriate care.
A group of trainers from the expert panel that developed the clinical pathway, and composed of:
- an ES medical doctor
- anaesthesiologist working on helicopter emergency unit
- two neurologists working in stroke unit
- a physician working in an emergency department
- a neurosurgeon
- two epidemiologist and evidence-based medicine expert
trained a group of health workers selected from all the entities participating in the intervention arm study to act as "facilitators" for peer education. The facilitators trained their colleagues in the workplace with the help of audiovisual materials produced by the teachers themselves.
At least one representative of the teachers' group took part in the meetings on the workplace to support groups' discussion and to answer possible questions. As a result of the training sessions every person working in any entity participating in the intervention arm of the trial have been trained on the content of the pathway.
No interventions will be implemented in the control group.
Objectives
The objective of this cluster randomised controlled trial is to evaluate the effectiveness of pre-hospital and emergency clinical pathway for patients with suspected stroke in: improving early identification of stroke, promoting appropriate triage coding, achieving coherence between diagnosis, interventions and referral.
Outcomes
The outcomes relate to organizational aspects (transportation to the appropriate hospital, accurate diagnosis and interventions, timely treatments) and will be measured as:
Primary outcome
- proportion of patients with a positive CT for ischaemic stroke and symptoms onset <= 6 hour admitted to the stroke unit
Secondary outcomes
- proportion of patients with a positive CT for hemorrhagic stroke and symptoms onset <= 6 hour admitted to a neurosurgical ward
- proportion of patients with a positive CT for hemorrhagic or ischaemic stroke or symptoms onset >6 hour admitted to the nearest hospital
- proportion of ICD9CM code for stroke in emergency setting confirmed in the hospital discharge data
- proportion of patients with stroke confirmed by CT results
- proportion of patients with ischaemic stroke receiving treatment within 6 hour of symptoms onset
All the outcomes will be reported at cluster level and will be cross-checked by the integration of data from the available information systems.
Sample size
We calculated the sample size, i.e. the duration of recruitment, keeping into account the number of emergency calls and the number of emergency room admissions for suspect stroke.
For the ES the mean number of patients per 19 clusters was about 50 (see figure 2). The estimate percentage of correctly transferred patients, based on the Information System of Emergency Rooms in 2003, is 47%. We assumed an intra cluster correlation of ICC = 0.05. Under this assumption we decided 6 months duration, i.e. 25 patients per cluster, to obtain a power of 95% to detect a difference of 50% in rates between the two groups, i.e. reaching about 70% of correctly transferred patients in the treated group, with a = 0.05.
For the emergency rooms the mean number of patients per 14 clusters was about 110. The estimate of correctly transferred patients, based on the Information System of Emergency Rooms in 2003, is 14%. We assumed an intra cluster correlation of ICC = 0.05. Under this assumption and with a duration of 6 months, i.e. 55 patients per cluster, to obtain a power of 95% to detect a difference of 150% in rates between the two groups, i.e. reaching about 35% of correctly transferred patients in the treated group, with a = 0.05.
Sequence generation and allocation
To avoid contamination due to personnel turnover, the entities were grouped into 20 clusters according to geographical nearness and personnel sharing procedures. Clusters were stratified according to their characteristics:
• Couples of one emergency service and one emergency room (usually working together in the same geographical area) (n°10)
• Groups of emergency services (n°6)
• Only emergency rooms (n°4).
Clusters were attributed sequential numbers and sample function of STATA 7 (StataCorp LP 2005) was used to generate random numbers. We utilized the Italian lottery extracted number of 6th November 2004 in Rome as seeds number for generating the random sequences.
Data collection
To assess the impact of the adoption of the clinical pathway over the current practice, we will analyse the data currently available from the actual information systems and no additional information will be collected.
In this way, emergency health workers will not be charged with extra work deriving from registration of ad-hoc information, and they will be completely devoted to the implementation of the pathway itself.
The Agency of Public Health created and maintains the Information System of Emergency Rooms (ISER), the Stroke Surveillance System (SSS) and the Hospital Information System (HIS) sufficient to obtain data for measuring the outcomes. Moreover the Agency for Public Health has access to the Information System of ES 118 (IS118).
Data analysis
• Baseline characteristics of the entities will be compared to ensure randomisation success;
• Analysis will be performed in an Intention to treat basis;
• Other analysis by protocol will be performed as a sensitivity analysis;
• Preliminary analysis will be performed after three month of the beginning of the observation.
All the analysis will be processed with SAS (version 8) and STATA 7 (StataCorp LP 2005).
Ethics
The present protocol was designed following the indication of Helsinki adopted by the 18th World Medical Association General Assembly in June 1964 and following amended in 1975–2002 and clarified respectively for paragraph 29–30 in 2002 and 2004. The ethical committee of the Agency for Public Health approved the protocol with the document n.124 of 15 June 2004. The ethical committee was created according to Helsinki indications including, among others, a member of a patients' rights organization.
An informed consensus was signed by the responsible of each entity participating in the study who received a document containing all the relevant information and copies of the clinical pathway and the study protocol. No informed consensus will be requested to patients [7].
Stopping rules
The nature of the intervention to be experimented prevented us from identifying specific stopping rules and we could only commit ourselves to submitting any possible problems to the ethical committee. The core of the pathway consists of transferring patients likely to benefit from complex interventions to specialized structures, rather than providing them with minimal assistance in the nearest hospital. This means that health workers belonging to experimental clusters may need to transfer patients for longer distance than they used to do. The risks we can foresee are unexpected events during transportation and troubles deriving from the unavailability of ambulances. Health workers have been warned to promptly report any kind of problems to the study coordinator.
Discussion
The emergency and pre-hospital field has not been studied sufficiently in randomised controlled trials, likely reasons are organizational difficulties, critical conditions of patients and the related ethical problems [8].
We are therefore pioneering this kind of study in a difficult environment and we are prepared to face many obstacles. First, there are the changes in the network of the emergency services. During the last twelve months many new entities were created and others were suppressed generating problems in the identification of clusters. Lack of personnel and high turnover increase risk of contamination due to personnel sharing during the summer and the holydays.
Other problems have to do with the local law imposing that patients taken from the scene should be brought to the nearest hospital where the emergency physicians will determine the need (opportunity) to transfer elsewhere for appropriate cure. This procedure delay interventions and determines a very low percentage of cases directly admitted to the stroke unit: 14% of the ambulances transport (calls), and 47% of the emergency room admissions. However health workers participating in the experimentation expressed their worry about possible legal consequences of their decisions caused by the implementation of the protocol study, and this may represent an important obstacle to adherence.
Accuracy of data registration is crucial for process evaluation but our quality control system revealed that, even though a priori criteria for data collection are homogeneous, results are sometimes heterogeneous. Nevertheless, as misreporting is non differential the deriving underestimation may not severely affect results.
Abbreviations
CME = continuous medical education
CT = computer tomography
ES = Emergency Service
ER = Emergency Room
ALS = Advanced Life Support
BLS = Basic Life Support
CHSS = Cincinnati pre-Hospital Stroke Scale
ICD9CM = International Classification Diseases 9th revision Clinical Modification
ICC = intra cluster correlation
ISER = Information System of Emergency Rooms
SSS = Stroke Surveillance System
HIS = Hospital Information System
IS118 = Information System of ES 118
Competing interests
The author(s) declare they have no competing interests.
Authors' contributions
MF wrote the manuscript and assisted in the design of the study;
ADL contributed to the manuscript, ideated and coordinated the study;
PGR assisted in the cluster creation, sequence generation and allocation and analysis planning and contributed to the manuscript;
GL provided data analysis from the information systems and planning of information retrieval for the assessment of outcomes;
GG gave the input of the project, provided overview of all the steps of the study, and contributed to the manuscript final review.
All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
We thank Tom Jefferson for having suggested the adoption of this study design and ideated the study.
Carlo Francia and Stefania Gabriele took care of organisation of training sessions, and contributed to manage the meetings with facilitators.
Andrea Angelini took care of the organization of training sessions and preparation of audiovisual materials.
Fabio Azzeri, Cinzia Barletta, Alessandro Caminiti, Stefano Castaldi, Antonio D'Urso, Marialuisa Sacchetti, Danilo Toni, Andrea Vignati taught in the training sessions, collaborated in preparing audiovisual materials and profused their expertise and enthusiasm to the healthworkers.
Mauro Beccaceci, Marcello Cappuccini, Pierluigi Cervelli, Patrizia Fratini, Gennaro Scialò, Luciano Sistimini, contributed to planning and conducting the workplace education and coordinated the activities in ERs and EMS involved in the study
Vittorio Altomani, Antonio De Santis, Pierluigi Tasciotti heads of ES and Sergio Iacoponi, Gianfranco Leone, Alessandro Masella, Mario Pagliei, Francesco Rocco Pugliese, Patrizio Ricciotti, Maurizio Saccucci heads of ERs authorised their centers to take part in the study and made all the efforts necessary to make the study possible.
==== Refs
Warburton E Stroke management Clin Evid 2004 12 236 252
Bamford J Sandercock P Dennis M Burn J Warlow C A prospective study of acute cerebrovascular disease in the community: the Oxfordshire Community Stroke Project – 1981–86. 2. Incidence, case fatality rates and overall outcome at one year of cerebral infarction, primary intracerebral and subarachnoid haemorrhage J Neurol Neurosurg Psychiatry 1990 53 16 22 2303826
Wade DT Hewer RL Functional abilities after stroke: measurement, natural history and prognosis J Neurol Neurosurg Psychiatry 1987 50 177 82 3572432
Kwan J Sandercock P In-hospital care pathways for stroke Cochrane Database Syst Rev 2004 4
Stroke Unit Trialists' Collaboration Organised inpatient (stroke unit) care for stroke Cochrane Database Syst Rev 2000 2
Donner A Klar N Design and Analysis of Cluster Randomization Trials in Health Research 2000 London: Arnold
Ferri M Jefferson T De Luca A Are the results of hospital based studies generalisable to the pre-hospital setting? Journal of Emergency Primary Health Care 2004 2
Medical Research Council 2000 A Framework For Development And Evaluation Of Rcts For Complex Interventions To Improve Health
|
16219099
|
PMC1262716
|
CC BY
|
2021-01-04 16:31:52
|
no
|
BMC Health Serv Res. 2005 Oct 11; 5:66
|
utf-8
|
BMC Health Serv Res
| 2,005 |
10.1186/1472-6963-5-66
|
oa_comm
|
==== Front
BMC Infect DisBMC Infectious Diseases1471-2334BioMed Central London 1471-2334-5-741617658010.1186/1471-2334-5-74Case ReportLeprosy & gangrene: A rare association; role of anti phospholipid antibodies Akerkar Shashank M [email protected] Lata S [email protected] Lecturer, Rheumatology,Dept of Medicine,Seth GSMC & KEM Hospital, Parel, Mumbai, India – 4000122 Chief Rheumatology,Head Dept of Medicine,Seth GSMC & KEM Hospital, Parel, Mumbai, India – 4000122005 21 9 2005 5 74 74 1 4 2005 21 9 2005 Copyright © 2005 Akerkar and Bichile; licensee BioMed Central Ltd.2005Akerkar and Bichile; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Leprosy still remains an important public health problem for many parts of the world. An association of gangrene with leprosy is a rare one & can have a number of causative mechanisms. We present a case with Leprosy & gangrene with positive anti phopholipid antibody titers.
Case presentation
A 50-year-old non-diabetic, non-hypertensive lady presented with 2 months history of progressive gangrene of bilateral toes. She was found to have madarosis & hypopigmented, hypoaesthetic macular lesions on the upper limb & thighs. Bilateral ulnar & popliteal nerves were thickened. A skin biopsy of the lesions revealed borderline tuberculoid leprosy, slit skin smears revealed a bacteriological index of 1+. She did not have any evidence of thromboembolic episode or atherosclerosis. ACLA was positive at presentation & also on another occasion 6 weeks later. ACLAs were of the IgM type on both occasions. Lupus Anticoagulant & β2 GPI antibody were negative. DOPPLER of the lower limb arteries did not reveal any abnormality. Patient was successfully treated with multi-drug antileprotics & anticoagulants.
Conclusion
Infectious APLAs should be recognized as a cause of thrombosis in Leprosy. Appropriate anticoagulation can salvage limb function.
==== Body
Background
Leprosy still remains an important public health problem for many parts of the world. An association of gangrene with leprosy is a rare one & can have a number of causative mechanisms. We present a case with Leprosy & gangrene with positive anti phopholipid antibody titers.
Case presentation
A 50-year-old non-diabetic, non-hypertensive lady presented with 2 months history of progressive blackish discoloration of the toes bilaterally. Examination revealed gangrene of the Right great toe, 2nd toe & early gangrenous changes in the 3rd toe. All the peripheral arteries were well felt, there was no radiofemoral delay. There was no cardiac murmur or a carotid bruit.
She was found to have madarosis & hypopigmented, hypoaesthetic macular lesions on the upper limb & thighs. Bilateral ulnar & popliteal nerves were thickened. A skin biopsy of the lesions revealed borderline tuberculoid leprosy. Slit skin smears revealed a bacteriological index of 1+. Erythrocyte sedimentation rate was 105, lipid profile & fasting sugars were normal & anti neutrophil cytoplsmic antibody (ANCA) negative.
Anti Cardiolipin antibody (ACLA) was positive at presentation (IgG-8; IgM-28.5; ELISA Genesis Diagnostics, Cambridgeshire, UK) & also on another occasion 6 weeks later (IgG-7.5; IgM-29; ELISA Genesis Diagnostics, Cambridgeshire, UK). Thus, ACLAs were of the IgM type on both occasions. Lupus Anticoagulant (PT, aPPT, Mixing studies, DRVVT) & β2 GPI antibody were negative (IgG-1; IgM-2.5; ELISA Genesis Diagnostics, Cambridgeshire, UK). DOPPLER of the lower limb arteries did not reveal any abnormality. Tests for other hypercoagulable states (protein C, protein S, Antithrombin III, homocystein, factor V Leiden) were normal.
The patient improved with the multi drug anti leprotics & anticoagulants. By 6 weeks, there was no progression of/ fresh gangrene & the pre gangrenous changes in the 3rd toe had resolved.
Discussion
Antiphospholipid antibodies (APLA) are a group of autoantibodies, which have been reported in Antiphospholipid syndrome (APS), which is characterized by raised levels of ACLA, thrombosis, recurrent fetal loss & thrombocytopenia. APLA is a generic term that describes closely related but not identical autoantibodies found in APS: ACLA, anti β2 GPI & those with lupus anticoagulant activity. The syndrome can occur in its primary form or secondarily in association with other autoimmune disorders. Although raised levels of these antibodies were first reported only in autoimmune diseases, their prevalence is now known to be more widespread. Elevated levels of these antibodies have been found in various infections like Syphillis, HIV disease, HCV disease, tuberculosis, cytomegalovirus infection [1]. Loizou et al studied 112 leprosy patients & found elevated titers of APLA in 29%, anti β2 GPI in 89%, & anti-Prothrombin in 21% of them [2]. Initially, it seemed that infection induced APLA are not associated with the thrombotic manifestations of APS. This was attributed to the fact that the binding of autoimmune APLA to phospholipid is enhanced by the cofactor β2 GPI (i.e. β2 GPI dependent) while the binding of infection induced APLA is not enhanced by this cofactor (i.e. β2 GPI independent). Recent studies, however show that the APLA in leprosy patients are heterogeneous with respect to their β2 GPI requirement: in 10 of 31 leprosy sera, the APLA were β2 GPI dependent & 16 of 31 were β2 GPI independent [3]. The clinical implications of this β2 GPI dependency are seen in Lucio's phenomenon in which the histopathological findings are related to microvascular thrombosis in the absence of inflammatory infiltration of the vessel wall. The β2 GPI dependency of APLA in this condition has been confirmed by Levi et al [4]. Apart from this evidence of microscopic thrombosis, frank gangrene in association with leprosy is a rare entity. It has been hypothesized that certain infections in genetically predisposed individuals may induce these APLA. Phospholipid binding peptides of bacterial & viral origin that have structural similarity to the phospholipid sites have been detected & found to induce APLA with properties similar to autoimmune APL in mice [5]. The elevated levels of IgM subtype of APLA seen in our patient is in accordance with other studies of APLA in leprosy [6,7].
Gangrene of the extremities in leprosy can have mechanisms other than APLA alone. Vascular changes in the form of intimal thickening & medial infiltration are known to occur in leprosy. Embolisation & resultant grafting of the Virchow cells has been found to lead to obstruction of the vessels [8]. Four such cases of arterial obstruction have been described; 2 of them being occlusion of the posterior tibial artery by lepromatous infiltration [9]. Arteriographic abnormalities such as occlusion, narrowing, tortuosity, dilatation, poststenotic dilatation, irregularity and incomplete filling of the lumen have been found in the digital circulation in more than 75–94% of leprosy patients [10].
Nerve trunk hypertrophy secondary to lepromatous process can lead to arterial entrapment in the osteoligamentous channels. This entrapment as well as the irritation of sympathetic fibers can lead to spasm of the vessels & resultant vascular compromise to the distal extremity. This has been confirmed with angiography & reversal of the spasm as well as the vascular compromise seen after release of the roof of the osteoligamentous channel [8].
Our patient did not have any clinical or laboratory markers of atherosclerosis or embolism, DOPPLER of the lower limbs did not reveal any vascular obstruction involving the medium size arteries. In the absence of any other hypercoagulable states, APLA remains the most probable cause of the digital gangrene.
Conclusion
Infectious APLA should be recognized as a cause of thrombosis in Leprosy. Appropriate anticoagulation can salvage limb function. However, other mechanisms of gangrene need careful evaluation & appropriate management.
List of abbreviations
APLA – anti phospholipid antibody
ACLA – anti cardiolipin antibody
β2 GPI – β2 Glycoprotein I
ANCA – Anti Neutrophil Cytoplasmic Antibody
APS – Anti Phospholipid Syndrome
PT – Prothrombin Time
aPPT – Activated Partial Thromboplastin Time
DRVVT – Dilute Russel's Viper Venom Time
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SA & LB carried out the study and conceived of the study. SA drafted the manuscript & LB reviewed the same. Both authors read and approved the final manuscript.
Figure 1 Gangrenous changes.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Written consent was obtained from the patient for the clinical photograph as well as publication of study.
==== Refs
Roubey RA Immunology of the antiphospholipid antibody syndrome Arthritis Rheum 1996 39 1444 54 8814055
Loizou S Singh S Wypkema E Wypkema E Asherson RA Anticardiolipin, anti-β2-glycoprotein I and antiprothrombin antibodies in black South African patients with infectious disease Ann Rhem Dis 2003 62 1106 1111 10.1136/ard.62.11.1106
Hojnik M Gilburd B Ziporen L Blank M Tomer Y Scheinberg M Tincani A Rozman B Shoenfeld Y Anticardiolipin Antibodies in infections are heterogeneous in their dependency on β2 GPI : Analysis of anticardiolipin antibodies in leprosy Lupus 1994 3 515 521 7704010
Levy RA Pierangeli SA Espinola RG Antiphospholipid beta-2 glycoprotein 1 dependency assay to determine antibody pathogenicity Arthritis Rheum 2000 43 1476
Gharavi EE Chaimovich H Cucucrull E Celli C Tang H Wilson W Gharavi AE Induction of Antiphospholipid antibodies by immunization with synthetic bacterial & viral peptides Lupus 1999 8 449 55 10483013
de Larranaga GF Forastiero RR Martinuzzo ME Carreras LO Tsariktsian G Sturno MM Alonso BS High prevalence of antiphospholipid antibodies in leprosy: evaluation of antigen reactivity Lupus 2000 9 594 600 11035434 10.1191/096120300678828712
Panunto-Castelo A Almeida IC Rosa JC Greene LJ Roque-Barreira M The Rubino test for leprosy is a beta2-glycoprotein 1-dependent antiphospholipid reaction Immunology 2000 101 147 53 11012766 10.1046/j.1365-2567.2000.00081.x
Carayon A Dharmendra Vascular changes in leprosy Textbook of Leprosy 1985 Samant & company 853 871
Carayon A Camain R Confrontation de 18 histologiede la clinique et de l'angiographie d'une multi nevrite tuberculoide reactionnelle Bull Soc Med Afr Noire Lgue Frse XI 273
Kaur S Wahi PL Chakravarti RN Peripheral vascular deficit in leprosy Int J Lepr Other Mycobact Dis 1976 44 332 9 987995
|
16176580
|
PMC1262717
|
CC BY
|
2021-01-04 16:28:16
|
no
|
BMC Infect Dis. 2005 Sep 21; 5:74
|
utf-8
|
BMC Infect Dis
| 2,005 |
10.1186/1471-2334-5-74
|
oa_comm
|
==== Front
BMC Infect DisBMC Infectious Diseases1471-2334BioMed Central London 1471-2334-5-781619120110.1186/1471-2334-5-78Research ArticleGB virus-C – a virus without a disease: We cannot give it chronic fatigue syndrome Jones James F [email protected] Prasad S [email protected] Salvatore T [email protected] William C [email protected] Viral Exanthems and Herpesvirus Branch, Division of Viral and Rickettsial Diseases, Centers for Disease Control and Prevention, 1600 Clifton Rd, mailstop A-15, Atlanta, Georgia, 30333, USA2 Laboratory Branch, Division of HIV/AIDS Prevention, Centers for Disease Control and Prevention, 1600 Clifton Rd, mailstop G-19, Atlanta, Georgia, 30333, USA2005 28 9 2005 5 78 78 28 6 2005 28 9 2005 Copyright © 2005 Jones et al; licensee BioMed Central Ltd.2005Jones et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Chronic fatigue syndrome (CFS) is an illness in search of an infectious etiology. GB virus-C (GBV-C) virus is a flavivirus with cell tropism and host defense induction qualities compatible with a role in producing the syndrome. The GBV-C genome is detectable in 4% of the population and 12% of the population is seropositive. The present study evaluated the association between infection with GBV and CFS.
Methods
We used a commercial EIA to detect antibodies against the GBV-C E2 protein and a quantitative real-time RT-PCR assay to detect active GBV-C infection. Sera were from a case control study of CFS in Atlanta, Georgia. The Fisher's exact two-tailed test was used for statistical analysis.
Results
Two of 12 CFS patients and one of 21 controls were seropositive for prior GBV-C infection and one control had viral RNA detected, indicating active infection. The results are not statistically different.
Conclusion
We found no evidence that active or past infection with GBV is associated with CFS.
==== Body
Background
Chronic fatigue syndrome (CFS) affects 400,000 to 900,000 adults in the United States [1,2]. At least a quarter of those suffering from CFS are unemployed or receiving disability because of the illness; the average affected family forgoes $20,000 annually in lost earnings and wages (approximately half of the average United States household income); and, the annual value of lost productivity in the United States is approximately $9 billion [2-4]. Despite the public health burden imposed by CFS, diagnostic, treatment and prevention strategies have proven difficult to devise because the etiology, pathophysiology and risk factors of CFS remain unknown [reviewed in [5]].
Because the symptoms resemble those of infectious diseases and because CFS may follow an acute infectious disease, a considerable body of work has attempted to identify persistent or reactivated infection in patients with this illness [reviewed in [5]]. These efforts have included analysis of seroprevalence, evaluation of viral RNA and DNA (i.e., latent/active infection), and most recently, molecular identification procedures for unique, previously uncharacterized pathogens [6-12]. None of these reports have developed convincing evidence for a significant association between any infectious agent and CFS.
However, negative, inconclusive, and conflicting case control studies do not mean that infectious etiologies for CFS should be dismissed. A recently discovered flavivirus, GB virus-C/HGV (GBV-C) [13-17], has many properties that demanded a study assessing its possible association with CFS. GBV-C preferentially replicates in peripheral blood mononuclear cells (PBMC), primarily B and T lymphocytes, and in bone marrow in vivo [18,19]. Of note, some flaviviruses activate the classic complement pathway [20] and we have found a high prevalence of split products (C3a, C4a, and C5a) in patients with CFS compared to controls [21]. We have also examined mononuclear cell populations for gene expression patterns and have found differences between patients and control subjects [22,23]. GBV-C viremia may persist for several years following primary infection and serologically antibodies are generated against the envelope protein, E2 [reviewed in [15,18]]. The antibodies are also long-lived and may protect against re-infection [24]. GBV-C viremia and presence of antibodies are usually mutually exclusive, and only a small percentage of exposed individuals exhibit both viremia and anti-E2 antibody [25]. As yet, no illness has been associated with GBV infection.
Therefore, thinking that complement activation and altered mononuclear cell gene expression might reflect continued infection with this virus, we chose to evaluate banked serum specimens from subjects enrolled in a case control study for GBV-C RNA, an indication of persistent active infection that might be associated with CFS and for specific antibodies associated with clearance of a GBV-C infection (anti-E2) as an indicator of past infection with this virus.
Methods
This study adhered to human experimentation guidelines of the U.S. Department of Health and Human Services and the Helsinki Declaration. The CDC Institutional Review Board approved study protocols. All participants were volunteers who gave informed consent.
Study subjects
Subjects and study design have been presented in detail elsewhere [25,26]. In brief, 26 patients with CFS (23 women and 3 men) were recruited from a physician surveillance study in Atlanta [26]. Patients met the 1988 CFS research case definition [27] and had been ill for no more than 10 years. Two age-(± 5 years), race-, and sex-matched nonfatigued control subjects were contacted by means of random-digit dialing in the Atlanta area. Subjects' clinical, seroepidemiologic [6], and immunological characteristics have been reported [28]. For the current pilot study, testable sera were available from 12 cases and 21 controls.
Laboratory procedures
Detection of GBV-C infection
Plasma samples were tested for antibodies against the GBV-C E2 envelope protein by a μPLATE Anti-HGenv microtiter assay (Roche Diagnostics Corp., Indianapolis, Ind.). Furthermore, nucleic acids were extracted from 200 μL of plasma by using the QIAamp Viral RNA Mini Kit (Qiagen Inc., Valencia, Calif.), as per the manufacturer's recommendations. A real-time quantitative reverse-transcriptase polymerase chain reaction (qRT-PCR) was developed to detect GBV-C and determine the GBV-C viral load. Five microliters of extracted nucleic acids were used in the qRT-PCR assay, performed in a final volume of 50 μL using the Quantitect Probe RT-PCR kit (Qiagen). The primers used were: GBV-C 03.1-F – 5' GCACGGTCCACAGGTGTT 3' (nucleotides 226-243 of the sequence with GenBank accession number U44402[14] and GBV-C 03.2-R – 5' GTACGTGGGCGTCGTTTG 3' (nucleotides 313–330). The probe used had the sequence 5' CCGACGTCAGGCTCGTCGTTAAAC 3' (nucleotides 268–291), and was labeled with 6-carboxy-fluorescein (FAM) on the 5' end and a dark quencher on the 3' end. This combination of primers resulted in an amplicon of 105 base pairs (bp). Serial dilutions of in vitro transcribed RNA generated from a linearized plasmid encoding nucleotides 136–400 of the 5' UTR of GBV-C with GenBank accession number U44402[14] was used as a quantitative standard curve.
Statistics
Based on the estimated prevalence of GBV-C (< 4%) in healthy blood donors in the US [29], the available sample size of 12 CFS subjects and 21 controls should allow detection of 60 % positivity in viremia or positive serology at the 95% confidence level with a power of 80: with at least 50% of persons with CFS demonstrating viremia. These estimates are based on the assumption that if GBV-C infection is the primary cause CFS, the majority of CFS subjects would be positive in one of the assay systems. Results were analyzed using Fisher's exact test (2-tailed). Odds ratios were calculated from a Chi square table and confidence intervals defined using the Wald statistic.
Results and discussion
Two cases (16.7%) and 1 control (4.8%) were seropositive (2-tailed Fisher's exact test p = .54). Only one person (a control subject) had detectable GBV-C RNA. Although the odds ratio of this difference is 5.14, the confidence interval (CI 0.32–49.59) demonstrates the difference was statistically insignificant.
This exploratory study did not identify significant differences in active or cleared GBV-C infections between individuals fulfilling a clinical definition of CFS or in control subjects. These data would not exclude GBV-C infection in small percentages (<20%) of CFS patients.
Since previous studies that used seropositivity or gene expression alone have failed to demonstrate an association between specific infections and CFS, we reasoned that a combined serological and virological approach would be more appropriate since only ongoing infection with GBV-C is associated with viremia. These results are consistent with values reported in a variety of population studies regarding the prevalence of GBV-C. Previous studies have addressed Herpesviridae, enteroviruses, retroviruses, hepatitis C, HTLV-II [5] and recently parvovirus [30] and Borna virus [31], but no definitive causative links have been made. Thus, another attempt to link a majority of CFS cases with a specific infectious agent did not identify a specific association.
Conclusion
CFS remains a clinically identifiable, but daunting, medical and psychological problem.
Competing interests
The author(s) declare that they have no competing interests
Authors' contributions
JJ and SB conceived the study. JJ was responsible for specimen identification, evaluation of results and preparation of the manuscript. PK performed the antibody and PCR procedures and participated in the preparation of the manuscript. SB supervised the laboratory procedures and participated in the preparation of the manuscript. WR was responsible for the primary patient study and participated in the design and evaluation of the study, and participated in the preparation of the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Elizabeth R. Unger, MD, PhD contributed the title of the manuscript.
==== Refs
Jason LA Richman JA Rademaker AW Jordan KM Plioplys AV Taylor RR McCready W Huang CF Plioplys S A community-based study of chronic fatigue syndrome Arch Intern Med 1999 159 2129 37 10527290 10.1001/archinte.159.18.2129
Reyes M Nisenbaum R Hoaglin DC Emmons C Stewart G Randall B Stewart JA Abbey S Jones JF Gantz N Minden S Reeves WC Prevalence and incidence of chronic fatigue syndrome in Wichita, Kansas Arch Intern Med 2003 163 1530 6 12860574 10.1001/archinte.163.13.1530
Solomon L Nisenbaum R Reyes M Papanicolaou DA Unger ER Reeves WC Functional status of persons with chronic fatigue syndrome in the Wichita population Health Qual Life Outcomes 2003 1 48 14577835 10.1186/1477-7525-1-48
Reynolds KJ Vernon SD Bouchery E Reeves WC The economic impact of chronic fatigue syndrome Cost Eff Resour Alloc 2004 2 4 15210053 10.1186/1478-7547-2-4
Afari N Buchwald D Chronic fatigue syndrome: A review Am J Psychiatry 2003 160 221 236 12562565 10.1176/appi.ajp.160.2.221
Mawle AC Nisenbaum R Dobbins JG Gary HE JrStewart JA Reyes M Steele L Schmid DS Reeves WC Seroepidemiology of chronic fatigue syndrome: A case control study Clin Infect Dis 1995 21 1386 9 8749620
Koelle DM Barcy S Huang ML Ashley RL Corey L Zeh J Ashton S Buchwald D Markers of viral infection in monozygotic twins discordant for chronic fatigue syndrome Clin Infect Dis 2002 35 518 25 12173124 10.1086/341774
Buchwald D Ashley RL Pearlman T Kith P Komaroff AL Viral serologies in patients with chronic fatigue and chronic fatigue syndrome J Med Virol 1996 50 25 30 8890037 10.1002/(SICI)1096-9071(199609)50:1<25::AID-JMV6>3.0.CO;2-V
Vernon SD Shukla SK Conradt J Unger ER Reeves WC Analysis of 16S rRNA gene sequences and circulating cell-free DNA from plasma of chronic fatigue :syndrome and non-fatigued subjects BMC Microbiol 2002 2 39 12498618 10.1186/1471-2180-2-39
Reeves WC Stamey FR Black JB Mawle AC Stewart JA Pellett PE Human herpesviruses 6 and 7 in chronic fatigue syndrome: a case-control study Clin Infect Dis 2000 31 48 52 10913395 10.1086/313908
Gelman JH Unger ER Mawle AC Nisenbaum R Reeves WC Chronic fatigue syndrome is not associated with expression of endogenous retroviral p15E Molec Diagnosis 2000 5 155 156
Vernon SD Shukla S Reeves WC Absence of Mycoplasma species DNA in chronic fatigue syndrome J Med Microbiol 2003 52 1027 1028 14532349 10.1099/jmm.0.05316-0
Simons JN Leary TP Dawson GJ Pilot-Matias TJ Muerhoff AS Schlauder GG Desai SM Mushahwar IK Isolation of novel virus-like sequences associated with human hepatitis Nat Med 1995 1 564 9 7585124 10.1038/nm0695-564
Linnen J Wages J Zhang-Keck Z-Y Fry KE Krawczynski KZ Alter H Koonin E Gallagher M Alter M Hadziyannis S Karayiannis P Fung K Nakatsuji Y Shih JW Young L Piatak M JrFernandez J Chen S Zou JC Morris T Hyams KC Ismay S Lifson JD Hess G Foung SK Thomas H Bradley D Margolis H Kim JP Molecular cloning and disease association of hepatitis G virus: a transfusion-transmissible agent Science 1996 271 505 8 8560265
Radkowski M Kubicka J Kisiel E Cianciara J Nowicki M Rakela J Laskus T Detection of active hepatitis C virus and hepatitis G virus/GB virus C replication in bone marrow in human subjects Blood 2000 95 3986 9 10845938
Zuckerman AJ Alphabet of hepatitis viruses Lancet 1996 347 558 9 8596314 10.1016/S0140-6736(96)91267-2
Leary TP Muerhoff AS Simons JN Pilot-Matias TJ Erker JC Chalmers ML Schlauder GG Dawson GJ Desai SM Sequence and genomic organization of GBV-C: a novel member of the flaviviridae associated with human non-A-E hepatitis J Med Virol 1996 48 60 7 8825712 10.1002/(SICI)1096-9071(199601)48:1<60::AID-JMV10>3.0.CO;2-A
Stapleton JT Williams CF Xiang J GB Virus Type C: a beneficial infection? J Clin Microbiol 2004 42 3915 9 15364968 10.1128/JCM.42.9.3915-3919.2004
Zampino R Pickering J Iqbal M Gaud U Thomas HC Karayiannis P Hepatitis G virus/GBV-C persistence: absence of hypervariable E2 region and genetic analysis of viral quasispecies in serum and lymphocytes J Viral Hepatitis 1999 6 209 18 10.1046/j.1365-2893.1999.00162.x
Diamond MS Shrestha B Mehlhop E Sitati E Engle M Innate and adaptive immune responses determine protection against disseminated infection by West Nile encephalitis virus Viral Immunol 2003 259 78 14583143 10.1089/088282403322396082
Sorensen B Streib JE Strand M Make B Giclas PC Fleshner M Jones J Complement activation in a model of chronic fatigue syndrome J Allergy Clin Immunol 2003 112 397 403 12897748 10.1067/mai.2003.1615
Steinau M Unger ER Jones JF Rajeevan MS Differential display PCR of peripheral blood for biomarker discovery in chronic fatigue syndrome J Mol Med 2004 82 750 5 15490094 10.1007/s00109-004-0586-4
Whistler T Unger ER Nisenbaum R Vernon SD Integration of gene expression, clinical, and epidemiologic data to characterize Chronic Fatigue Syndrome J Transl Med 2003 1 10 14641939 10.1186/1479-5876-1-10
Thomas DL Vlahov D Alter HJ Hunt JC Marshall R Astemborski J Nelson KE Association of antibody to GB virus C (hepatitis G virus) with viral clearance and protection from reinfection J Infect Dis 1998 177 539 42 9498429
Gutierrez RA Dawson GJ Knigge MF Melvin SL Heynen CA Kyrk CR Young CE Carrick RJ Schlauder GG Surowy TK Dille BJ Coleman PF Thiele DL Lentino JR Pachucki C Mushahwar IK Seroprevalence of GB virus C and persistence of RNA and antibody J Med Viro 1997 53 167 73 10.1002/(SICI)1096-9071(199710)53:2<167::AID-JMV10>3.0.CO;2-G
Reyes M Gary HE JrDobbins JG Randall B Steele L Fukuda K Holmes GP Connell DG Mawle AC Schmid DS Stewart JA Schonberger LB Gunn WJ Reeves WC Surveillance for chronic fatigue syndrome–four U.S. cities, September 1989 through August 1993 MMWR CDC Surveill Summ 1997 46 1 13 12412768
Holmes GP Kaplan JE Gantz NM Komaroff AL Schonberger LB Straus SE Jones JF Dubois RE Cunningham-Rundles C Pahwa S Tosato G Zegans LS Purtilo DT Brown N Schooley RT Brus R Chronic fatigue syndrome: a working case definition Ann Intern Med 1988 108 387 9 2829679
Mawle AC Nisenbaum R Dobbins JG Gary HE JrStewart JA Reyes M Steele L Schmid DS Reeves WC Immune responses associated with chronic fatigue syndrome: a case-control study J Infect Dis 1997 175 136 41 8985207
Simons JN Desai SM Mushahwar IK The GB viruses Curr Top Microbiol Immunol 2000 242 341 75 10592668
Kerr JR Bracewell J Laing I Mattey DL Bernstein RM Bruce IN Tyrrell DA Chronic fatigue syndrome and arthralgia following parvovirus B19 infection J Rheumatol 2002 29 595 602 11911112
Li YJ Wang DX Zhang FM Liu ZD Yang AY Ykuta K Detection of antibody against Borna disease virus-p24 in the plasma of Chinese patients with chronic fatigue syndrome by Western-blot analysis Zhonghua Shi Yan He Lin Chuang Bing Du Xue Za Zhi 2003 17 330 3 15340544
|
16191201
|
PMC1262718
|
CC BY
|
2021-01-04 16:28:16
|
no
|
BMC Infect Dis. 2005 Sep 28; 5:78
|
utf-8
|
BMC Infect Dis
| 2,005 |
10.1186/1471-2334-5-78
|
oa_comm
|
==== Front
BMC Infect DisBMC Infectious Diseases1471-2334BioMed Central London 1471-2334-5-791619428010.1186/1471-2334-5-79Research ArticleLow pH immobilizes and kills human leukocytes and prevents transmission of cell-associated HIV in a mouse model Olmsted Stuart S [email protected] Kristen V [email protected] Erina M [email protected] Steven T [email protected] Owen N [email protected] Richard B [email protected] Richard A [email protected] Thomas R [email protected] Department of Biophysics, Johns Hopkins University, Jenkins Hall, 3400 N. Charles St., Baltimore, MD 21218, USA2 RAND Corporation, 201 N. Craig St #202, Pittsburgh, PA 15213, USA3 ReProtect, Inc., 703 Stags Head Rd, Baltimore, MD 21286, USA4 Department of Molecular Microbiology and Immunology, Johns Hopkins Bloomberg School of Public Health, 615 North Wolfe Street, Suite E5132, Baltimore, MD 21205 USA2005 30 9 2005 5 79 79 18 5 2005 30 9 2005 Copyright © 2005 Olmsted et al; licensee BioMed Central Ltd.2005Olmsted et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Both cell-associated and cell-free HIV virions are present in semen and cervical secretions of HIV-infected individuals. Thus, topical microbicides may need to inactivate both cell-associated and cell-free HIV to prevent sexual transmission of HIV/AIDS. To determine if the mild acidity of the healthy vagina and acid buffering microbicides would prevent transmission by HIV-infected leukocytes, we measured the effect of pH on leukocyte motility, viability and intracellular pH and tested the ability of an acidic buffering microbicide (BufferGel®) to prevent the transmission of cell-associated HIV in a HuPBL-SCID mouse model.
Methods
Human lymphocyte, monocyte, and macrophage motilities were measured as a function of time and pH using various acidifying agents. Lymphocyte and macrophage motilities were measured using video microscopy. Monocyte motility was measured using video microscopy and chemotactic chambers. Peripheral blood mononuclear cell (PBMC) viability and intracellular pH were determined as a function of time and pH using fluorescent dyes. HuPBL-SCID mice were pretreated with BufferGel, saline, or a control gel and challenged with HIV-1-infected human PBMCs.
Results
Progressive motility was completely abolished in all cell types between pH 5.5 and 6.0. Concomitantly, at and below pH 5.5, the intracellular pH of PBMCs dropped precipitously to match the extracellular medium and did not recover. After acidification with hydrochloric acid to pH 4.5 for 60 min, although completely immotile, 58% of PBMCs excluded ethidium homodimer-1 (dead-cell dye). In contrast, when acidified to this pH with BufferGel, a microbicide designed to maintain vaginal acidity in the presence of semen, only 4% excluded dye at 10 min and none excluded dye after 30 min. BufferGel significantly reduced transmission of HIV-1 in HuPBL-SCID mice (1 of 12 infected) compared to saline (12 of 12 infected) and a control gel (5 of 7 infected).
Conclusion
These results suggest that physiologic or microbicide-induced acid immobilization and killing of infected white blood cells may be effective in preventing sexual transmission of cell-associated HIV.
==== Body
Background
Most HIV transmission occurs sexually [1,2], and semen [3-6] and cervicovaginal secretions [7] contain free HIV virions as well as HIV-infected leukocytes. Whether cell-free virus or infected cells are the primary means of transmission, or whether both are important remains unknown.
Free virus transmits infection in monkey [8], chimpanzee [9], and cat [10] vaginal-challenge models, although the amount of virus used [8] has generally been greater than the amount of infectious virus found in semen of HIV-infected men [4]. Studies of cell-vectored transmission include an unsuccessful attempt to establish a model of vaginal transmission with cryopreserved SIV-infected cells in the monkey [8], but, fresh HIV-infected cells transmitted infection in the chimpanzee when applied to the cervical os [9], fresh FIV-infected cells transmitted infection after vaginal deposition in the cat [10,11], and fresh HIV-infected human peripheral blood leukocytes transmitted infection to SCID mice after vaginal deposition [12,13]. Notably, foreign lymphocytes are able to migrate through vaginal mucosa and reach the iliac lymph nodes of mice [14,15] and HIV-1 infected mononuclear cells are capable of transmigrating through a monolayer of human epithelial cells [16].
Topical microbicides are being developed to reduce the sexual transmission of HIV and other STDs. The presence of both free HIV virions and HIV-infected cells in sexual secretions, and the demonstrated ability of both to transmit infection in diverse animal models, suggests that microbicide candidates should protect against cell-associated HIV as well as cell-free HIV.
Foremost, vaginal microbicides should not injure or disrupt the normal vaginal flora or the vaginal epithelium; thus, microbicides designed to enhance and maintain natural vaginal protective mechanisms merit careful consideration. One natural protective mechanism is the mild acidity found in the healthy vagina (~pH 4) that is generated predominately by the lactic acid produced by vaginal lactobacilli [17,18] and is thought to inhibit harmful flora and some STD pathogens. Two microbicides (BufferGel®(ReProtect, Inc., Baltimore, MD) [19-21] and Acidform (TOPCAD, Chicago, IL) [22]), have been developed with the goal of strengthening and maintaining vaginal acidity, by having sufficient buffer capacity to block the alkalinizing action of semen. Another microbicide, CAP, also has acidic buffering properties, which have been postulated to contribute to its activity [23-25]. BufferGel is being tested in an HIV prevention efficacy trial (HPTN 035), and a commercially-available acidifying lubricating gel (Replens®) is being used together with a cervical barrier (diaphragm) in an HIV prevention trial [26].
In these studies, we examined the ability of mild acidity to inhibit lymphocyte, monocyte and macrophage motility. We also determined PBMC viability in mildly acidic conditions, and measured the intracellular pH (pHi) to determine their ability to defend their cytoplasmic pH. We report the effect of Carbopol®, the buffering agent contained in BufferGel, on leukocyte motility and PBMC viability. Finally, we tested the ability of BufferGel to reduce transmission of cell-associated HIV in the HuPBL-SCID mouse model.
Methods
PBMC collection
Venous blood and semen were obtained from donors according to procedures approved by the Review Board on the Use of Human Subjects at the Johns Hopkins University. PBMCs were separated on Histopaque®-1077 step gradients (Sigma, St Louis, MO). Cells were maintained at 37°C during experiments and were used within 5 hours of collection.
Monocyte chemotaxis
Gradient purified PBMCs were washed twice in 0.9% saline and resuspended in 0.9% saline at a concentration of 108 cells/ml. Monocyte chemotactic experiments were performed using microchambers with polycarbonate filters with 5 μm pores (NeuroProbe, Cabin John, MD). The top wells of the microchamber were filled with neutral or acidified RPMI (Gibco, Rockville, MD) supplemented with 1% fetal bovine serum (FBS, Gibco) and then the cell suspension was added. RPMI was acidified with one or combinations of the following acids: acetic acid (JT-Baker, Phillipsburg, NJ), HCl (JT-Baker), or 2-(N-Morpholino)ethanesulfonic acid (MES, Sigma). The bottom wells were filled with RPMI supplemented with 1% FBS at the same pH and a chemoattractant; either 0.1 μM formyl-Met-Leu-Phe (FMLP, Sigma) or 5 μM platelet activating factor (PAF, Sigma). The chamber assembly was then incubated for 90 minutes at 37°C, 5% CO2. After incubation, pH in upper and lower chambers was verified to have remained stable (± 0.1 pH units, MI-414-6 pH microelectrode, Microelectrodes Inc., Bedford, NH), the chambers were disassembled, stained with Diff-Quick® (American Scientific Products, McGaw Park, IL) and the number of monocytes that had migrated to the lower side of the filter counted.
Monocyte, macrophage and lymphocyte chemokinesis
PBMCs separated and washed as above were resuspended in 15 ml RPMI containing either 1% FBS for monocyte or lymphocyte purification or 10% human serum (HS) for macrophage maturation. Monocytes were separated from lymphocytes by reversible fibronectin adherence [27] resulting in ~90% monocyte purity as judged by morphology after Diff-Quick staining. For experiments requiring macrophages, the monocytes were additionally incubated for 3 days in RPMI containing 10% HS [28], and matured macrophages were harvested. For lymphocyte experiments, monocytes were depleted from PBMCs with CD14-coated magnetic beads (Dynal Biotech, Lake Success, NY) according to the manufacturer's protocol. The remaining cells were predominantly lymphocytes (less than 2% monocytes by Diff-Quick staining).
Leukocytes were observed by video microscopy for movement on glass microscope slides (VWR, Bridgeport, NJ) using a published method [29] modified as described below. Monocytes and macrophages were pelleted in a microcentrifuge at 400 G and resuspended in RPMI containing 0.125% Carbopol (BF Goodrich, Cleveland, OH), serum and chemoattractant (1% FBS and 0.1 μM FMLP for monocytes and 10% HS and 0.1 nM FMLP for macrophages). Lymphocytes were resuspended in RPMI containing 1% FBS, 0.1 μM FMLP, and either 50 mM MES or 0.125% Carbopol. The media had previously been adjusted to the desired pH, and the pH of the cell suspensions were verified immediately before placing them under coverslips for video microscopy. Video recordings were made for 15 minutes and then observed at high speed to assess the motility of each of the 50–200 cells contained in a low power field.
PBMC viability
PBMCs were washed in Hanks' Balanced Salt Solution without Ca+2 and Mg+2 (HBSS; Gibco BRL), resuspended in 10 ml RPMI, and 4 × 106 cells were placed in microcentrifuge tubes. Tubes were spun at 400 G for 7 minutes and all supernatant removed using drawn out capillary tips, and cells were resuspended in seminal plasma obtained by centrifugation of fresh semen.
PBMCs in seminal plasma (0.05, 0.1, or 0.2 ml) were added to glass vials containing 0.1 g of BufferGel and stirred. Vials were maintained at 37°C during incubations. Additionally, PBMCs were resuspended in seminal plasma with or without 0.2% Carbopol. These samples were observed at pH 7.4 or acidified with HCl to pH 4.5. Each cell suspension was monitored and pH adjusted throughout the timed incubations (1–60 minutes) with the pH microelectrode, and adjusted with 0.1 N NaOH or HCl as necessary to maintain the pH to within 0.03 pH units of the starting pH. After timed incubations, samples were neutralized with 2 ml RPMI with 25 mM HEPES (pH ~8.5) (Gibco BRL) and NaOH (JT-Baker) was used as needed to bring pH to 7.0–7.4. Neutralized cell suspensions were incubated with 2 μM calcein AM and 12 μM ethidium homodimer (Molecular Probes, Eugene, OR) and observed with an epifluorescent Nikon E-800 microscope and live and dead cells counted.
PBMC intracellular pH
PBMCs were labeled with 50 μM Oregon Green™ 488 carboxylic acid diacetate or 50 μM carboxyfluorescein diacetate (Molecular Probes) for 35 minutes. Fluorescently labeled platelets were removed by four 120 G washes. PBMCs were resuspended at 5 × 105 cells/ml in either HBSS or a high potassium medium containing 17.8 mM NaCl, 125.2 mM KCl, 8.7 mM Na2HPO4, 1 mM Ca2Cl, 1 mM MgCl2 (all JT-Baker), 1.5 mM KH2PO4, 5 mM D-glucose (both Sigma). Cell viability was determined with trypan blue (Sigma) exclusion. Standard curves were obtained using Nigericin (Molecular Probes), a K+/H+ ionophore to equalize the intracellular and extracellular pH of leukocytes in the high potassium medium [30-33]. Aliquots of labeled leukocytes were acidified and fluorescence ratio measurements were taken (ex: 490 nm/440 nm, with em: 520 nm for Carboxyfluorescein and 555 nm for Oregon Green) as a function of time, and ratiometric measurements were made as a function of time on an LS50B fluorometer (Perkin Elmer, Norwalk, CT).
Intracellular pH measurements were also performed using microscopic fluorescent ratiometry using the same dyes and wavelengths as in the cuvette experiments. Aliquots of PBMCs were placed onto Labtek chambered slides (Nalgene-Nunc International, Rochester, NY) coated with 0.1% polylysine (Sigma), labeled and washed. Individual cells were imaged and fluorescence measured in standard and high potassium/ionophore buffer using an Axiovert light microscope (Zeiss, Thornwood, NY) and IP Lab software (Scanalytics, Fairfax, VA).
HuPBL-SCID mouse model
The HuPBL-SCID mouse was used to model vaginal transmission of HIV-1 as previously described [13,34]. Briefly, SCID mice were administered uninfected human, peripheral blood mononuclear cells (HuPBMC) to the peritoneal cavity one week prior to inoculation, and treated with 2.5 mg Depo-Provera (Upjohn Pharmaceutical, Kalamazoo, MI), which thinned the vaginal epithelium. Vaginal inoculation of 106 HIV-1-infected HuPBMC followed vaginal administration of 10 μL of PBS, KY jelly, or BufferGel. The inoculated HIV-infected cells, of which between 1 and 5% of the cells are infected with HIV-1 as demonstrated by limiting dilution PCR, were administered on day 10 post-infection. On day 14 following vaginal inoculation, the mice were euthanized and cells from the peritoneal cavity, of both human and murine origin, were collected. HIV-1-infected cells within this population may originate from the infected-cell inoculum, or may be the human target cells from the peritoneal transplant within which the virus has replicated. These cells were placed into culture with PHA-stimulated T cells for co-culture of HIV-1, and were detected by HIV-1 p24 antigen ELISA.
Statistics
Results from leukocyte chemotaxis and chemokinesis experiments were fit with sigmoidal curves by nonlinear regression. Curves were fit to the four parameter logistics equation [35]:
y = ((d-a)/(1+(x/c)b)) + a
where a is the maximum asymptote, b is the slope of the linear region of the curve, c is the midpoint of the linear region, and d is the minimum asymptote. For monocytes and macrophages, the maximum asymptote was set to 100 and for all three cell types, the minimum asymptote was set to 0, since values above 100 and below 0 are not physiological.
Statistical analyses were performed using SPSS® statistical software version 10.0 (SPSS Inc., Chicago, IL). Results from the PBMC viability experiments were analyzed by analysis of variance for the main effects of time and pH and for their interaction. Post hoc pair-wise group comparisons were made using Schéffe's multiple comparison procedure. A linear regression model was developed to determine whether there was a significant difference in viability between PBMCs acidified with HCl and BufferGel, after adjusting for the effects of time and the interaction between time and the acidifying agent. Results for the Hu-PBL SCID mouse model were analyzed with a two sided Fisher's exact test.
Results
Monocyte chemotactic response was observed using chemotactic chambers, over the pH range of pH 5.0 to 7.5, in RPMI acidified and buffered with 50 mM MES. Monocytes attracted by FMLP crawl through a filter separating two chambers filled with media. The percent response normalized to maximal response (pH 7.0–7.5) is plotted for four experiments (Fig. 1A). Chemotaxis was essentially blocked below pH 5.8. Similar results were observed when the media were acidified with buffering systems consisting of 20 mM acetic acid, 20 mM MES and 20 mM acetic acid, and 1 N HCl without additional buffering agent and when PAF was used as the chemoattractant (data not shown).
Figure 1 Leukocyte chemotactic and chemokinetic response as a function of pH. A. Monocyte chemotactic response, as measured by monocytes migrating through filters of nucleopore chambers, is plotted as a function of pH. Monocyte chemotactic response is normalized to the maximal response for each of four repeats of the experiment (maximum response was between pH 7.0 and 7.3 for each repeat). The cells were in RPMI containing 50 mM MES. Similar results were observed when the media were acidified with buffering systems consisting of 20 mM acetic acid, 20 mM MES and 20 mM acetic acid, and 1 N HCl and when PAF was used as an alternative chemoattractant (data not shown). B. Monocyte (circles) and macrophage (squares) chemokinesis, as measured by observing cells migrating on glass slides, is plotted as a function of pH. The percent response was determined by the number of motile cells divided by the total number of cells (50–200) observed at each pH. The cells were in RPMI containing 0.125% Carbopol. C. Lymphocyte (triangles) chemokinesis is plotted as a function of pH. The percent response was determined as in B. The cells were in RPMI containing 50 mM MES (open triangles) or RPMI containing 0.125% Carbopol (closed triangles). Nonlinear regression was used to fit a sigmoidal curves for each data set: A: R = 0.95; B: R = 0.95, monocytes, 0.99 macrophages; C: R = .98 lymphocytes with Carbopol, R = .96 with MES. The minimum value was set to 0 for all curves and the maximum value to 100 for monocyte and macrophage curves. The bracket from pH 3.7 to 4.4 on the x axis of panel C indicates the pH range in a healthy, lactobacilli dominated vagina.
Since Carbopol consists of micron-sized clusters of lightly cross-linked polymer, it clogged the filters of the microchamber assembly. Therefore, monocyte, macrophage, and lymphocyte chemokinesis were observed in RPMI containing FMLP and 0.125% Carbopol or 50 mM MES by videotaping cells migrating on glass slides. In all cases, nearly all motility ceased below pH 5.8 (Fig. 1B and 1C). The chemokinesis results for monocytes, macrophages, and lymphocytes were similar to the results for monocyte chemotaxis through filters. Although monocytes and macrophages were occasionally able to change shape between pH 5 and 6, progressive motility was never observed below pH 6.0. Lymphocytes exhibited lower maximal motility rates at neutral pH (40–50%) than monocytes or macrophages (85–100%).
To determine if the loss of cell motility was due to killing of the PBMCs, cell viability was tested using a fluorescent live/dead cell assay [36]. As shown in Fig. 2, PBMCs in seminal plasma acidified with HCl to pH 4.5 maintained their viability substantially longer than PBMCs in seminal plasma acidified with BufferGel to pH 4.5 (P = 0.001). When acidified with HCl, 80% of PBMCs were still viable at 30 minutes, but when acidified to this same pH with BufferGel, less than 10% were viable at 5 minutes and none were viable at 30 minutes. To test if this enhanced killing was due to Carbopol, or other constituents or the gel structure of BufferGel, PBMCs were resuspended in seminal plasma containing 0.2% Carbopol (approximately 20-fold lower concentration than in BufferGel) and adjusted to pH 4.5 with HCl, and the viability was measured. This is also relevant for situations where the concentration of Carbopol may be low due to uneven distribution of gel in the vagina. At 30 minutes, 50% of PBMCs in pH 4.5 RPMI containing 0.2% Carbopol were viable. However, PBMCs incubated with 0.2% Carbopol at pH 7.4 had no reduction in viability over one hour.
Figure 2 Viability of PBMCs as a function of time and pH. PBMCs in seminal plasma were mixed in various volume ratios (1:2, 1:1, 2:1) with BufferGel to give final pH values of 4.0, 4.5 and 5.0. Live PBMCs (as detected calcein-AM) and dead PBMCs (as detected by ethidium homodimer) were counted as a function of time and pH. As a comparison, PBMCs in seminal plasma were mixed with a 0.2% Carbopol (approximately 1/20th the amount in BufferGel), and acidified to pH 4.5 with HCl, or maintained at pH 7.4. Also, PBMCs in seminal plasma without Carbopol were acidified to pH 4.5 with HCl. Each point represents the mean (± SD) of three to six experiments. For experiments with BufferGel, results of analysis of variance for the main effects of pH and time and their interaction were P < 0.001. Scheffe's multiple comparison procedure was used to compare pH effects. The pH 5.0 curve is significantly different from the pH 4.0 and 4.5 curves (P < 0.001). By linear regression, the pH 4.5 BufferGel curve is significantly different from the pH 4.5 HCl curve (P < 0.001).
PBMC pHi was measured as a function of extracellular pH (pHe) to investigate potential mechanisms mediating the acid immobilization and killing of PBMCs. At every pH tested, pHi equilibrated with pHe within 2 minutes (Fig. 3). At pHe as low as 6.0, however, PBMCs were able to restore their pHi to approximately 7.5 within 10 minutes. However, at pH 5.5 and below, PBMCs were unable to restore their normal pHi, even after 60 minutes.
Figure 3 The effect of extracellular pH on intracellular pH of PBMCs as a function of time. PBMCs labeled intracellularly with pH sensitive fluorescent indicators (Oregon Green or carboxyfluorescein), were incubated under varying acidic conditions and monitored over time. Upon acidification, PBMC intracellular pH undergoes a rapid decrease to or nearly to extracellular pH. At and above pHe 6.0, PBMCs recover their normal intracellular pH (7.3) after 5–10 minutes. At pHe 5.5 and below, PBMCs are unable to recover their intracellular pH. Points are mean (± SD) of three experiments.
Control experiments using fluorescent microscopy were performed measuring the fluorescence ratio of individual cells and of the surrounding buffer, to detect artifacts due to possible dye leakage out of cells and inadvertent observation of dye in the extracellular buffer. These results showed that dye leakage was not a factor and that, as with the cuvette method, pHi dropped to pHe at both pHe 5.5 and 5.0 and pHi did not recover (data not shown).
Three groups of mice were challenged with cell-associated HIV in the HuPBL-SCID mouse model. Depo-Provera treated mice were treated with BufferGel, PBS (control), or KY Jelly (gel control) prior to infected-cell challenge. One of 12 mice pretreated with BufferGel became infected with HIV, 12 of 12 mice pretreated with PBS became infected, and 5 of 7 mice pretreated with KY Jelly became infected. BufferGel provided significant protection compared to KY jelly (P < 0.01) and saline (P < 0.0001).
Discussion
During reproductive years, the pH of a healthy human vagina is usually pH 4–4.5 [37]. However, Masters and Johnson showed that an ejaculate acts as a potent alkaline buffer, which abolishes vaginal acidity within seconds and keeps the vagina neutralized (pH 6–7) for several hours after intercourse [38]. During this time acid-sensitive sperm and microbes can reach their targets or enter the upper reproductive tract. Maintaining the normal acidic condition of the vagina during and after intercourse is thus a potential method for preventing conception as well as STDs.
Here we describe the effect of low pH on the motility and viability of leukocytes that may serve as vectors for sexual transmission of HIV and present data indicating that BufferGel reduces vaginal transmission of cell-associated HIV in the HuPBL-SCID mouse model. We found that monocytes, macrophages and lymphocytes all completely lose progressive motility at a pH slightly below 6.0 (Fig. 1). The results were similar in the presence of a variety of buffering species, including Carbopol, and with a second chemoattractant, platelet activating factor. Our results are consistent with those of Fischer, et al, who showed that lymphocytes lose cytotoxic activity and other immunological functions when the extracellular pH falls to 5.8 [39,40] and with Hill and Anderson who showed that lymphocyte proliferation is abolished at pH < 6 [41].
Acidity dramatically reduced leukocyte viability in this study. When acidified with BufferGel, PBMCs added to seminal plasma were rapidly killed at pH 4.0 and 4.5, and killed more slowly at pH 5.0 (Fig. 2). Both BufferGel and Carbopol alone enhanced acid-killing of PBMCs compared to acidification with HCl alone.
In the SCID mouse model of vaginal HIV transmission, pretreatment of mice with BufferGel, as compared with saline, significantly reduced the transmission of cell-associated HIV. In addition, BufferGel provided significant protection compared to another gel, KY Jelly. We and others have shown similar protection using this model with other microbicide candidates [13,42-44].
Importantly, despite the cell killing potential of acidity especially in the presence of BufferGel, BufferGel has proven to be non-toxic to cervicovaginal epithelium in vivo in two high-dose tolerance Phase I trials [19,45] and in a highly sensitive acute vaginal toxicity mouse model [46]. We hypothesize that this advantageous differential toxicity (toxic to potentially infectious human leukocytes, but non-toxic to human epithelia in vivo) may be due to a greater ability of the epithelium to maintain the pHi of its surface cells because these cells can export protons to underlying vascularized tissue. In contrast, individual cells or pathogens surrounded by an acidic environment within the vaginal lumen must pump out protons against a high gradient over their entire surface area. Thus, although demonstrably cytotoxic to human white blood cells in vitro, and able to prevent cell-associated HIV transmission in the vagina lumen in an animal model, acidic buffering appears to have minimal or no cytotoxicity to the human genital tract epithelium.
Our results demonstrate that mild acidity inactivates the motility and viability of leukocytes that may act as vectors for sexual transmission of HIV [47,48] and reduces transmission of HIV in a mouse model. These results are consistent with other studies that have shown that the natural vaginal acidity may help reduce HIV transmission in women. Women with abundant vaginal lactobacilli have a lower HIV susceptibility compared to women with bacterial vaginosis [49,50], a condition in which vaginal lactobacilli, and therefore, vaginal acidity, are lost. It has been suggested that reduction of cervicovaginal viral load by physiological vaginal acidity may reduce both male to female, and female to male transmission of HIV [51,52]. In a study of HIV infected women, vaginal lavages with a pH < 4.5 showed a trend to contain less cultivatable HIV virus than lavages with a pH ≥ 4.5 (P = 0.08) when tested with the multinuclear-activation galactosidase indicator (MAGI) assay [52] possibly indicating a reduction in HIV load and a reduction in risk of transmission at lower pH values. Moreover, another study showed that exposure to acidic cervico-vaginal secretions reduce HIV viability as detected by cocultivation with PBMCs [51].
To gain understanding about how acid pH mediates the immobilization and killing of leukocytes we measured the pHi of PBMCs in acidified medium. Intracellular fluorescent pH probes have been used to measure the pHi of human leukocytes in a number of studies [39,53-57]. Most previous studies have concentrated on leukocyte activation and small changes in pHi (△pH ~0.5) and minor reductions in pHe. One study, exploring leukocyte function near tumors (known to have a low pHe), showed that PBMCs incubated for four hours at low pHe were unable to maintain their pHi below pHe 6.5 [39]. In our experiments with PBMCs, a significant and continued perturbation of pHi was observed at and below a pHe of 5.5, indicating that PBMCs are unable to defend their pHi below this pHe (Fig. 3). At pHe of 6.0 and above, PBMCs restored their normal pHi within 10 minutes. However, even after 60 minutes, PBMCs were unable to restore their normal pHi after exposure to a pHe of 5.5. Considering that pHi exerts profound influence on the apoptosis pathway [58], intracellular enzyme functions, protein stability, and other molecular interactions, we believe the observed prolonged perturbation of pHi is the likely cause of the observed loss of motility and viability at low pHe.
Conclusion
We found that human leukocytes lose motility and viability at pH levels typically found in a healthy vagina and that BufferGel reduces transmission of HIV in HuPBL-SCID mice. These results support the hypothesis that physiologic vaginal flora and vaginal acidity may reduce female to male HIV transmission via HIV-infected cell vectors. Our results further suggest that by helping to maintain acidity that would otherwise be abolished by semen [38,51], BufferGel, a broad spectrum microbicide/spermicide [20,21], may enhance the natural protection of vaginal acidity by killing or immobilizing infected cell vectors in semen, and may thus reduce male to female transmission.
Competing interests
Cone and Moench are developers of BufferGel and Khanna is an employee of ReProtect Inc. All three hold equity in ReProtect, Inc.
Authors' contributions
SO participated in the design of the study, conducted pHi experiments, assisted with statistical analysis, assisted in data analysis and drafted the manuscript. RC conceived of the study, assisted in data analysis, and helped draft the manuscript. KK designed and conducted the HuPBL-SCID mouse experiments, assisted in data analysis, and helped draft the manuscript. EN conducted viability and chemokinesis experiments, and assisted in data analysis. SW participated in the design of the study, conducted chemotaxis and chemokinesis experiments, and participated in analysis of the data. OJ conducted pHi experiments and assisted in data analysis. RM helped design the HuPBL-SCID mouse experiments and assisted in data analysis. TM conceived of and designed the study, supervised experiments, assisted in data analysis, and helped draft the manuscript. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The authors thank Leslie A. Meyn, M.S., Magee-Womens Research Institute, Pittsburgh, PA, for statistical analysis and Usman Bacha, Johns Hopkins University, for fluorescent microscopy on individual cells. This work was supported by National Institute of Health Training Grant GM07231-23 (SSO, STW), National Institutes of Health Program Project Grant P01-AI45967, Gustave Martin Innovative Research Award from the Department of Molecular Microbiology and Immunology, Johns Hopkins Bloomberg School of Public Health (KK) and ReProtect, Inc.
==== Refs
Auerbach DM Darrow WW Jaffe HW Curran JW Cluster of cases of the acquired immune deficiency syndrome. Patients linked by sexual contact Am J Med 1984 76 487 492 6608269 10.1016/0002-9343(84)90668-5
Chamberland ME Castro KG Haverkos HW Miller BI Thomas PA Reiss R Walker J Spira TJ Jaffe HW Curran JW Acquired immunodeficiency syndrome in the United States: an analysis of cases outside high-incidence groups Ann Intern Med 1984 101 617 623 6333197
Pilcher CD Tien HC Eron JJJ Vernazza PL Leu SY Stewart PW Goh LE Cohen MS Brief but efficient: acute HIV infection and the sexual transmission of HIV J Infect Dis 2004 189 1785 1792 15122514 10.1086/386333
Vernazza PL Gilliam BL Dyer J Fiscus SA Eron JJ Frank AC Cohen MS Quantification of HIV in semen: correlation with antiviral treatment and immune status AIDS 1997 11 987 993 9223732 10.1097/00002030-199708000-00006
Quayle AJ Xu C Mayer KH Anderson DJ T lymphocytes and macrophages, but not motile spermatozoa, are a significant source of human immunodeficiency virus in semen J Infect Dis 1997 176 960 968 9333154
Tachet A Dulioust E Salmon D De Almeida M Rivalland S Finkielsztejn L Heard I Jouannet P Sicard D Rouzioux C Detection and quantification of HIV-1 in semen: identification of a subpopulation of men at high potential risk of viral sexual transmission AIDS 1999 13 823 831 10357382 10.1097/00002030-199905070-00012
Hart CE Lennox JL Pratt-Palmore M Wright TC Schinazi RF Evans-Strickfaden T Bush TJ Schnell C Conley LJ Clancy KA Ellerbrock TV Correlation of human immunodeficiency virus type 1 RNA levels in blood and the female genital tract J Infect Dis 1999 179 871 882 10068582 10.1086/314656
Sodora DL Gettie A Miller CJ Marx PA Vaginal transmission of SIV: assessing infectivity and hormonal influences in macaques inoculated with cell-free and cell-associated viral stocks AIDS Res Hum Retroviruses 1998 14 Suppl 1 S119 23 9581895
Girard M Mahoney J Wei Q van der Ryst E Muchmore E Barre-Sinoussi F Fultz PN Genital infection of female chimpanzees with human immunodeficiency virus type 1 AIDS Res Hum Retroviruses 1998 14 1357 1367 9788677
Burkhard MJ Obert LA O'Neil LL Diehl LJ Hoover EA Mucosal transmission of cell-associated and cell-free feline immunodeficiency virus AIDS Res Hum Retroviruses 1997 13 347 355 9071435
Moench TR Whaley KJ Mandrell TD Bishop BD Witt CJ Cone RA The cat/feline immunodeficiency virus model for transmucosal transmission of AIDS: nonoxynol-9 contraceptive jelly blocks transmission by an infected cell inoculum AIDS 1993 7 797 802 8395855
Di Fabio S Giannini G Lapenta C Spada M Binelli A Germinario E Sestili P Belardelli F Proietti E Vella S Vaginal transmission of HIV-1 in hu-SCID mice: a new model for the evaluation of vaginal microbicides AIDS 2001 15 2231 2238 11698695 10.1097/00002030-200111230-00003
Khanna KV Whaley KJ Zeitlin L Moench TR Mehrazar K Cone RA Liao Z Hildreth JE Hoen TE Shultz L Markham RB Vaginal transmission of cell-associated HIV-1 in the mouse is blocked by a topical, membrane-modifying agent J Clin Invest 2002 109 205 211 11805132 10.1172/JCI200213236
Ibata B Parr EL King NJ Parr MB Migration of foreign lymphocytes from the mouse vagina into the cervicovaginal mucosa and to the iliac lymph nodes Biol Reprod 1997 56 537 543 9116158
Zacharopoulos VR Perotti ME Phillips DM A role for cell migration in the sexual transmission of HIV-1? Curr Biol 1997 7 534 537 9210380 10.1016/S0960-9822(06)00225-9
Carreno MP Chomont N Kazatchkine MD Irinopoulou T Krief C Mohamed AS Andreoletti L Matta M Belec L Binding of LFA-1 (CD11a) to intercellular adhesion molecule 3 (ICAM-3; CD50) and ICAM-2 (CD102) triggers transmigration of human immunodeficiency virus type 1-infected monocytes through mucosal epithelial cells J Virol 2002 76 32 40 11739669 10.1128/JVI.76.1.32-40.2002
Boskey ER Cone RA Whaley KJ Moench TR Origins of vaginal acidity: high D/L lactate ratio is consistent with bacteria being the primary source Hum Reprod 2001 16 1809 1813 11527880 10.1093/humrep/16.9.1809
Boskey ER Telsch KM Whaley KJ Moench TR Cone RA Acid production by vaginal flora in vitro is consistent with the rate and extent of vaginal acidification Infect Immun 1999 67 5170 5175 10496892
Mayer K Peipert J Fleming T Fullem A Moench T Cu-Uvin S Bentley M Chesney M Rosenberg Z Safety and tolerability of buffergel, a novel vaginal microbicide, in women in the united states Clin Infect Dis 2001 32 476 482 11170957 10.1086/318496
Olmsted SS Dubin NH Cone RA Moench TR The rate at which human sperm are immobilized and killed by mild acidity Fertil Steril 2000 73 687 693 10731526 10.1016/S0015-0282(99)00640-8
Zeitlin L Hoen TE Achilles SL Hegarty TA Jerse AE Kreider JW Olmsted SS Whaley KJ Cone RA Moench TR Tests of Buffergel for contraception and prevention of sexually transmitted diseases in animal models Sex Transm Dis 2001 28 417 423 11460027
Amaral E Faundes A Zaneveld L Waller D Garg S Study of the vaginal tolerance to Acidform, an acid-buffering, bioadhesive gel Contraception 1999 60 361 366 10715372 10.1016/S0010-7824(99)00102-X
Neurath AR Li YY Mandeville R Richard L In vitro activity of a cellulose acetate phthalate topical cream against organisms associated with bacterial vaginosis J Antimicrob Chemother 2000 45 713 714 10797102 10.1093/jac/45.5.713
Neurath AR Strick N Li YY Water dispersible microbicidal cellulose acetate phthalate film BMC Infect Dis 2003 3 27 14617380 10.1186/1471-2334-3-27
Neurath AR Strick N Li YY Lin K Jiang S Design of a "microbicide" for prevention of sexually transmitted diseases using "inactive" pharmaceutical excipients Biologicals 1999 27 11 21 10441398 10.1006/biol.1998.0169
Meehan A Chidanyika A Naidoo S Didier L Ramjee G Chipato T Dinat N Blanchard K team MIRA Padian N Prevalence and Risk Factors for HIV Infection in Zimbabwean and South African Women: Preliminary analysis of the MIRA trial screening population.: ; The XV International AIDS Conference; Bangkok, Thailand. 2004
Jones BM Nicholson JK Holman RC Hubbard M Comparison of monocyte separation methods using flow cytometric analysis J Immunol Methods 1989 125 41 47 2575116 10.1016/0022-1759(89)90076-8
Dougherty GJ McBride WH Zembala M and Asherson GL Monocyte Differentiation in vitro Human Monocytes 1989 San Diego, Academic Press 49 58
Huu TP Chemotaxis of leucocytes: an improved method of direct microscopic observation Ann Immunol (Paris) 1977 128C 763 770 20033
Bright GR Fisher GW Rogowska J Taylor DL Fluorescence ratio imaging microscopy: temporal and spatial measurements of cytoplasmic pH J Cell Biol 1987 104 1019 1033 3558476 10.1083/jcb.104.4.1019
Chaillet JR Boron WF Intracellular calibration of a pH-sensitive dye in isolated, perfused salamander proximal tubules J Gen Physiol 1985 86 765 794 4078557 10.1085/jgp.86.6.765
Tanasugarn L McNeil P Reynolds GT Taylor DL Microspectrofluorometry by digital image processing: measurement of cytoplasmic pH J Cell Biol 1984 98 717 724 6198329 10.1083/jcb.98.2.717
Thomas JA Buchsbaum RN Zimniak A Racker E Intracellular pH measurements in Ehrlich ascites tumor cells utilizing spectroscopic probes generated in situ Biochemistry 1979 18 2210 2218 36128 10.1021/bi00578a012
Khanna KV Whaley KJ Zeitlin L Ford DH Markham RB Vaginal transmission of HIV-1 in Hu-PBL-SCID mice.: ; Poster 11215 at Twelth World AIDS Conference; Geneva, Switzerland. 1998
Rodbard D Hutt DM Statistical analysis of radioimmunoassays and immunoradiometric (labeled antibody) assays: ; Istanbul. 1973 1 International Atomic Energy Agency 165 192
Burghardt RC Barhoumi R Doolittle DJ Philips TD Hayes AW Application of Cellular Fluorescence Imaging for In Vitro Toxicology Testing Principles and Methods of Toxicology 1994 3rd New York, Raven Press 1231 1258
Larsen B Vaginal flora in health and disease. Clin Obstet Gynecol 1993 36 107 121 8435935
Masters WH Johnson VE The physiology of the vaginal reproductive function Western J Surg 1961 69 105 120 13767982
Fischer B Muller B Fischer KG Baur N Kreutz W Acidic pH inhibits non-MHC-restricted killer cell functions Clin Immunol 2000 96 252 263 10964544 10.1006/clim.2000.4904
Fischer B Muller B Fisch P Kreutz W An acidic microenvironment inhibits antitumoral non-major histocompatibility complex-restricted cytotoxicity: implications for cancer immunotherapy J Immunother 2000 23 196 207 10746546 10.1097/00002371-200003000-00004
Hill JA Anderson DJ Human vaginal leukocytes and the effects of vaginal fluid on lymphocyte and macrophage defense functions Am J Obstet Gynecol 1992 166 720 726 1536258
Di Fabio S Giannini G Spada M Binelli A Germinario E Lapenta C Carlini F Belardelli F Vella S Efficacy of PRO 2000/5 Gel in a hu-SCID Mouse Model for Vaginal Transmission of Cell-associated HIV: ; Microbicides 2004; London. 2004
Di Fabio S Van Roey J Giannini G van den Mooter G Spada M Binelli A Pirillo MF Germinario E Belardelli F de Bethune MP Vella S Inhibition of vaginal transmission of HIV-1 in hu-SCID mice by the non-nucleoside reverse transcriptase inhibitor TMC120 in a gel formulation Aids 2003 17 1597 1604 12853741 10.1097/00002030-200307250-00003
D'Cruz OJ Samuel P Uckun FM PHI-443: a novel noncontraceptive broad-spectrum anti-human immunodeficiency virus microbicide Biol Reprod 2004 71 2037 2047 15306558 10.1095/biolreprod.104.032870
van De Wijgert J Fullem A Kelly C Mehendale S Rugpao S Kumwenda N Chirenje Z Joshi S Taha T Padian N Bollinger R Nelson K Phase 1 trial of the topical microbicide BufferGel: safety results from four international sites J Acquir Immune Defic Syndr 2001 26 21 27 11176265
Achilles SL Shete PB Whaley KJ Moench TR Cone RA Microbicide Efficacy and Toxicity Tests in a Mouse Model for Vaginal Transmission of Chlamydia trachomatis Sex Transm Dis 2002 29 655 664 12438901
Levy JA The transmission of AIDS: the case of the infected cell Jama 1988 259 3037 3038 3285042 10.1001/jama.259.20.3037
Anderson DJ Yunis EJ "Trojan Horse" leukocytes in AIDS N Engl J Med 1983 309 984 985 6621629
Taha TE Hoover DR Dallabetta GA Kumwenda NI Mtimavalye LA Yang LP Liomba GN Broadhead RL Chiphangwi JD Miotti PG Bacterial vaginosis and disturbances of vaginal flora: association with increased acquisition of HIV AIDS 1998 12 1699 1706 9764791 10.1097/00002030-199813000-00019
Martin HL Richardson BA Nyange PM Lavreys L Hillier SL Chohan B Mandaliya K Ndinya-Achola JO Bwayo J Kreiss J Vaginal lactobacilli, microbial flora, and risk of human immunodeficiency virus type 1 and sexually transmitted disease acquisition J Infect Dis 1999 180 1863 1868 10558942 10.1086/315127
Tevi-Benissan C Belec L Levy M Schneider-Fauveau V Si Mohamed A Hallouin MC Matta M Gresenguet G In vivo semen-associated pH neutralization of cervicovaginal secretions Clin Diagn Lab Immunol 1997 4 367 374 9144379
Cummins JEJ Villanueva JM Evans-Strickfaden T Sesay SM Abner SR Bush TJ Green TA Lennox JL Wright T Folks TM Hart CE Dezzutti CS Detection of infectious human immunodeficiency virus type 1 in female genital secretions by a short-term culture method J Clin Microbiol 2003 41 4081 4088 12958229 10.1128/JCM.41.9.4081-4088.2003
Grinstein S Cohen S Lederman HM Gelfand EW The intracellular pH of quiescent and proliferating human and rat thymic lymphocytes J Cell Physiol 1984 121 87 95 6332816 10.1002/jcp.1041210112
Grinstein S Cohen S Rothstein A Cytoplasmic pH regulation in thymic lymphocytes by an amiloride-sensitive Na+/H+ antiport J Gen Physiol 1984 83 341 369 6325586 10.1085/jgp.83.3.341
Swallow CJ Grinstein S Rotstein OD Cytoplasmic pH regulation in macrophages by an ATP-dependent and N,N'-dicyclohexylcarbodiimide-sensitive mechanism. Possible involvement of a plasma membrane proton pump J Biol Chem 1988 263 19558 19563 2904439
Swallow CJ Grinstein S Rotstein OD Regulation of cytoplasmic pH in resident and activated peritoneal macrophages Biochim Biophys Acta 1990 1022 203 210 2155021
Swallow CJ Grinstein S Rotstein OD Grinstein S and Rotstein OD Regulation and Functional Significance of Cytoplasmic pH in Phagocytic Leukocytes Mechanisms of leukocyte activation 1990 San Diego, Academic Press 227 247
Meisenholder GW Martin SJ Green DR Nordberg J Babior BM Gottlieb RA Events in apoptosis. Acidification is downstream of protease activation and BCL-2 protection J Biol Chem 1996 271 16260 16262 8663307 10.1074/jbc.271.27.16260
|
16194280
|
PMC1262719
|
CC BY
|
2021-01-04 16:28:14
|
no
|
BMC Infect Dis. 2005 Sep 30; 5:79
|
utf-8
|
BMC Infect Dis
| 2,005 |
10.1186/1471-2334-5-79
|
oa_comm
|
==== Front
BMC Infect DisBMC Infectious Diseases1471-2334BioMed Central London 1471-2334-5-811620215610.1186/1471-2334-5-81Research ArticleIdentification of moaA3 gene in patient isolates of Mycobacterium tuberculosis in Kerala, which is absent in M. tuberculosis H37Rv and H37Ra Sarojini Suma [email protected] Smitha [email protected] Indulakshmi [email protected] Sathish [email protected] Mycobacterial Research Group, Rajiv Gandhi Centre for Biotechnology, Thiruvananthapuram, Kerala, India2005 4 10 2005 5 81 81 30 3 2005 4 10 2005 Copyright © 2005 Sarojini et al; licensee BioMed Central Ltd.2005Sarojini et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Tuberculosis is endemic to developing countries like India. Though the whole genome sequences of the type strain M. tuberculosis H37Rv and the clinical strain M. tuberculosis CDC1551 are available, the clinical isolates from India have not been studied extensively at the genome level. This study was carried out in order to have a better understanding of isolates from Kerala, a state in southern India.
Results
A PCR based strategy was followed making use of the deletion region primers to understand the genome level differences between the type strain H37Rv and the clinical isolates of M. tuberculosis from Kerala. PCR analysis of patient isolates using RD1 region primers revealed the amplification of a 386 bp region, in addition to the expected 652 bp amplicon. Southern hybridization of genomic DNA with the 386 bp amplicon confirmed the presence of this new region in a majority of the patient isolates from Kerala. Sequence comparison of this amplicon showed close homology with the moaA3 gene of M. bovis. In M. bovis this gene is present in the RvD5 region, an IS6110 mediated deletion that is absent in M. tuberculosis H37Rv.
Conclusion
This study demonstrates the presence of moaA3 gene, that is absent in M. tuberculosis H37Rv and H37Ra, in a large number of local isolates. Whether the moaA3 gene provides any specific advantage to the field isolates of the pathogen is unclear. Field strains from Kerala have fewer IS6110 sequences and therefore are likely to have fewer IS6110 dependent rearrangements. But as deletions and insertions account for much of the genomic diversity of M. tuberculosis, the mechanisms of formation of sequence polymorphisms in the local isolates should be further examined. These results suggest that studies should focus on strains from endemic areas to understand the complexities of this pathogen.
==== Body
Background
Tuberculosis remains one of the most life threatening diseases and has been declared as a global emergency in 1993 by the World Health Organization (WHO) [1]. Failure in adhering to the strict drug regimen has led to the emergence of multi drug resistant isolates of this pathogen. This, combined with the problem of HIV, worsens the TB menace in developing countries. However, the main reason for the failure to eradicate tuberculosis lies in the biological properties of the infecting organism and its ability to persist in a latent state in the macrophages.
Bacterial strains within a single species exhibit variations in their properties such as pathogenicity, host specificity, virulence, adaptation to particular habitats and drug resistance. The passaging of M. tuberculosis H37Rv and M. bovis BCG for several decades outside the human host have induced changes in the genome of the pathogen and have also altered their virulence characteristics. Whole genome sequences of the type strain M. tuberculosis H37Rv [2], the clinical strain M. tuberculosis CDC1551 [3] and M. bovis [4] are already available. Many researchers have used a number of comparative analysis techniques like subtractive hybridization and microarray to identify differences in the genomes of laboratory strains and vaccine strains. Three genomic regions which are absent in M. bovis BCG but present in M. bovis and M. tuberculosis were first described using subtractive hybridization [5]. DNA microarray based studies between H37Rv and BCG have shown that 16 RDs (Regions of Differences) are deleted in BCG [6]. Similarly, whole genome comparison studies have shown six deletion regions in M. tuberculosis H37Rv – RvD1 to RvD5 and TbD1 [7].
Our earlier studies on the clinical isolates from Kerala on IS6110 polymorphism have shown that a large number of isolates have few or no copy of the sequence [8]. This made about fifty percent of strains untypable using IS6110. These results have been corroborated by studies from other endemic areas in India as well as outside [9-12]. This has prompted us to speculate on whether there are major differences in the genome in the field strains from endemic areas. Therefore, we examined these strains to see the distribution of the different RD regions. Here we report the presence of a genomic region in the clinical isolates of M. tuberculosis, which is absent in the type strains H37Rv and H37Ra.
Methods
Mycobacterial strains and DNA isolation
The type strains used for the study included Mycobacterium tuberculosis H37Rv, H37Ra and M. bovis BCG. These were grown in Middlebrook 7H9 Broth (Difco Laboratories) supplemented with OADC enrichment (Difco Laboratories) and 0.05% glycerol (USB Corporation). Field strains of Mycobacterium tuberculosis were those isolated from sputum samples of tuberculosis patients from different parts of Kerala. The strains were biochemically tested for Catalase, Niacin and Nitrate for identification. They were characterised by IS6110 fingerprinting. Drug resistance pattern was also studied using the four major frontline drugs viz, isoniazid, rifampicin, ethambutol and streptomycin.
DNA was isolated from cells pelleted from liquid culture using glass beads in a minibead beater. The DNA was precipitated after phenol:chloroform extraction using 3M Sodium acetate (pH 5.2) and 100% ethanol and dissolved in TE buffer, pH 8.0.
PCR amplification of RD1 region
Oligonucleotide primer pairs used for the study were RD1DLa: 5'-AGA TGA AGA CCG ATG CCG CTA C -3' and RD1DRa: 5'-CCC GTG TTT CGC TAT TCT ACG C-3'. PCR was performed in a final volume of 30 μl using 1.25 units of Taq DNA Polymerase (Promega Corporation) for each reaction. After initial denaturation, amplification was done using a PCR thermal cycler (BioRad) for 35 cycles of 94°C/40 sec, 64°C/1 min, 72°C/1 min followed by a final extension of 72°C/7 min. To identify the flanking sequences of the 386 bp region, another set of primers were used (moaFP: 5'-CCCATCGTGGTCGTTCACC-3' and moaRP: 5'-CGATGGCAGCGGTTTACAG-3') which was expected to amplify a 1254 bp product.
Southern hybridization
Genomic DNA from M. tuberculosis H37Rv, H37Ra, M. bovis BCG and the clinical isolates digested using EcoR I (New England Biolabs) was separated on agarose gels, transferred to nylon membrane (Hybond) and probed with α32P [dCTP] labelled PCR product. After overnight hybridization at 65°C, the blot was washed with increasing stringency of SSC-SDS and exposed to an activated Phosphor screen (Kodak). The screen was then scanned using Personal Molecular Imager FX (BioRad) and the picture was visualized using the software Quantity One (BioRad).
Cloning and sequencing
PCR products separated on agarose gel were eluted using GFX ™ PCR DNA and Gel Band Purification kit (Amersham Pharmacia Biotech Inc). The eluted DNA was cloned into pGEMT Easy vector (Promega Corporation). Plasmid DNA for sequencing was purified using Nucleospin Plasmid kit (Macherey-Nagel) in accordance with manufacturers' instructions. Plasmids were digested with EcoR I to check for the presence of inserts. DNA sequencing by cycle sequencing method with the fluorescent dye terminator (Big Dye Terminator Cycle Sequencing Ready Reaction Kit, (PE Biosystems)) was carried out with T7 and SP6 promoter primers in an automated sequencer (ABI Prism 310).
Results
Screening of RD1 by PCR
RD1, the most significant region of difference between M. tuberculosis and M. bovis BCG is a 9505 bp long region absent in all the different BCG substrains. PCR primers were designed to amplify regions within RD1 to find out polymorphism between type strains and the clinical isolates. PCR using RD1DLa and RD1DRa primers was expected to amplify a 652 bp fragment (comprising of Rv3874 and Rv3875, coding for cfp10 and esat 6) in M. tuberculosis H37Rv and the clinical isolates. In H37Rv and H37Ra the expected 652 bp band was observed. In BCG the 652 bp band was absent as expected, but a 386 bp fragment was amplified. The clinical isolates showed both 652 and 386 bp fragments. A set of twenty patient isolates from Kerala was used for the initial screening. Of these, only one isolate (RGTB43), did not have the 386 bp amplicon. (Fig. 1). Later we screened a total of one hundred isolates from Kerala by PCR and all except three showed the presence of the 386bp amplicon (Data not shown). A second PCR using primes designed from the surrounding regions of moaA3 gene was done to confirm the presence of the full ORF in clinical isolates. The expected amplicon of 1254 bp was obtained in all those clinical isolates which had the 386 bp fragment (Fig 2).
Figure 1 PCR of clinical isolates of M. tuberculosis for RD1 region. Lane 1: 100 bp marker, lane 2: H37Rv, lane 3: H37Ra, lane 4: M. bovis BCG, lanes 5–24: Clinical isolates RGTB 29, 37, 40, 43, 55, 60, 70, 86, 87, 93, 95, 109, 110, 123, 142, 144, 154, 167, 177, 193 respectively.
Figure 2 PCR of clinical isolates of M. tuberculosis for moaA3 gene. Lane 1: Marker- λ DNA double digest (EcoR I/Hind III), lane 2: H37Rv, lane 3: H37Ra, lane 4: M. bovis BCG, lanes 5–24: Clinical isolates RGTB 29, 37, 40, 43, 55, 60, 70, 86, 87, 93, 95, 109, 110, 123, 142, 144, 154, 167, 177, 193 respectively, lane 25: Negative Control.
Southern blot
To confirm the results of the PCR, EcoR I digested genomic DNA from M. bovis BCG, M. tuberculosis H37Rv and H37Ra and pooled DNA from all the 20 clinical isolates (called local pool) were subjected to Southern hybridization using radioactively labelled 386 bp fragment from M. bovis BCG. Local pool and M. bovis BCG showed a signal corresponding to about 1.0 kb (Fig. 3A) while H37Rv and H37Ra were negative. Southern hybridization of each of the individual clinical isolate was then carried out for confirming the result and all, except RGTB 43, showed a positive signal (Fig. 3B). The strain RGTB 43 was negative by PCR as well.
Figure 3 Southern hybridization of EcoR I digested genomic DNA probed with radiolabelled 386 bp PCR product. Panel A- Lane1: Marker – λ DNA double digest (EcoR I/Hind III) lane2: H37Rv, lane 3: H37Ra, lane 4: M. bovis BCG, lane 5: pool of DNA from local isolates. Panel B – Southern hybridization to DNA from individual isolates. Lane 1: Marker – λ DNA double digest (EcoR I/Hind III), lane 2: M. bovis BCG, Lanes 3–22: Isolates RGTB 70, 86, 87, 93, 95, 109, 110, 123, 142, 144, 154, 167, 177, 193, 29, 37, 40, 43, 55, 60 respectively.
DNA sequence homology
Sequencing of the 386 bp amplicon cloned into pGEMT Easy vector was carried out using T7 and SP6 promoter primers. The sequence data obtained was compared to the whole genome of M. bovis [19]M. tuberculosis [20] and 100% sequence homology was obtained with M. bovis whereas M. tuberculosis H37Rv showed only 61%. The upstream and downstream sequences of this 386 bp region were identified by searching the M. bovis genome database. It was found that the sequenced fragment did not belong to the RD1 region. Instead, it was found to be part of RvD5, a deletion in the type strain H37Rv (Fig. 4). This region corresponds to moaA3 gene in M. bovis which codes for molybdenum cofactor biosynthesis protein A, MoaA1.
Figure 4 A diagrammatic representation of moaA3 and the surrounding regions. Comparison of the region comprising moaA3 gene and the surrounding genes in M. bovis, H37Rv and the clinical isolate, CDC1551. All data from references [19, 20, 21, 22]. The coamplified PCR product is shown as a thick line in the moaA3 locus. The amplicon extends from 575 to 960 of moaA3 gene in M. bovis. The genome coordinates of the moaA3 gene in M. bovis and M. tuberculosis CDC1551 are shown. The thin dotted lines indicate corresponding genes in M. bovis and CDC1551. The bold dotted lines indicate similar genes in CDC1551 and H37Rv. Rv3324A is differentially shown since it's a truncated gene and has only partial nucleotide similarity to CDC1551 moaB3 gene.
The moaA3 gene was present in 97 of the 100 clinical isolates tested (details not presented). These isolates had varying IS6110 and drug resistance profiles suggesting the possible absence of a relationship between the moaA3 fragment, IS6110 copy number and drug resistance profile.
Discussion
Studies using subtractive hybridization [5] and microarrays [6] have identified 16 regions, (ranging in size from 2–12.7 kb), in M. tuberculosis H37Rv which are absent in M. bovis BCG. Deletions are also reported in H37Rv – RvD1 to RvD5 and TbD1 [7]. These results suggest that generation of deletions may be a major mechanism for creating genetic diversity among the members of the complex. On this basis, we sought to screen the clinical isolates of M. tuberculosis from Kerala for differences in the RD regions. Initially, we used primers spanning RD1 region, since RD1 is the most important region of difference and is deleted in all the substrains of M. bovis BCG [6]. The loss of RD1 is one major genetic event that contributes to the attenuation of BCG, and its reintroduction into an attenuated strain resulted in a significant increase in virulence [13].
Of the nine open reading frames predicted within the 9.5 kb RD1 region, ORFs coding for cfp10 and esat6 are considered to be very important as there is vigorous host response to these proteins. Amplification using primers that span this region was expected to give a 652 bp PCR product. But the PCR results revealed an extra amplicon of 386 bp in the local isolates and BCG. Further characterization by sequencing and homology search indicated that this region is a part of the moaA3 gene which codes for molybdopterin cofactor protein A in M. bovis. The sequence of the 386 bp amplicon obtained from the local strains showed 100% homology with M. bovis as compared to 61% with M. tuberculosis H37Rv. The PCR primers that we made spanning the RD1 region was similar to portions of the moaA3 gene in the RvD5 region, which resulted in the amplification of the 386 bp fragment. This amplicon spanned the nucleotides 575 to 960 of the moaA3 gene in M. bovis (Fig 4). The moaA3 gene is absent in H37Rv, but another gene in the biosynthetic pathway, moaC3, was the closest to the 386 bp amplicon, with a homology of 61%. Database searches revealed that the moaA3 gene is present in the CDC1551 in the RvD5 region as well. To confirm the location of the moaA3 gene in our isolates, a second PCR designed to amplify the flanking sequences of moaA3 gene was performed. The results confirmed the location of the moaA3 gene in the clinical isolates from Kerala. In M. bovis (Mb3355) the gene is 1065 bp long while in CDC 1551, the moaA3 gene (MT3427) is 1189 bp long, due to an additional 123 bp in the C terminal region. In the overlapping region, CDC1551 has 100% homology with M. bovis moaA3. Genome comparison studies have shown that moaA3 is one among the few genes that is present in CDC1551 and absent in H37Rv [3]. Southern hybridization studies done in our lab confirmed that moaA3 gene is absent in the type strains M. tuberculosis H37Rv and H37Ra and is present in most of the clinical isolates in Kerala as well as in M. bovis BCG. Since moaA3 gene has been seen in the RvD5 region in both M bovis and in CDC1551, we have presumed that these genes are in the same region in our local strains as well, but these results need confirmation. The regions surrounding the moaA3 gene and the IS6110 elements flanking the RvD5 region in these local isolates merit further investigation.
Molybdopterin is a cofactor required for nitrate reductase and many other enzymes involved in anaerobic metabolism. Genes involved in the molybdopterin cofactor biosynthesis pathway are present in almost all organisms. M. tuberculosis H37Rv dedicates 21 genes to the biosynthesis of this cofactor [2]. But there is no gene homologous to the moaA3 found in M. bovis. This cofactor is thought to be involved in the biosynthesis of molybdopterin precurser Z from guanosine in M. bovis. Studies in Escherichia coli have suggested that these molybdoenzymes have the ability to hydroxylate or dehydroxylate certain compounds enabling the bacteria to detoxify them [14]. In addition, E. coli with defective moa show a decrease in the frequency of adaptive mutations [15]. Thus, one may infer that the moaA3 gene might have a role in the intracellular survival of the local M. tuberculosis strains or may provide some selective advantage to them. A recent study using a promoter trap vector has identified two of the genes, moaX and moeB1 as upregulated in mouse lungs upon infection [16]. A systematic study is required to understand the exact role of this protein in the lifecycle of this pathogen. At the same time, the effect due to the lack of moaA3 on M. tuberculosis H37Rv may be difficult to quantify as the remaining array of moa genes could be expected to complement any lost activity.
The RvD5 region from which the amplicon was generated is an IS6110 mediated deletion in the type strain H37Rv [7]. IS6110, a powerful genetic marker for strain differentiation [17] has also been shown to play an important role in mediating genomic rearrangements and deletions in mycobacteria. In fact, four of the five genomic deletions in M. tuberculosis H37Rv (except RvD1) are predicted to be IS6110 mediated recombinations [7]. But IS6110 mediated alterations may not provide much selective advantage to the bacteria from endemic areas such as Kerala, as a large percentage of the isolates have very few copies of IS6110 [8]. Insertions and deletions are important in the evolution of a bacterial species. M. tuberculosis, considered an evolutionarily "young" pathogen, would not be expected to have undergone extensive variations in its genome [18]. But, in spite of this, differences could be detected between the laboratory strains and clinical isolates, both by sequence analysis as in the case of CDC1551 [3] and by PCR as in this study. In a scenario of few/no copies of IS6110 other insertion sequences or mobile genetic elements could be involved in these variations. Therefore, a detailed study of the genome of field strains from different endemic regions would provide more insights into the diversity of this pathogen.
Conclusion
This study, demonstrates the presence of the moaA3 gene in a large number of local isolates. This gene has been shown to be present in M. bovis, but not in H37Rv or H37Ra. The results obtained suggest that the population of strains in endemic areas is different from type strains, as suggested earlier by our analysis of IS6110. The field strains may also vary between different endemic regions. So the strains from endemic areas need to be examined in greater detail to understand the complexities of this pathogen. Such an understanding is essential for us to be able to plan adequate control measures for tackling what is appearing to be the world's number one killer.
Abbreviations
RD: Region of Difference
RGTB: Rajiv Gandhi Centre for Biotechnology Tuberculosis isolates
PCR: Polymerase Chain Reaction
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SS carried out most of the experiments, data analysis and wrote the manuscript. SS# did part of the experimental work. IR did the IS6110 work and contributed to the writing of the manuscript. SM conceived and co-designed the study, provided inputs for writing and supervised the study. All authors read and approved the final manuscript.
Table 1 Details of clinical isolates of M. tuberculosis. The resistance (R)/sensitivity(S) profile of the isolates to the four frontline anti- tuberculosis drugs and their IS6110 copy number are shown below.
RGTB No: Drug resistance profile IS6110 copy no:
Isoniazid Ethambutol Rifampicin Streptomycin
29 R R S S 1
37 S S S S 1
40 S R S S 0
43 S R S S 9
55 R R S S 1
60 R R S R 2
70 S S S S 1
86 R R R S 1
87 R R S S 3
93 R S S S 1
95 S S S S 6
109 S S S S 12
110 R S S S 8
123 S S S S 0
142 R R S R 10
144 S S S S 1
154 R S R S 1
167 S R S S 14
177 S R S S 1
193 R R R S 1
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Suma Sarojini and Smitha Soman are recipients of Senior Research Fellowship and Junior Research Fellowship respectively from Council for Scientific and Industrial Research (CSIR), Govt. of India. This study received financial assistance under program support from the Department of Biotechnology, Government of India. Rajiv Gandhi Centre for Biotechnology is under the Kerala State Council for Science, Technology and Environment. We thank Laiza K. Paul for excellent technical assistance.
==== Refs
TB a global emergency-WHO report on the TB epidemic 1994 Geneva , World Health Organization
Cole ST Brosch R Parkhill J Garnier T Churcher C Harris D Gordon SV Eiglmeier K Gas S Barry CE Tekaia F Badcock K Basham D Brown D Chillingworth T Connor R Davies R Devlin K Feltwell T Gentles S Hamlin N Holroyd S Hornsby T Jagels K Krogh A McLean J Moule S Murphy L Oliver K Osborne J Quail MA Rajandream MA Rogers J Rutter S Seeger K J S Squares R Squares S Sulston JE Taylor K Whitehead S Barrell BG Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence Nature 1998 393 537 544 9634230 10.1038/31159
Fleischmann RD Alland D Eisen JA Carpenter L White O Peterson J DeBoy R Dodson R Gwinn M Haft D Hickey E Kolonay JF Nelson WC Umayam LA Ermolaeva M Salzberg SL Delcher A Utterback T Weidman J Khouri H Gill J Mikula A Bishai W Jacobs Jr WRJ Venter JC Fraser CM Whole-genome comparison of Mycobacterium tuberculosis clinical and laboratory strains J Bacteriol 2002 184 5479 5490 12218036 10.1128/JB.184.19.5479-5490.2002
Garnier T Eiglmeier K Camus JC Medina N Mansoor H Pryor M Duthoy S Grondin S Lacroix C Monsempe C Simon S Harris B Atkin R Doggett J Mayes R Keating L Wheeler PR Parkhill J Barrell BG Cole ST Gordon SV Hewinson RG The complete genome sequence of Mycobacterium bovis Proc Natl Acad Sci U S A 2003 100 7877 7882 12788972 10.1073/pnas.1130426100
Mahairas GG Sabo PJ Hickey MJ Singh DC Stover CK Molecular analysis of genetic differences between Mycobacterium bovis BCG and virulent M. bovis J Bacteriol 1996 178 1274 1282 8631702
Behr MA Wilson MA Gill WP Salamon H Schoolnik GK Rane S Small PM Comparative genomics of BCG vaccines by whole-genome DNA microarray Science 1999 284 1520 1523 10348738 10.1126/science.284.5419.1520
Brosch R Philipp WJ Stavropoulos E Colston MJ Cole ST Gordon SV Genomic analysis reveals variation between Mycobacterium tuberculosis H37Rv and the attenuated M. tuberculosis H37Ra strain Infect Immun 1999 67 5768 5774 10531227
Radhakrishnan I K MY Kumar RA Mundayoor S Implications of low frequency of IS6110 in fingerprinting field isolates of Mycobacterium tuberculosis from Kerala, India J Clin Microbiol 2001 39 1683 11354035 10.1128/JCM.39.4.1683.2001
Siddiqi N Shamim M Amin A Chauhan DS Das R Srivastava K Singh D Sharma VD Katoch VM Sharma SK Hanief M Hasnain SE Typing of drug resistant isolates of Mycobacterium tuberculosis from India using the IS6110 element reveals substantive polymorphism Infect Genet Evol 2001 1 109 116 12798025 10.1016/S1567-1348(01)00014-4
Mathema B Bifani PJ Driscoll J Steinlein L Kurepina N Moghazeh SL Shashkina E Marras SA Campbell S Mangura B Shilkret K Crawford JT Frothingham R Kreiswirth BN Identification and evolution of an IS6110 low-copy-number Mycobacterium tuberculosis cluster J Infect Dis 2002 185 641 649 11865421 10.1086/339345
Warren RM Victor TC Streicher EM Richardson M van der Spuy GD Johnson R Chihota VN Locht C Supply P van Helden PD Clonal expansion of a globally disseminated lineage of Mycobacterium tuberculosis with low IS6110 copy numbers J Clin Microbiol 2004 42 5774 5782 15583312 10.1128/JCM.42.12.5774-5782.2004
Das SD Narayanan S Hari L Hoti SL Thangathurai RK Charles N Jaggarajamma K Narayanan PR Differentiation of highly prevalent IS6110 single-copy strains of Mycobacterium tuberculosis from a rural community in South India with an ongoing DOTS programme Infect Genet Evol 2005 5 67 77 15567140 10.1016/j.meegid.2004.06.007
Pym AS Brodin P Brosch R Huerre M Cole ST Loss of RD1 contributed to the attenuation of the live tuberculosis vaccines Mycobacterium bovis BCG and Mycobacterium microti Mol Microbiol 2002 46 709 717 12410828 10.1046/j.1365-2958.2002.03237.x
Leimkuhler S Rajagopalan KV In vitro incorporation of nascent molybdenum cofactor into human sulfite oxidase J Biol Chem 2001 276 1837 1844 11042213 10.1074/jbc.M007304200
Janion C A new look at adaptive mutations in bacteria Acta Biochim Pol 2000 47 451 457 11051210
Dubnau E Chan J Mohan VP Smith I responses of mycobacterium tuberculosis to growth in the mouse lung Infect Immun 2005 73 3754 3757 15908407 10.1128/IAI.73.6.3754-3757.2005
van Soolingen D Hermans PW de Haas PE Soll DR van Embden JD Occurrence and stability of insertion sequences in Mycobacterium tuberculosis complex strains: evaluation of an insertion sequence-dependent DNA polymorphism as a tool in the epidemiology of tuberculosis J Clin Microbiol 1991 29 2578 2586 1685494
Sreevatsan S Pan X Stockbauer KE Connell ND Kreiswirth BN Whittam TS Musser JM Restricted structural gene polymorphism in the Mycobacterium tuberculosis complex indicates evolutionarily recent global dissemination[] Proc Natl Acad Sci U S A 1997 94 9869 9874 9275218 10.1073/pnas.94.18.9869
|
16202156
|
PMC1262720
|
CC BY
|
2021-01-04 16:28:15
|
no
|
BMC Infect Dis. 2005 Oct 4; 5:81
|
utf-8
|
BMC Infect Dis
| 2,005 |
10.1186/1471-2334-5-81
|
oa_comm
|
==== Front
BMC ImmunolBMC Immunology1471-2172BioMed Central London 1471-2172-6-211616475310.1186/1471-2172-6-21Research ArticleMetallothionein mediates leukocyte chemotaxis Yin Xiuyun [email protected] David A [email protected] Michael A [email protected] Department of Molecular and Cell Biology, 91 North Eagleville Rd., U-3125, University of Connecticut, Storrs, CT USA 06269-31252005 15 9 2005 6 21 21 25 5 2005 15 9 2005 Copyright © 2005 Yin et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Metallothionein (MT) is a cysteine-rich, metal-binding protein that can be induced by a variety of agents. Modulation of MT levels has also been shown to alter specific immune functions. We have noticed that the MT genes map close to the chemokines Ccl17 and Cx3cl1. Cysteine motifs that characterize these chemokines are also found in the MT sequence suggesting that MT might also act as a chemotactic factor.
Results
In the experiments reported here, we show that immune cells migrate chemotactically in the presence of a gradient of MT. This response can be specifically blocked by two different monoclonal anti-MT antibodies. Exposure of cells to MT also leads to a rapid increase in F-actin content. Incubation of Jurkat T cells with cholera toxin or pertussis toxin completely abrogates the chemotactic response to MT. Thus MT may act via G-protein coupled receptors and through the cyclic AMP signaling pathway to initiate chemotaxis.
Conclusion
These results suggest that, under inflammatory conditions, metallothionein in the extracellular environment may support the beneficial movement of leukocytes to the site of inflammation. MT may therefore represent a "danger signal"; modifying the character of the immune response when cells sense cellular stress. Elevated metallothionein produced in the context of exposure to environmental toxicants, or as a result of chronic inflammatory disease, may alter the normal chemotactic responses that regulate leukocyte trafficking. Thus, MT synthesis may represent an important factor in immunomodulation that is associated with autoimmune disease and toxicant exposure.
==== Body
Background
Initiation of an immune response is accompanied by physiological changes that can produce a stressful environment for both the cells involved in the immune response, and for bystander cells that are part of adjacent but uninvolved tissues. These stresses can be further increased by the presence of infectious microorganisms. The changes to the environment include increases in reactive oxygen and reactive nitrogen species, products of cellular metabolism, and agents that initiate apoptotic or necrotic cell death.
Cells react to stressful environments with a broad range of different homeostatic responses. These responses can include the synthesis of a host of stress response proteins, including the heat shock proteins, acute phase cytokines, and metallothionein. Metallothionein is a novel member of this type of response with a unique biochemistry and an intriguing array of physiological roles. Metallothionein is small (about 7 kDa) and extremely thiol-rich [1]. The thiols participate in complexing with divalent metal cations [2]. When metallothionein binds to essential divalent metals (e.g. zinc and copper) it may serve as a metal reservoir for apoenzymes and zinc-finger transcription regulators [3,4]. Metallothionein that is induced by other divalent metal cations (e.g. mercury, cadmium,) protects essential cellular functions [5] and enhances the survival of both cells and whole organisms that are exposed to toxic heavy metals. The thiol-rich nature of metallothionein also enables it to regulate the redox potential of cells, and thus serves as a way of indirectly regulating redox-sensitive transcription via NF-kB [6]. There are also reports that link metallothionein to a much more direct interaction with NF-kB [7,8]. Metallothionein has also been found to be released to the extracellular environment in a number of different compartments, including cell culture media, serum, urine, bronchoalveolar spaces, liver sinusoids, and inflammatory lesions [9-12]. Extracellular metallothionein has been shown to have significant immunomodulatory effects both in vivo and in vitro [13-16] however the molecular mechanism(s) of this effect have yet to be elucidated.
Leukocyte movement is an essential component of the normal response to inflammatory signals. A variety of chemotactic agents can be produced by local immune cells, damaged bystander cells, and by invading microorganisms. In aggregate, these soluble signals determine the infiltration and departure of cells that participate in the inflammation, and serve as essential regulatory components of the immune response. Stress responses alter these patterns of leukocyte trafficking in various ways. For example, psychological stress in humans has been shown to increase both the magnitude of the cellular influx at an inflammatory site and the chemotactic index of peripheral blood mononuclear cells [17]. Restraint stress in hamsters has similarly increased leukocyte trafficking and delayed type hypersensitivity responses [18]. Xenobiotics may alter leukocyte trafficking in similar ways to diminish immune competence.
We have found that metallothionein has significant chemotactic activity for both cell lines and primary leukocytes. For most of the work described in this report we have used a new assay of chemotactic cell movement. The ECIS/taxis assay is sensitive enough to detect the response of a single cell, and allows automated, real-time quantification of cell movement (21). These results suggest that cell movement in stressful environments may be influenced by the presence of metallothionein, and that this protein could be an important therapeutic target for the manipulation of inflammation in vivo.
Results
Cysteines present in the primary amino acid sequence of metallothionein are arranged in cys, cys-cys, cys-X-cys, and cys-X3-cys motifs. These motifs are also found in chemokine molecules, and serve to differentiate chemokine families from one another. A comparison of the sequences of metallothionein and the chemokine Ccl17 is shown in Figure 1A. In addition to having similar cysteine motifs, Ccl17 is also located near the MT genes in both mice and humans (Figure 1B). Combined with our previously published observations of the impact extracellular metallothionein has on developing T-dependent humoral immunity [15] information suggested that metallothionein might display chemotactic activity.
Naïve splenocytes and thioglycollate-elicited leukocytes respond chemotactically to metallothionein in the Boyden chamber assay at a level similar to that observed when the cells are exposed to guinea pig serum used as a source of activated complement (Figure 2A). To avoid the heterogeneity of primary cell populations, the chemotactic dose response of cells to metallothionein was characterized using Jurkat T cells. Previous studies have shown that these T cells express CXCR4 (the stromal derived factor-1α (SDF-1α) receptor), and exposure to an SDF-1α gradient has previously been shown to induce a chemotactic response [19]. Using the Boyden chamber assay, we found that these cells also respond to metallothionein (Figure 2B) in a dose-dependent manner. The dose response curve shows a peak response at 14.3 μM, and the shape of the curve is consistent with that of other chemokines [20].
In addition to the Boyden chamber assay, chemotaxis was assessed using a recently developed technology called ECIS/taxis. This assay employs a miniature under-agarose chemotaxis chamber in which cell arrival on a surface microelectrode is measured by changes to electrical current flow through the electrode [21]. Unlike the Boyden chamber assay, the under-agarose configuration allows the establishment of stable and shallow chemotactic gradients that more accurately reflect the subtle gradients found in vivo than the Boyden chamber assay. In ECIS/taxis measurements, the total normalized resistance to current flow is proportional to the number of cells that occupy the surface of the target electrode. If no chemoattractant is added, few cells move out of the well and those that do never reach the target electrode and no change in resistance is recorded (Figure 3). In the presence of a gradient of metallothionein, the initial arrival of a small number of cells at the target electrode is indicated by the appearance of small, rapid resistance fluctuations. Over time as more cells accumulate on the electrode, a gradual increase in resistance is observed. The concentration of metallothionein added to the chemoattractant well that elicited the highest resistance increase (Figure 3) and the fastest movement of responding Jurkat T cells (data not shown) was 14.3 μM. This dose optimum is similar to that measured with the Boyden chamber assay. The average speed of the fastest Jurkat T cells, 1.56 ± 0.12 μm/min, was calculated from the time of arrival of the first cells at the target electrode. This speed is comparable to the speed of the Jurkat T cell response to a gradient of SDF-1α (1.43 ± 0.1 μm/min, Figure 3). We also tested metallothionein's effect on WBC 264-9C cells. These cells have been shown to exhibit chemotactic movement toward a source of activated complement [22]. In the ECIS/taxis assay, WBC 264-9C cells are chemotactic toward both activated complement and metallothionein (Figure 4). The metallothionein response was dose-dependent, and the optimal dose was similar to that found with Jurkat T cells (data not shown). The average speed of WBC 264-9C cells responding to a gradient of guinea pig serum used as a source of activated complement was 1.32 ± 0.2 μm/min, compared to 0.75 ± 0.03 μm/min for cells responding to metallothionein. However, the population response was nevertheless robust since total resistance (indicating the absolute number of responding cells) ultimately reached a level that approximates that produced by cells responding to activated complement.
In light of the potential for contaminants (e.g. small liver peptides that co-purify with the MT) in the commercially available metallothionein preparations, we affinity-purified metallothionein from commercial preparations using an anti-metallothionein monoclonal antibody (UC1MT) coupled to CNBr-activated Sepharose. The purified metallothionein also stimulates Jurkat T cell chemotaxis (Figure 5A). In addition, the chemotactic response to the original metallothionein (data not shown) or the affinity-purified metallothionein (Figure 5A) could be blocked by pre-incubation of the purified metallothionein with either UC1MT or E9 monoclonal anti-metallothionein antibodies. Neither of the two anti-metallothionein antibodies nor the isotype-matched IgG1 (MOPC 21) stimulated cell movement to the target electrode on their own (data not shown).
G protein activation has been shown to play a role in the chemotactic response and this pathway can be inhibited by cholera toxin (CTX) [23] or pertussis toxin (PTX) [24,25]. Jurkat T cells (106 cells/ml) were pre-incubated with 0.133 μg/ml CTX or 200 ng/ml pertussis toxin and then placed in a metallothionein gradient. The chemotactic effect of metallothionein on Jurkat T cells could be blocked by CTX (figure 5B) and by PTX (figure 5C). PTX was also capable of blocking the chemotaxis of Jurkat cells to SDF-1α.
A more direct assessment of chemotaxis was done using time-lapse video microscopy of cells moving in the presence of a metallothionein or SDF-1α gradient. When Jurkat T cells were exposed to a gradient similar to that present in the ECIS/taxis assay, they could be observed to move out of the cell well and continue up the gradient toward the chemoattractant well. Tracks of the outlines of these cells show persistent directional movement (Figure 6D, E). In the absence of a gradient few cells exit the cell well (data not shown), and those that do show little directional movement (Figure 6A–C). The speed, persistence and chemotactic indices of individual cell movements are consistent with the speeds calculated using the ECIS/taxis measurements of population movement (Table 1). In order to assess the role of chemokinesis in this process, cells were overlaid with a pre-formed agarose sheet containing a uniform concentration of metallothionein, SDF-1α or medium alone. The MT-exposed cells moved more rapidly than control cells, but the movement lacked directional persistence and was much slower than movement in a spatial gradient (Table 1 and Figure 6A–C). Similar results were obtained using a checkerboard analysis of cell movement in the Boyden chamber format (Table 2). Metallothionein added to the same side of the filter as the cells, or to both sides of the filter in equal concentration resulted in fewer cells reaching the lower surface of the filter than in wells where the metallothionein was added to the opposite side of the filter from the cells. This data supports the conclusion that metallothionein induces both chemotaxis and chemokinesis in Jurkat T cells.
Another hallmark of cellular responses to chemokines is a change in the amount and distribution of polymerized actin. Signal transduction through G protein coupled receptors causes reorganization of the actin cytoskeleton, leading to the formation of new F-actin rich lamellipods that extend in the direction of movement. This reorganization can be assessed by in vitro measurements of polymerized actin from cell extracts of stimulated cells with phalloidin [26]. Metallothionein stimulated a 19% increase in total F-actin within 30 seconds and a 79% increase by 2 minutes (Figure 7). The extent and timing of this response is consistent with receptor activation in other cell types [27,28].
Discussion
Cells of the immune system operate in a complex microenvironment where they are presented with a host of different and often conflicting signals [29,30]. The ways in which cells integrate and respond to these signals can ultimately govern the way in which the immune system will respond to antigen exposure. In some cases, the outcome is an activated immune response that is designed to eliminate the source of antigen. In other instances, the cells become anergic or undergo apoptosis and thus fail to initiate or participate in an immune response to that antigen. An early aspect of many immune responses is the directional movement of cells toward a site of infection or other injury. This directional movement is a response to chemotactic factors produced by some infectious organisms, to chemokines produced by cells already at the site of inflammation, or to other agents. Cells that express receptors for these signals can detect the gradient(s) of diffusing chemoattractants, and move toward the source of the agent. This chemotactic response is an essential aspect of lymphocyte trafficking.
In this report, we show that metallothionein can direct the chemotaxis of primary and transformed leukocytes. While metallothionein has been historically thought of as an intracellular protein, there are numerous reports that describe its presence in serum, urine [31], broncho-alveolar spaces [10], liver sinusoids [32], and other extracellular locations. While the mechanism(s) by which metallothionein is released from cells has yet to be determined, heat shock protein 70 [33], Interleukin 1β [34] and fibroblast growth factor [35] are among a set of proteins that lack signal sequences and nevertheless are released from cells by a non-traditional secretory mechanism. These results indicate that stress response proteins may gain access to the extracellular environment via mechanisms other than cell lysis, and suggest that a thorough understanding of the immunomodulatory roles played by metallothionein must include the extracellular compartment.
Metallothionein is synthesized in response to acute phase cytokines (e.g. IL-1, IL-6, and TNF-a) that are secreted at sites of inflammation [36-38] in a variety of contexts in which immune activities are changing. Metallothionein is also synthesized in cells exposed to glucocorticoids, a signal that is often associated with stressful environments [39]. Furthermore, metallothionein can be induced by reactive oxygen species, by endotoxin [40], and in cells exposed to divalent metal cations [1]. With all of these different initiators, it is not surprising that elevated metallothionein levels are detected in the context of neoplastic disease [41,42], autoimmune disease [43], chronic inflammation [44], and infection [45]. Previous work from our laboratory and others has shown that metallothionein can have significant immunomodulatory activities. For example, metallothionein can diminish T dependent humoral responses and it can alter the proliferative capacity of lymphocytes [16], diminish cytotoxic T cell function [46], and it can alter the effector function of macrophages [47]. Inadequate expression of metallothionein in the context of inflammatory disease can dramatically shorten life span [43], and exogenous metallothionein can diminish the severity of a collagen-induced arthritis [48].
There are a multitude of studies which show that different forms of stress originating from external sources can alter normal immune function [49]. Psychological, physical and chemical agents which induce stress each affect the immune system. In some instances, these stressors suppress effective immune functioning, which renders the individual susceptible to infectious pathogens. In other instances, the immune modifications result in undesirable increases in immune recognition of self antigens, ultimately resulting in autoimmune disease. These stressors are known to induce metallothionein synthesis and may alter immune functions in part via their effect on metallothionein.
We have demonstrated metallothionein-induced chemotactic cell movement in the traditional Boyden chamber assay, by computerized analysis of time-lapse images of cell movement, and using the ECIS/taxis assay. We have shown that the response of Jurkat T cells to a metallothionein gradient corresponds well with chemotactic responses of leukocytes to other agents. Jurkat T cells and WBC 264-9C migrate in response to a metallothionein gradient at speeds which are similar to those found in other systems [50,51]. In addition, the pattern of the dose response to metallothionein is a bell shaped curve similar to other classical chemokines [20]. One important consideration is whether the chemotactic response to metallothionein occurs at physiologically relevant concentrations. Chemoattractants can act over an extremely wide concentration range (e.g. 4 logs) [52,53] because the cells sense the local spatial differential in chemoattractant concentration. Our work shows that 1 to 10 μM metallothionein can stimulate chemotaxis of cells in both Boyden and under-agarose assays. Higher concentrations of metallothionein used in the ECIS/taxis assay refer to the concentrations added to the micro-volume chemoattractant wells, which are then diluted in the process of diffusion away from the source. The metallothionein amounts used in these experiments represent biologically reasonable concentrations, given that metallothionein has been measured at concentrations of 1 μM in serum (which would be substantially diluted from the source tissue concentration) in normal patients, and in individuals undergoing some form of stress (inflammation, cancer, toxicant exposure, etc.) [54].
The chemotactic response to metallothionein can be blocked by monoclonal antibodies to metallothionein while isotype-matched antibody has no effect. This blockade of the response is an important control, since commercial metallothionein preparations contain contaminating peptides from the liver tissue source (D. Lawrence, personal communication). Since both cholera toxin and pertussis toxin block the metallothionein-initiated chemotaxis, it is likely that G protein mediated signaling is involved in the response. Another common aspect of chemotactic signaling is an activation of the actin polymerization machinery in response to a sharp increase in chemoattractant concentration [26,55]. Metallothionein causes an increase both in total F-actin content and in peripheral F-actin. It will be of great interest to determine the receptor for metallothionein and the signal transduction pathway that leads to actin polymerization.
It is intriguing to speculate that once outside the cell, metallothionein serves as one of the many signals designed to draw immunocompetent cells to sites of cellular stress. Our observation(s) that metallothionein and anti-metallothionein injections modify immune activity in vivo suggest that there is an appropriate range of extracellular metallothionein in which leukocytes ordinarily function [14,15]. A pair of recent reports suggest that cytosolic constituents of apoptotic cells are released to the extracellular compartment and support the progression of the inflammatory process [56,57]. These reports further suggested that release of the cytosolic components of these dying cells might represent one of the signals central to the "Danger Hypothesis" proposed by Matzinger et al. [58,59]. This hypothesis holds that an active immune response cannot be mounted without a signal indicating that cellular damage has occurred. While other reports have suggested that heat shock proteins can fill this role [60], metallothionein is another potential candidate for the danger signal.
Methods
Cells
Jurkat T cells, (TIB-152, American Type Culture Collection (ATCC), Bethesda, MD) were maintained in complete RPMI 1640 media with L-glutamine containing 10% heat-inactivated FBS (Mediatech, Herndon, VA), 1% Sodium Bicarbonate, 100 units/ml penicillin, 0.1 mg/ml streptomycin, 1% sucrose, and 1 mM sodium pyruvate as recommended by ATCC. WBC 264-9C cell lines (HB-8902, ATCC) were kept in Minimum Essential Medium (Eagle) with Earle's balanced salt solution (BSS) containing 10% heat-inactivated fetal bovine serum. All cells were cultured in a humidified incubator with 5% CO2 in air at 37°C. The WBC 264-9C is a macrophage-like cell line that is chemotactic to N-formylmethionyl-leucyl-phenylalanine [61]. Media was replenished every three days.
Reagents
SDF-1α (Synthetic Human SDF-1α) was purchased from BD Biosciences (Bedford, MA). BSA (DNase, RNase, and Protease-free) and Hema-3 stain set kit were purchased from Fisher Scientific Inc. (Pittsburgh, PA). A mixture of Cd, Zn-metallothionein I and II purified from rabbit liver, mouse IgG1, kappa (MOPC21) purified immunoglobulin, and pertussis toxin, were purchased from Sigma Chemical Co. (St Louis, MO). SeaKem® GTG® Agarose was obtained from BioWhittaker Molecular Applications (Rockland, ME). Metallothionein monoclonal antibodies UC1MT (IgG1, kappa) [14,15], available from StressGen, Inc., Victoria, BC and E-9 (IgG1, kappa), purchased from Zymed Laboratories Inc. (San Francisco, CA) were used in some experiments. Cholera toxin was purchased from List Biological Laboratories, Inc. (through Cedarlane, Ltd., Hornby, ONT Canada). Guinea pig serum was purchased from Colorado Serum Company (Denver, CO).
Affinity purification of metallothionein
UC1MT was first purified on ProteinG-Sepharose (Sigma) according to manufacturer's instructions. The purified antibody was then coupled to CNBr-activated Sepharose (Sigma) according to manufacturer's instructions. Metallothionein I and II, prepared in PBS, was mixed with the immobilized UC1MT and allowed to bind. After unbound proteins were washed away from the affinity matrix, the specifically captured protein was eluted with 0.1 M glycine HCl, pH 2.8, adjusted pH to 7.4 and dialyzed against PBS.
Boyden chamber assay
The micro-Boyden assay was done using a 48 well chamber apparatus (NeuroProbe, Cabin John, MD). Polyvinylpyrrolidone (PVP)-free polycarbonate membrane filters with 5 μm pores were obtained from the same source. The lower chambers of the apparatus were loaded with 30 μl of diluted chemoattractant in media, PBS vehicle in media, or media alone and then covered with the membrane and the upper chambers. Fifty microliters of cell suspension (2 × 106 cells/ml) was then added to the upper chambers. After incubating for 2 hours in a humidified incubator at 37°C in 5% CO2, the filters were collected, cells that remained on the upper surface of the filter were removed and the filters were processed according to manufacturer's instructions. The numbers of migrated cells were counted under 400× magnification. For each of six replicate wells, the numbers of cells in at least six fields were determined and the mean and standard deviation was calculated.
ECIS/taxis assay
This assay was done as previously described with minor modifications [21]. Linear electrode ECIS chambers (Applied Biophysics, Inc. Troy, NY) were used in the assays described here. Target electrodes were 0.02 × 2 mm, and were used in an orientation in which the long axis of the target electrode was oriented perpendicularly to the direction of cell migration. All the chambers containing electrodes were pre-treated with 10 mM cysteine for 15 min at room temperature to stabilize the electrical performance of the gold electrodes, washed three times with sterile distilled water, and dried in a standard biosafety laminar-flow hood. Then 250 μl of molten 0.5% agarose gel (dissolved in RPMI 1640 with 10% FBS) was added to each chamber and allowed to cool. Two wells were cut with a sharpened 14 gauge cannula equally distant on either side of the electrode and separated a combined intrawell distance of approximately 1.9 to 2 mm. Then 7 μl of cell suspension (15 × 106 cells/ml for Jurkat T cells and 10 × 106 cells/ml for WBC 264-9C cells) was placed into the cell well and an equal volume of chemoattractant or vehicle control was dispensed into the opposite well. A 1 volt AC current of 4000 Hz is passed through the electrode, and the resistance of the circuit was calculated. Cell movement was assessed by measurements of changes in the resistance caused by arrival of cells at the target electrode. Data is reported as the change in resistance at the target electrode normalized to the initial resistance of the system. In addition to the general increase in resistance caused by cells covering the electrode, rapid fluctuations in resistance are indicative of changes in the shape and surface adherence of cells, and of continuing cell viability and movement.
Trough chemotaxis assay
For some chemotaxis experiments,3.5 ml of 0.5% agarose (dissolved in medium with 10% FBS and 20 mM HEPES) was loaded into a 35 mm Petri dish. After the agarose solidified, 2 wells separated by about 2 mm were cut in the agarose. One well was loaded with 7 μl of chemoattractant or media and the opposing well was loaded with 7 μl of cell suspension. The Petri dish was then sealed with Parafilm to retain moisture and incubated on the microscope stage at 37°C. Temperature was maintained by enclosing the microscope in a Styrofoam box in which a constant temperature airstream was provided by an Air-Therm feedback regulated heater (WPI, Inc., Sarasota FL). Chemokinesis in the under-agarose environment was investigated by seeding cells to the surface of the Petri dish in liquid media. After the cells had settled, the overlying media was removed and the cells were overlaid with a pre-gelled layer of agarose containing a uniform concentration of the different stimuli or medium alone. Images of the cells were taken at regular intervals using a CCD-72 analog video camera (Dage, Michigan City, IN) and Scion frame grabber controlled by Scion Image (Scion, Inc., Frederick, Maryland) software. The images were compiled into movies using public domain Image J software [62]. The trajectories of cells were analyzed from these movies using Dynamic Image Analysis System (DIAS) software (Solltech, Inc., Oakdale, IA). Trajectories of a number of cells from each condition (see n in table 1) were tracked and analyzed from the movies. Each cell was tracked for the same total time interval and the data is presented as the mean of all cells analyzed.
Measurements of actin polymerization: Cells were spun at 200 × g for 5 minutes and resuspended in RPMI 1640 containing 10% FBC at a density of 2 × 106 cells/ml. Cells were stimulated with either SDF-1α (data not shown) or 2 uM metallothionein at 37°C and then fixed with 3.7% formaldehyde for 15 minutes in Buffer F on a rotator at room temperature (5 mM KCl, 138 mM NaCl, 4 mM NaHCO3, 0.4 mM KH2PO4, 1.1 mM Na2HPO4, 2 mM MgCl2, 2 mM EGTA, 5 mM PIPES, pH 7.2). The fixed cells were centrifuged and resuspended in 1 ml of 0.5% Triton X-100 in Buffer F for 20 minutes and stained in 1 uM TRITC-Phalloidin in Buffer F on a rotator for 1 hour at room temperature. They were then pelleted, washed with 5 ml Buffer F twice, and re-suspended in 850 μl of Buffer F. The TRITC-Phalloidin fluorescence in the cell suspension was measured using a Spectramax M2 fluorimeter (Molecular Devices, Sunnyvale, CA) at 544 nm excitation and 580 nm emission. In each well, the raw fluorescence of 9 points were measured and used to calculate the average well fluorescence.
Authors' contributions
XY designed and carried out all of the experiments drafted the manuscript. DAK and MAL conceived of the study, and participated in its design and coordination and helped to author the manuscript. All authors read and approved the final manuscript.
Acknowledgements
This work was supported by grants from NIBIB (EB000208) to MAL and DAK, from NIGMS to DAK (GM40599) and from NIEHS (ES007408) to MAL. We thank Dr. Lawrence E. Hightower and Dr. Lisa A. Borghesi for careful reading of the manuscript.
Figures and Tables
Figure 1 Structural features of metallothionein gene and protein. A: Clustal alignment of Ccl17 and metallothionein (MT) protein sequences. Amino acid similarity is set at 85% (gray). Identical amino acids are boxed. B: Mouse chromosome 8 showing expanded region from map position 43 to 47 shown on left. Homologous human genes and their chromosomal map positions are indicated on the right. Synteny map generated by the Mouse Genome Database [63].
Figure 2 Chemotactic responses of cells to metallothionein in the modified Boyden chamber assay. A. Mouse splenocytes or Thioglycollate-elicited cells respond to guinea pig serum (as a source of activated complement) and to metallothionein. Cells that had migrated to the lower side of the filters were counted in at least six fields of view per well. Data are means of six fields in each of six replicates. The data shown is representative of three experiments and is expressed as the average ± standard deviation. B. Dose response curve of chemotaxis to metallothionein. Different concentrations of metallothionein were loaded into the lower wells of the Boyden chambers and Jurkat T cells were added to the upper wells.
Figure 3 Dose response of Jurkat T cells to metallothionein measured by ECIS/taxis. Jurkat T cells were added to the cell wells and the indicated chemoattractant was loaded into the opposing well. SDF-1α is used as a positive control. The resistance measured at the electrode between the two wells is shown. Inset: A 3 hour window of data on an expanded scale is shown to highlight the arrival of cells on the electrode. Arrows indicated the time of arrival of the first cells leading to the appearance of small fluctuations in resistance.
Figure 4 ECIS/taxis assessment of the chemotactic response by WBC 264-9C cells. Cell wells were loaded and guinea pig serum (1:2 dilution), PBS, or 20 μM metallothionein was added to the chemoattractant wells. Serum complement is activated by exposure to agarose to generate a control chemotactic gradient. Inset: A 3 hours window of data on an expanded scale showing the arrival of cells on the electrode.
Figure 5 Inhibitors of MT chemotaxis. A: Monoclonal anti-metallothionein antibody blocks the chemotactic response of Jurkat T cells to 20 μM metallothionein gradient. Affinity-purified metallothionein (20 μM) was incubated with antibody (clone UC1MT or E9) and then added to the chemoattractant well. Controls were performed with metallothionein alone (MT), medium alone (-), or metallothionein preincubated with isotype-matched control antibody (not shown). B: Cholera toxin (CTX) inhibits chemotactic response to metallothionein. Jurkat T cells were preincubated with or without 0.133 μg/106 cells/ml CTX at 37°C for 1 h and then added to the cell well. Chemotactic responses of toxin-treated cells to MT (CTX-MT) were compared to untreated cells (MT) and untreated cells without an MT gradient (-). C: Pertussis toxin (PTX) blocks the chemotactic response to metallothionein. Jurkat T cells were preincubated with or without 200 ng/106 cells/ml PTX at 37°C for 16 h. Chemotactic responses of toxin-treated cells to MT (PTX-MT) was compared to untreated cells (MT) and untreated cells without an MT gradient (-). Each figure is a representative more than four independent experiments performed in triplicate.
Figure 6 Cell movement in the presence and absence of metallothionein and SDF-1α. Images at 3 minute intervals (over a 75 minute time period) are presented for untreated cells (A), cells treated with a uniform concentration of MT(0.5 μM) (B), or SDF-1α (100 ng/ml) (C), or cells moving under agar in a chemotactic gradient of MT (20 μM) (D) or SDF-1α (200 ng/ml) (E). Each panel has the same number of images and represents the same total time interval. The line at the bottom of panels D and E represents the direction of the chemoattractant gradient from high (wide) to low (narrow). Only cells that have exited the cell well and are clearly visible under the agarose were analyzed.
Figure 7 Actin reorganization in metallothionein-stimulated cells. Jurkat T cells were treated with metallothionein at 2 μM and then harvested and fixed at various times after stimulation. Cells were stained with rhodamine phalloidin and F-actin fluorescence quantified using a plate fluorimeter. (RFU = relative fluorescent units). Values reported represent the average of 3 replicates ± standard deviation. *** represents significant difference from the medium alone control (p < 0.001). This data is representative of 3 independent experiments.
Table 1 DIAS analysis of chemokinetic and chemotactic movement. Jurkat T cells were cultured under agarose in the presence of uniform concentrations or in a gradient of the stimuli. Cells were imaged over time and their motile behavior quantified. Persistence is an indicator of the rate of directional change. Chemotactic index is a measure of the proportion of movement in a designated direction (1 = toward a source, 0 = random movement, -1 = away from the source).
Stimulus (n = number of analyzed cells) Average translational speed (μm/min) Persistence (μm/minute-degree) Chemotactic index
Media control (7) 0.66 ± 0.29 0.18 ± 0.11 0.19
Uniform concentration of metallothionein (10) 2.110 ± 0.73 0.5 ± 0.22 0.07
Uniform concentration of SDF-1α (10) 1.69 ± 0.82 0.53 ± 0.34 0.09
Metallothionein gradient (7) 6.6 ± 3.2 2.042 ± 1.09 0.73
SDF-1α gradient (7) 3.31 ± 1.48 1.17 ± 0.49 0.72
Table 2 Checkerboard analysis of chemokinetic cell movement induced by metallothionein. Metallothionein has both chemokinetic and chemotactic activities. Significant chemotactic movement is measured when metallothionein is presented from below the filter. Chemokinetic movement is indicated when metallothionein is present above the filter, or in both chambers. The data is representative of three independent experiments.
Cell number migrated to lower side of membrane
Metallothionein (μM) above filter
Metallothionein (μM) below filter 0 2.5 5.0 10
0 14.0 ± 4.3 90.8 ± 17.5 43.0 ± 3.2 20.6 ± 9.2
2.5 55.3 ± 17.0 61.3 ± 23.9 38.7 ± 12.5 25.5 ± 13.2
5.0 192.8 ± 45.4 93.6 ± 20.1 67.0 ± 12.5 37.0 ± 12.1
10 80.2 ± 23.7 96.0 ± 21.8 68.0 ± 21.2 30.8 ± 10.3
==== Refs
Hamer DH Metallothionein Annu Rev Biochem 1986 55 913 951 3527054
Furey WF Robbins AH Clancy LL Winge DR Wang BC Stout CD Crystal structure of Cd,Zn metallothionein Science 1986 231 704 710 3945804
Ejnik J Munoz A Gan T Shaw CF Petering DH Interprotein metal ion exchange between cadmium-carbonic anhydrase and apo- or zinc-metallothionein J Biol Inorg Chem 1999 4 784 790 10631610 10.1007/s007750050351
Zeng J Heuchel R Schaffner W Kagi JH Thionein (apometallothionein) can modulate DNA binding and transcription activation by zinc finger containing factor Sp1 FEBS Lett 1991 279 310 312 2001744 10.1016/0014-5793(91)80175-3
Theocharis SE Margeli AP Koutselinis A Metallothionein: a multifunctional protein from toxicity to cancer Int J Biol Markers 2003 18 162 169 14535585
Sakurai A Hara S Okano N Kondo Y Inoue J Imura N Regulatory role of metallothionein in NF-kappaB activation FEBS Lett 1999 455 55 58 10428471 10.1016/S0014-5793(99)00839-X
Butcher HL Kennette WA Collins O Zalups RK Koropatnick J Metallothionein mediates the level and activity of nuclear factor {kappa}B (NF-kB) in murine fibroblasts J Pharmacol Exp Ther 2004 15039454
Kim CH Kim JH Lee J Ahn YS Zinc-induced NF-kappaB inhibition can be modulated by changes in the intracellular metallothionein level Toxicol Appl Pharmacol 2003 190 189 196 12878048 10.1016/S0041-008X(03)00167-4
Bruwer M Schmid KW Metz KA Krieglstein CF Senninger N Schurmann G Increased expression of metallothionein in inflammatory bowel disease Inflamm Res 2001 50 289 293 11475329
Hart BA Garvey JS Detection of metallothionein in bronchoalveolar cells and lavage fluid following repeated cadmium inhalation Environ Res 1986 40 391 398 3732210 10.1016/S0013-9351(86)80114-1
Nordberg GF Garvey JS Chang CC Metallothionein in plasma and urine of cadmium workers Environ Res 1982 28 179 182 7049686 10.1016/0013-9351(82)90167-0
Penkowa M Espejo C Ortega-Aznar A Hidalgo J Montalban X Martinez Caceres EM Metallothionein expression in the central nervous system of multiple sclerosis patients Cell Mol Life Sci 2003 60 1258 1266 12861391
Borghesi LA Lynes MA Nonprotective effects of extracellular metallothionein Toxicol Appl Pharmacol 1996 139 6 14 8685909 10.1006/taap.1996.0137
Canpolat E Lynes MA In vivo manipulation of endogenous metallothionein with a monoclonal antibody enhances a T-dependent humoral immune response Toxicol Sci 2001 62 61 70 11399794
Lynes MA Borghesi LA Youn J Olson EA Immunomodulatory activities of extracellular metallothionein. I. Metallothionein effects on antibody production Toxicology 1993 85 161 177 8303711 10.1016/0300-483X(93)90040-Y
Lynes MA Garvey JS Lawrence DA Extracellular metallothionein effects on lymphocyte activities Mol Immunol 1990 27 211 219 2342485 10.1016/0161-5890(90)90132-J
Redwine L Snow S Mills P Irwin M Acute psychological stress: effects on chemotaxis and cellular adhesion molecule expression Psychosom Med 2003 65 598 603 12883110 10.1097/01.PSY.0000079377.86193.A8
Bilbo SD Dhabhar FS Viswanathan K Saul A Yellon SM Nelson RJ Short day lengths augment stress-induced leukocyte trafficking and stress-induced enhancement of skin immune function Proc Natl Acad Sci U S A 2002 99 4067 4072 11904451 10.1073/pnas.062001899
Bleul CC Farzan M Choe H Parolin C Clark-Lewis I Sodroski J Springer TA The lymphocyte chemoattractant SDF-1 is a ligand for LESTR/fusin and blocks HIV-1 entry Nature 1996 382 829 833 8752280 10.1038/382829a0
Corcione A Tortolina G Bonecchi R Battilana N Taborelli G Malavasi F Sozzani S Ottonello L Dallegri F Pistoia V Chemotaxis of human tonsil B lymphocytes to CC chemokine receptor (CCR) 1, CCR2 and CCR4 ligands is restricted to non-germinal center cells Int Immunol 2002 14 883 892 12147625 10.1093/intimm/dxf054
Hadjout N Laevsky G Knecht DA Lynes MA Automated real-time measurement of chemotactic cell motility Biotechniques 2001 31 1130 1138 11730019
Fung M Lu M Fure H Sun W Sun C Shi NY Dou Y Su J Swanson X Mollnes TE Pre-neutralization of C5a-mediated effects by the monoclonal antibody 137-26 reacting with the C5a moiety of native C5 without preventing C5 cleavage Clin Exp Immunol 2003 133 160 169 12869020 10.1046/j.1365-2249.2003.02213.x
Maghazachi AA al-Aoukaty A Schall TJ C-C chemokines induce the chemotaxis of NK and IL-2-activated NK cells. Role for G proteins J Immunol 1994 153 4969 4977 7525719
Su SB Silver PB Zhang M Chan CC Caspi RR Pertussis toxin inhibits induction of tissue-specific autoimmune disease by disrupting G protein-coupled signals J Immunol 2001 167 250 256 11418656
Backlund PSJ Meade BD Manclark CR Cantoni GL Aksamit RR Pertussis toxin inhibition of chemotaxis and the ADP-ribosylation of a membrane protein in a human-mouse hybrid cell line Proc Natl Acad Sci U S A 1985 82 2637 2641 3857605
Segall JE Tyerech S Boselli L Masseling S Helft J Chan A Jones J Condeelis J EGF stimulates lamellipod extension in metastatic mammary adenocarcinoma cells by an actin-dependent mechanism Clin Exp Metastasis 1996 14 61 72 8521618
Condeelis J Hall A Bresnick A Warren V Hock R Bennett H Ogihara S Actin polymerization and pseudopod extension during amoeboid chemotaxis Cell Motil Cytoskeleton 1988 10 77 90 3052871 10.1002/cm.970100113
Rao KM Varani J Actin polymerization induced by chemotactic peptide and concanavalin A in rat neutrophils J Immunol 1982 129 1605 1607 6286771
Heit B Tavener S Raharjo E Kubes P An intracellular signaling hierarchy determines direction of migration in opposing chemotactic gradients J Cell Biol 2002 159 91 102 12370241 10.1083/jcb.200202114
Gouwy M Struyf S Catusse J Proost P Van Damme J Synergy between proinflammatory ligands of G protein-coupled receptors in neutrophil activation and migration J Leukoc Biol 2004 15075362
Evering WE Haywood S Bremner I Wood AM Trafford J The protective role of metallothionein in copper-overload: II. Transport and excretion of immunoreactive MT-1 in blood, bile and urine of copper-loaded rats Chem Biol Interact 1991 78 297 305 2070434 10.1016/0009-2797(91)90060-K
Danielson KG Ohi S Huang PC Immunochemical localization of metallothionein in rat liver and kidney J Histochem Cytochem 1982 30 1033 1039 6752263
Hightower LE Guidon PTJ Selective release from cultured mammalian cells of heat-shock (stress) proteins that resemble glia-axon transfer proteins J Cell Physiol 1989 138 257 266 2918030 10.1002/jcp.1041380206
Andrei C Dazzi C Lotti L Torrisi MR Chimini G Rubartelli A The secretory route of the leaderless protein interleukin 1beta involves exocytosis of endolysosome-related vesicles Mol Biol Cell 1999 10 1463 1475 10233156
Backhaus R Zehe C Wegehingel S Kehlenbach A Schwappach B Nickel W Unconventional protein secretion: membrane translocation of FGF-2 does not require protein unfolding J Cell Sci 2004 117 1727 1736 15075234 10.1242/jcs.01027
Penkowa M Hidalgo J IL-6 deficiency leads to reduced metallothionein-I+II expression and increased oxidative stress in the brain stem after 6-aminonicotinamide treatment Exp Neurol 2000 163 72 84 10785446 10.1006/exnr.2000.7383
Hernandez J Carrasco J Belloso E Giralt M Bluethmann H Kee Lee D Andrews GK Hidalgo J Metallothionein induction by restraint stress: role of glucocorticoids and IL-6 Cytokine 2000 12 791 796 10843766 10.1006/cyto.1999.0629
Schroeder JJ Cousins RJ Interleukin 6 regulates metallothionein gene expression and zinc metabolism in hepatocyte monolayer cultures Proc Natl Acad Sci U S A 1990 87 3137 3141 2326272
Karin M Herschman HR Glucocorticoid hormone receptor mediated induction of metallothionein synthesis in HeLa cells J Cell Physiol 1980 103 35 40 7430258 10.1002/jcp.1041030106
Suzuki KT Yamamura M Induction of hepatic zinc-thionein in rat by endotoxin Biochem Pharmacol 1980 29 2260 7426032 10.1016/0006-2952(80)90210-5
Zelger B Hittmair A Schir M Ofner C Ofner D Fritsch PO Bocker W Jasani B Schmid KW Immunohistochemically demonstrated metallothionein expression in malignant melanoma Histopathology 1993 23 257 263 8225244
Schmid KW Ellis IO Gee JM Darke BM Lees WE Kay J Cryer A Stark JM Hittmair A Ofner D Presence and possible significance of immunocytochemically demonstrable metallothionein over-expression in primary invasive ductal carcinoma of the breast Virchows Arch A Pathol Anat Histopathol 1993 422 153 159 8385380 10.1007/BF01607167
Lynes MARCAMCRCKCLJCYJSIBSLD Klaassen C Metallothionein-mediated changes in cell populations of autoimmune mice Metallothionein IV 1999 Basel , Birkhauser Verlag 437 444
Min KS Kim H Fujii M Tetsuchikawahara N Onosaka S Glucocorticoids suppress the inflammation-mediated tolerance to acute toxicity of cadmium in mice Toxicol Appl Pharmacol 2002 178 1 7 11781073 10.1006/taap.2001.9323
Sobocinski PZ Canterbury WJJ Mapes CA Dinterman RE Involvement of hepatic metallothioneins in hypozincemia associated with bacterial infection Am J Physiol 1978 234 E399 406 645856
Youn J Lynes MA Metallothionein-induced suppression of cytotoxic T lymphocyte function: an important immunoregulatory control Toxicol Sci 1999 52 199 208 10630572 10.1093/toxsci/52.2.199
Youn J Borghesi LA Olson EA Lynes MA Immunomodulatory activities of extracellular metallothionein. II. Effects on macrophage functions J Toxicol Environ Health 1995 45 397 413 7643428
Youn J Hwang SH Ryoo ZY Lynes MA Paik DJ Chung HS Kim HY Metallothionein suppresses collagen-induced arthritis via induction of TGF-beta and down-regulation of proinflammatory mediators Clin Exp Immunol 2002 129 232 239 12165078 10.1046/j.1365-2249.2002.01922.x
Padgett DA Glaser R How stress influences the immune response Trends Immunol 2003 24 444 448 12909458 10.1016/S1471-4906(03)00173-X
Allen WE Zicha D Ridley AJ Jones GE A role for Cdc42 in macrophage chemotaxis J Cell Biol 1998 141 1147 1157 9606207 10.1083/jcb.141.5.1147
Jones GE Prigmore E Calvez R Hogan C Dunn GA Hirsch E Wymann MP Ridley AJ Requirement for PI 3-kinase gamma in macrophage migration to MCP-1 and CSF-1 Exp Cell Res 2003 290 120 131 14516793 10.1016/S0014-4827(03)00318-5
Foxman EF Campbell JJ Butcher EC Multistep navigation and the combinatorial control of leukocyte chemotaxis J Cell Biol 1997 139 1349 1360 9382879 10.1083/jcb.139.5.1349
Varnum B Soll DR Effects of cAMP on single cell motility in Dictyostelium J Cell Biol 1984 99 1151 1155 6088555 10.1083/jcb.99.3.1151
Nakayama A Fukuda H Ebara M Hamasaki H Nakajima K Sakurai H A new diagnostic method for chronic hepatitis, liver cirrhosis, and hepatocellular carcinoma based on serum metallothionein, copper, and zinc levels Biol Pharm Bull 2002 25 426 431 11995918 10.1248/bpb.25.426
McRobbie SJ Newell PC Chemoattractant-mediated changes in cytoskeletal actin of cellular slime moulds J Cell Sci 1984 68 139 151 6092396
Shi Y Zheng W Rock KL Cell injury releases endogenous adjuvants that stimulate cytotoxic T cell responses Proc Natl Acad Sci U S A 2000 97 14590 14595 11106387 10.1073/pnas.260497597
Shi Y Rock KL Cell death releases endogenous adjuvants that selectively enhance immune surveillance of particulate antigens Eur J Immunol 2002 32 155 162 11754356 10.1002/1521-4141(200201)32:1<155::AID-IMMU155>3.0.CO;2-P
Matzinger P Tolerance, danger, and the extended family Annu Rev Immunol 1994 12 991 1045 8011301
Matzinger P An innate sense of danger Semin Immunol 1998 10 399 415 9840976 10.1006/smim.1998.0143
Moseley P Stress proteins and the immune response Immunopharmacology 2000 48 299 302 10960671 10.1016/S0162-3109(00)00227-7
Aksamit RR A human-mouse hybrid cell line that stably expresses chemotaxis to N-formylmethionyl-leucyl-phenylalanine Biochem Biophys Res Commun 1986 138 1001 1008 3741419 10.1016/S0006-291X(86)80595-2
Rasband WS ImageJ National Institiutes of Health, Bethesda, MD, USA 1997 http://rsb.info.nih.gov/ij/
Mouse Genome Database (MGD) TJLBHM http://www.informatics.jax.org
|
16164753
|
PMC1262721
|
CC BY
|
2021-01-04 16:03:32
|
no
|
BMC Immunol. 2005 Sep 15; 6:21
|
utf-8
|
BMC Immunol
| 2,005 |
10.1186/1471-2172-6-21
|
oa_comm
|
==== Front
BMC Med GenetBMC Medical Genetics1471-2350BioMed Central London 1471-2350-6-341618804310.1186/1471-2350-6-34Research ArticleEvaluation of the toll-like receptor 6 Ser249Pro polymorphism in patients with asthma, atopic dermatitis and chronic obstructive pulmonary disease Hoffjan Sabine [email protected] Susanne [email protected] Qumar [email protected] Elisabeth [email protected] Umut [email protected] Gernot [email protected] Karin [email protected] Gerhard [email protected] Albrecht [email protected] Jörg T [email protected] Department of Human Genetics, Ruhr-University Bochum, Germany2 Private medical practice, Gladbeck, Germany3 Department of Neuroanatomy and Molecular Brain Research, Ruhr-University Bochum, Germany4 Department of Internal Medicine lll, Pneumology, Allergology and Sleep Medicine, Ruhr-University Bochum, Germany5 Department of Experimental Pneumology, Ruhr-University Bochum, Germany2005 28 9 2005 6 34 34 3 6 2005 28 9 2005 Copyright © 2005 Hoffjan et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
For allergic disorders, the increasing prevalence over the past decade has been attributed in part to the lack of microbial burden in developed countries ('hygiene hypothesis'). Variation in genes encoding toll-like receptors (TLRs) as the receptor system for the first innate immune response to microbial stimuli has been implicated in various inflammatory diseases. We evaluated here the role of a coding variation, Ser249Pro, in the TLR6 gene in the pathogenesis of asthma, atopic dermatitis (AD) and chronic obstructive pulmonary disease (COPD).
Methods
Genotyping of the Ser249Pro polymorphism in 68 unrelated adult patients and 132 unrelated children with asthma, 185 unrelated patients with COPD, 295 unrelated individuals with AD and 212 healthy control subjects was performed by restriction enzyme digestion.
Results
We found a weak association of the 249Ser allele with childhood asthma (p = 0.03). Yet, significance was lost after Bonferroni correction. No association was evident for AD or COPD.
Conclusion
Variation in TLR6 might play a role in the pathogenesis of childhood asthma.
==== Body
Background
Asthma, atopic dermatitis (AD) and chronic obstructive pulmonary disease (COPD) are common chronic inflammatory diseases with prevalence rates between 5 and 15%, making them major public-health problems worldwide [1,2]. For all three diseases, a multifactorial background has been suggested with genetic as well as environmental factors influencing disease susceptibility [3-5]. There is now strong evidence that exposure to microbial products in early childhood plays an important role in the postnatal maturation of the immune system and that the increase in the prevalence of allergic diseases over the last few decades might be in part due to the decreased microbial burden in industrialized countries (the so-called hygiene hypothesis) [6,7]. The receptor system that constitutes the first, innate immune response to microbial stimuli consists of the family of toll-like receptors (TLRs), highly conserved receptor complexes that recognize pathogen-associated molecular patterns (PAMPs) [8]. Currently, 10 different TLRs are known each of which recognizes a different spectrum of PAMPs. As activators of cytokine production in response to infections, TLRs are believed to be involved in the establishment of T helper (Th)1 immune responses in early life, counter-balancing the Th2-dominated cytokine spectrum at birth [9]. Genetic variation in these receptors could thus influence susceptibility for both Th1- (e.g., autoimmune) and Th2-mediated (e.g., allergic) diseases [9].
While several studies have evaluated the role of variation in TLR4, the receptor for endotoxin, for allergic diseases [10-12], little is known about the role of TLR6 variation so far. TLR6 forms heterodimers with TLR2 for recognition of bacterial lipopeptides [13]. Expression of TLR6 has been demonstrated in human mast cells [14] which play an important role in allergic diseases [15]. Interestingly, activation of mast cells via, both, TLR4 and TLR2/TLR6 resulted in additive effects on cytokine production [16]. Thus, TLR6 appears as interesting candidate gene for allergic disorders. The TLR6 gene is located on chromosome 4p13 and it comprises a single exon encoding a 796 amino-acid polypeptide. Recently, a common coding polymorphism was identified in this gene that leads to an exchange from serine to proline at position 249 in the extracellular domain of the TLR6 protein [17]. In preliminary analyses, the T allele (249Ser) was associated with protection from asthma in African American samples, while the same trend was not significant for European Americans [17]. Although the association reported in this study is intriguing considering the hygiene hypothesis, a replication of this finding for asthma in a different sample has not yet been accomplished. We further speculated that TLR6 coding variation might also play a role in the pathogenesis of COPD since it is – like asthma – a common chronic lung disease characterized by chronic airway inflammation and airway hyper-responsiveness [1]. In addition, there is growing evidence that at least part of the genetic background might be common among COPD and asthma [18]. We therefore studied the TLR6 Ser249Pro polymorphism in four different German cohorts: adult asthmatics, pediatric asthmatics, patients with COPD and patients with AD in comparison with matched German controls.
Methods
Subjects
185 unrelated adult patients with COPD and 68 unrelated adult patients with asthma were recruited while hospitalized at the Bergmannsheil clinics, Ruhr-University Bochum, Germany, and 132 unrelated children with asthma in the Pediatric Pneumology Studycenter (PPS) located at the same place. The COPD diagnosis was based on the classification of the Global Initiative for Chronic Obstructive Lung Disease (GOLD) criteria [19]. Patients included in the study had at least a forced expiratory volume in one second (FEV1) of < 80% of predicted and a FEV1/FVC (forced vital capacity) ratio of < 70% of predicted, classified as moderate to severe COPD (GOLD stages II-IV). Adult asthmatics had a doctor's diagnosis of asthma according to the standards of the American Thoracic Society [20]. The exclusion criteria were α1-antitrypsin deficiency, dyspnea of other origin (including cardiovascular disorders, pneumonia, interstitial lung disease, pleural disease, upper airways obstruction, neuromuscular disease and anaemia) and bronchial carcinoma. For the asthmatic children, the asthma diagnosis was based on the ISAAC questionnaire [21]. Disease severity was assessed by a symptom score recently validated [22]. 294 unrelated patients with AD were recruited by a consultant specialist for AD (QP, Gladbeck, Germany), including 175 children and 119 adults. The AD diagnosis was based on the presence of clinical features, including purities, eczema with age-dependent differences in location, xerosis and chronic or relapsing dermatitis. In addition, all investigated AD patients had a positive family history for atopic diseases.
212 control samples from adults without known allergies, asthma or AD were collected in the same private practice as the AD patients. We specifically chose to use non-allergic adults as controls because for diseases as frequent as asthma and AD, the risk remains very high for asymptomatic children to develop an allergic disease during childhood or even adulthood [23,24]. The control subjects underwent clinical examination in order to exclude symptoms of asthma, AD or COPD, had no self-reported allergies or allergic symptoms and no first degree relatives with known allergic diseases. Lung function tests or IgE measures were not performed for the controls. All patient and control subjects were Caucasians of German origin. Informed consent was obtained from all subjects. The study was approved by the Ethics Committee of the University of Bochum and the Declaration of Helsiniki protocols were followed. DNA was extracted from EDTA anti-coagulated peripheral blood by using a standard salting-out method [25].
Genotyping
Genotyping of the TLR6 Ser249Pro polymorphism was performed by polymerase chain reaction (PCR) with subsequent restriction enzyme digestion. PCR reactions were performed in a total volume of 10 μl, containing 50 ng DNA, 200 mmol of each dNTP, 3 mmol MgCl2, 10 pmol of each primer (forward: GCATTTCCAAGTCGTTTCTATGT; reverse: GCAAAAACCCTTCACCTTGTT), and 0.4 U Taq polymerase (Genecraft, Münster, Germany). Thermal cycling was performed (Biometra, Göttingen, Germany). After two initial cycles at 6°C and 3°C above the annealing temperature, 27 cycles of 95°C (30 sec), 57°C (60 sec) and 72°C (60 sec) were run. The PCR product was digested with Ava II (0.01 U/ng DNA) at 37°C for three hours. The fragments were subsequently separated on 2% agarose gels in 1 × TBE buffer (30 min, 200 V) and visualized using ethidium bromide staining.
Statistical analyses
Genotype and allele frequencies were ascertained by direct counting and subsequently analyzed according to the χ2 method. Deviations from Hardy-Weinberg equilibrium were evaluated using the FINETTI program. P < 0.05 was considered to be significant. Power calculations were performed using the Power and sample size program [26].
Results and discussion
Clinical data concerning the four patient groups are summarized in table 1. Since the nucleotide substitution from C to T at position 744 in the TLR6 gene results in the creation of a new restriction site, the Ser249Pro polymorphism was genotyped by restriction enzyme digestion. Genotype frequencies for case and control groups are shown in table 2. The frequencies in all five groups were in Hardy-Weinberg equilibrium. There was a significant association between the T allele (249Ser) and childhood asthma (p = 0.03, table 3). Yet, significance was lost after Bonferroni correction. For the other three case groups, no significant association was observed.
Although it seems puzzling to identify the opposite allele being associated with childhood asthma as compared to the one reported by Tantisira et al. [17], there are plausible explanations for this phenomenon. First, the association with childhood asthma reported here is only a weak association that did not withstand Bonferroni correction. Yet, we consider the study to be explorative and hypothesis-generating and find the results intriguing in the light of the emerging role of TLRs in allergic diseases [9]. In addition, the reported association refers to a coding variation that might at least theoretically have functional consequences. It remains largely unknown what sort of genetic variants explain inherited variation in complex traits, but recent evidence suggests that common, non-coding genetic variants will explain at least some of the inherited variation in susceptibility to common disease [27]. Thus, testing a single functional variation would be more likely biased into the direction of a negative result rather than of a positive association.
Second, the previous association of the 249Ser allele with protection from asthma was defined in a sample of African American patients and controls. The minor allele frequency of the Ser249Pro polymorphism in this cohort was 0.08 in asthmatics and 0.19 in controls [17] and thus much lower than in our four German cohorts with a minor allele frequency of on average 0.40 (range 0.35 to 0.49). Yet, the allele frequencies we observed in this study are comparable to the ones for European Americans [28]. It appears possible that the Ser249Pro polymorphism might not be disease-causing by itself, but instead be in linkage disequilibrium with the true disease-causing variation. In this case, different alleles of Ser249Pro could be linked to the relevant allele in different populations (e.g., African Americans and Caucasians). Further association studies in these and other populations are needed to answer this question.
Third, environmental factors might play a role in disease susceptibility associated with opposite alleles. One such example has been described for CD14, which associates with TLR4 to form the receptor complex that recognizes lipopolysaccharide (LPS, endotoxin). A single nucleotide polymorphism (SNP) in the CD14 promoter at position -159 (C/T) was identified [29]. The -159T allele was associated with high levels of soluble CD14 and decreased total serum IgE in a cohort of children from Tucson, Arizona [29]. Yet, no association of this SNP with allergy or IgE levels was evident in a large German cohort [30]. In the Hutterites, an isolated population from South Dakota, the -159T allele was instead associated with increased risk for atopy [31]. Vercelli recently postulated an intriguing explanation for this phenomenon: she suggested that the level of endotoxin exposure influences the 'switch' from the Th2-biased cytokine profile at birth to a Th1-biased cytokine profile in early childhood, and that endotoxin levels might interact with the CD14 genotype to confer either risk to or protection from atopic phenotypes later in life [32]. Thus, environmental factors – and even endotoxin load – might also be responsible for the discrepancy between the findings of Tantisira et al. [17] and our results, concerning the role of the Ser249Pro polymorphism in asthma pathogenesis. Since we did not yet measure endotoxin levels, we are unable to explore this question at the moment.
Association studies of polymorphisms in other TLR genes have also revealed somewhat contradictory results. Two missense mutations (Asp299Gly and Thr399Ile) in the gene encoding TLR4 which mediates the biological response to LPS were associated with a decreased response to inhaled endotoxin in humans [33]. Since exposure to LPS in early life has been suggested to exert protective effects on the development of allergic diseases [34-36], the TLR4 gene appeared to be an outstanding functional candidate for the susceptibility to asthma and allergic diseases. Yet only one out of four association studies found a direct association of the TLR4 Asp299Gly polymorphism with asthma [10]. In three other studies, no differences in the overall risk for asthma between carriers of the wild type and mutant genotypes were obvious [11,12,37]. On the other hand, an impact of this polymorphism on the severity of atopy in asthmatics [11] was reported as well as a modified response to endotoxin [12], indicating gene/environment interplay. Similarly, a polymorphism at position -16934 in the TLR2 gene was significantly associated with asthma and atopy in farmers' children but not in children that did not grow up on farms [38], pointing again to a gene/environment interaction. Thus, variation in TLR genes appears to have modifying effects on asthma susceptibility.
The fact that we did not find an association of the Ser249Pro polymorphism with adult asthma might be due to the small sample size of adult asthmatics. Yet, in order to achieve a significance level of 0.05 with a statistical power of 80%, a sample size of over 500 patients would have been required. Replication studies in larger cohorts as well as functional studies are clearly needed. To our knowledge, this is the first evaluation of the TLR6 Ser249Pro polymorphism in patients with atopic dermatitis, and we did not find association with this allergic skin disease, suggesting that the observed association might be restricted to (childhood) asthma, rather than allergic diseases in general. This finding is not totally unexpected since the results of genome-wide screens indicate that loci linked to AD do not overlap substantially with loci linked to asthma and related phenotypes, but rather with loci for psoriasis, another chronic skin disease [4]. Furthermore, we did not find evidence that TLR6 Ser249Pro contributes to susceptibility for COPD.
Conclusion
We found a weak association between the TLR6 Ser249Pro polymorphism and risk for childhood asthma, while no association was evident for AD or COPD. Analysis of the effect of this variation on disease severity and potential gene/environment interactions might be a promising approach since gene/environment interactions have been described for other TLR genes. Further studies of the TLR6 Ser249Pro polymorphism in ethnically well-defined cohorts as well as functional studies of this variation are warranted to evaluate its role in the pathogenesis of allergic diseases.
List of abbreviations
AD atopic dermatitis
COPD chronic obstructive pulmonary disease
FEV1 forced exspiratory volume in the first second
FVC forced vital capacity
LPS lipopolysaccharide
PAMP pathogen-associated molecular pattern
PCR polymerase chain reaction
TLR toll-like receptor
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SH was in charge of the design and coordination of the atopic dermatitis association study, performed the statistical analysis and drafted the manuscript. SS coordinated the asthma and COPD association studies. EPP participated in the design and coordination of the AD study. QP, UA, GR and KRR participated in the recruitement of patients and clinical data collection. GSW, AB and JTE participated in the design and coordination of the whole study and helped to draft the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This work was supported by the "Forschungsförderung der Ruhr-Universität Bochum Medizinische Fakultät" (FoRUM) grants #F377-03 and #175-99 and by the "Bundesministerium für Bildung und Forschung" (BMBF) grant #01GC 0101/TP6.
We thank Daniela Falkenstein and Natascha Wirkus for technical assistance, Wolfram Klein for helpful discussions and the patients for participating in this study.
Figures and Tables
Table 1 Clinical data of the asthma, AD and COPD cohorts as well as the control group
Adult patients with asthma Children with asthma AD COPD Healthy controls
number of subjects 68 132 294 182 212
gender: f/m 36/32 52/80 180/114 62/120 131/81
age (years) [median(range)] 38 (18–68) 9.5 (2–15) 12 (0.5–72) 68 (33–81) 59 (22–87)
FEV1 (%pred.) (mean ± STD) 73 ± 25 97 ± 18 -- 47 ± 19 --
FEV1/FVC ratio (mean ± STD) 65 ± 16 98 ± 15 -- 49 ± 15 --
pack years [median(range)] 2 (0–44) -- -- 30 (0–200) --
Table 2 Genotype frequencies of the TLR6 Ser249Pro polymorphism in the case and control groups.
TLR6 Ser249Pro Controls Asthma adults Asthma children AD COPD
N = 212 N = 68 N = 132 N = 294 N = 185
Pro/Pro 75 28 32 108 76
Pro/Ser 104 32 72 138 80
Ser/Ser 33 8 28 48 29
p-value - n.s. 0.07 n.s. n.s.
Table 3 Allele frequencies of the TLR6 Ser249Pro polymorphism in asthma children vs. controls.
Asthma children
N = 132 Controls
N = 212
Pro 136 (51.5%) 254 (59.9%)
Ser 128 (48.5%) 170 (40.1%)
p = 0.03
==== Refs
Skrepnek GH Skrepnek SV Epidemiology, clinical and economic burden, and natural history of chronic obstructive pulmonary disease and asthma Am J Manag Care 2004 10 S129 38 15354678
Leung DY Bieber T Atopic dermatitis Lancet 2003 361 151 160 12531593 10.1016/S0140-6736(03)12193-9
Hoffjan S Ober C Present status on the genetic studies of asthma Curr Opin Immunol 2002 14 709 717 12413520 10.1016/S0952-7915(02)00393-X
Bowcock AM Cookson WO The genetics of psoriasis, psoriatic arthritis and atopic dermatitis Hum Mol Genet 2004 13 Spec No 1 R43 55 14996755 10.1093/hmg/ddh094
Molfino NA Genetics of COPD Chest 2004 125 1929 1940 15136409 10.1378/chest.125.5.1929
Martinez FD The coming-of-age of the hygiene hypothesis Respir Res 2001 2 129 132 11686875 10.1186/rr48
Weiss ST Eat dirt--the hygiene hypothesis and allergic diseases N Engl J Med 2002 347 930 931 12239263 10.1056/NEJMe020092
Janssens S Beyaert R Role of Toll-like receptors in pathogen recognition Clin Microbiol Rev 2003 16 637 646 14557290 10.1128/CMR.16.4.637-646.2003
Cook DN Pisetsky DS Schwartz DA Toll-like receptors in the pathogenesis of human disease Nat Immunol 2004 5 975 979 15454920 10.1038/ni1116
Fageras Bottcher M Hmani-Aifa M Lindstrom A Jenmalm MC Mai XM Nilsson L Zdolsek HA Bjorksten B Soderkvist P Vaarala O A TLR4 polymorphism is associated with asthma and reduced lipopolysaccharide-induced interleukin-12(p70) responses in Swedish children J Allergy Clin Immunol 2004 114 561 567 15356557 10.1016/j.jaci.2004.04.050
Yang IA Barton SJ Rorke S Cakebread JA Keith TP Clough JB Holgate ST Holloway JW Toll-like receptor 4 polymorphism and severity of atopy in asthmatics Genes Immun 2004 5 41 45 14735148 10.1038/sj.gene.6364037
Werner M Topp R Wimmer K Richter K Bischof W Wjst M Heinrich J TLR4 gene variants modify endotoxin effects on asthma J Allergy Clin Immunol 2003 112 323 330 12897738 10.1067/mai.2003.1648
Takeda K Akira S Toll-like receptors in innate immunity Int Immunol 2005 17 1 14 15585605 10.1093/intimm/dxh186
McCurdy JD Olynych TJ Maher LH Marshall JS Cutting edge: distinct Toll-like receptor 2 activators selectively induce different classes of mediator production from human mast cells J Immunol 2003 170 1625 1629 12574323
Boyce JA The role of mast cells in asthma Prostaglandins Leukot Essent Fatty Acids 2003 69 195 205 12895603 10.1016/S0952-3278(03)00081-4
Hume DA Underhill DM Sweet MJ Ozinsky AO Liew FY Aderem A Macrophages exposed continuously to lipopolysaccharide and other agonists that act via toll-like receptors exhibit a sustained and additive activation state BMC Immunol 2001 2 11 11686851 10.1186/1471-2172-2-11
Tantisira K Klimecki WT Lazarus R Palmer LJ Raby BA Kwiatkowski DJ Silverman E Vercelli D Martinez FD Weiss ST Toll-like receptor 6 gene (TLR6): single-nucleotide polymorphism frequencies and preliminary association with the diagnosis of asthma Genes Immun 2004 5 343 346 15266299 10.1038/sj.gene.6364096
Meyers DA Larj MJ Lange L Genetics of asthma and COPD. Similar results for different phenotypes Chest 2004 126 105S 110S; discussion 159S-161S 15302770 10.1378/chest.126.2_suppl_1.105S
Hurd S Pauwels R Global Initiative for Chronic Obstructive Lung Diseases (GOLD) Pulm Pharmacol Ther 2002 15 353 355 12220939 10.1006/pupt.2002.0381
Standards for the diagnosis and care of patients with chronic obstructive pulmonary disease (COPD) and asthma. This official statement of the American Thoracic Society was adopted by the ATS Board of Directors, November 1986 Am Rev Respir Dis 1987 136 225 244 3605835
Asher MI Keil U Anderson HR Beasley R Crane J Martinez F Mitchell EA Pearce N Sibbald B Stewart AW International Study of Asthma and Allergies in Childhood (ISAAC): rationale and methods Eur Respir J 1995 8 483 491 7789502 10.1183/09031936.95.08030483
Bufe A Ziegler-Kirbach E Stoeckmann E Heidemann P Gehlhar K Holland-Letz T Braun W Efficacy of sublingual swallow immunotherapy in children with severe grass pollen allergic symptoms: a double-blind placebo-controlled study Allergy 2004 59 498 504 15080830 10.1111/j.1398-9995.2004.00457.x
Bel EH Clinical phenotypes of asthma Curr Opin Pulm Med 2004 10 44 50 14749605 10.1097/00063198-200401000-00008
De Marco R Locatelli F Cerveri I Bugiani M Marinoni A Giammanco G Incidence and remission of asthma: a retrospective study on the natural history of asthma in Italy J Allergy Clin Immunol 2002 110 228 235 12170262 10.1067/mai.2002.125600
Miller SA Dykes DD Polesky HF A simple salting out procedure for extracting DNA from human nucleated cells Nucleic Acids Res 1988 16 1215 3344216
The Vanderbilt Medical Center Power and Sample Size Calculation
Newton-Cheh C Hirschhorn JN Genetic association studies of complex traits: design and analysis issues Mutat Res 2005 573 54 69 15829237
Innate Immunity in Heart, Lung and Blood disease
Baldini M Lohman IC Halonen M Erickson RP Holt PG Martinez FD A Polymorphism* in the 5' flanking region of the CD14 gene is associated with circulating soluble CD14 levels and with total serum immunoglobulin E Am J Respir Cell Mol Biol 1999 20 976 983 10226067
Sengler C Haider A Sommerfeld C Lau S Baldini M Martinez F Wahn U Nickel R Evaluation of the CD14 C-159 T polymorphism in the German Multicenter Allergy Study cohort Clin Exp Allergy 2003 33 166 169 12580907 10.1046/j.1365-2222.2003.01549.x
Ober C Tsalenko A Parry R Cox NJ A second-generation genomewide screen for asthma-susceptibility alleles in a founder population Am J Hum Genet 2000 67 1154 1162 11022011
Vercelli D Learning from discrepancies: CD14 polymorphisms, atopy and the endotoxin switch Clin Exp Allergy 2003 33 153 155 12580903 10.1046/j.1365-2222.2003.01606.x
Arbour NC Lorenz E Schutte BC Zabner J Kline JN Jones M Frees K Watt JL Schwartz DA TLR4 mutations are associated with endotoxin hyporesponsiveness in humans Nat Genet 2000 25 187 191 10835634 10.1038/76048
Braun-Fahrlander C The role of the farm environment and animal contact for the development of asthma and allergies Clin Exp Allergy 2001 31 1799 1803 11737027 10.1046/j.1365-2222.2001.01269.x
Gehring U Bischof W Fahlbusch B Wichmann HE Heinrich J House dust endotoxin and allergic sensitization in children Am J Respir Crit Care Med 2002 166 939 944 12359650 10.1164/rccm.200203-256OC
Gehring U Bischof W Schlenvoigt G Richter K Fahlbusch B Wichmann HE Heinrich J Exposure to house dust endotoxin and allergic sensitization in adults Allergy 2004 59 946 952 15291902 10.1111/j.1398-9995.2004.00551.x
Raby BA Klimecki WT Laprise C Renaud Y Faith J Lemire M Greenwood C Weiland KM Lange C Palmer LJ Lazarus R Vercelli D Kwiatkowski DJ Silverman EK Martinez FD Hudson TJ Weiss ST Polymorphisms in toll-like receptor 4 are not associated with asthma or atopy-related phenotypes Am J Respir Crit Care Med 2002 166 1449 1456 12406828 10.1164/rccm.200207-634OC
Eder W Klimecki W Yu L von Mutius E Riedler J Braun-Fahrlander C Nowak D Martinez FD Toll-like receptor 2 as a major gene for asthma in children of European farmers J Allergy Clin Immunol 2004 113 482 488 15007351 10.1016/j.jaci.2003.12.374
|
16188043
|
PMC1262722
|
CC BY
|
2021-01-04 16:03:34
|
no
|
BMC Med Genet. 2005 Sep 28; 6:34
|
utf-8
|
BMC Med Genet
| 2,005 |
10.1186/1471-2350-6-34
|
oa_comm
|
==== Front
BMC Med GenetBMC Medical Genetics1471-2350BioMed Central London 1471-2350-6-341618804310.1186/1471-2350-6-34Research ArticleEvaluation of the toll-like receptor 6 Ser249Pro polymorphism in patients with asthma, atopic dermatitis and chronic obstructive pulmonary disease Hoffjan Sabine [email protected] Susanne [email protected] Qumar [email protected] Elisabeth [email protected] Umut [email protected] Gernot [email protected] Karin [email protected] Gerhard [email protected] Albrecht [email protected] Jörg T [email protected] Department of Human Genetics, Ruhr-University Bochum, Germany2 Private medical practice, Gladbeck, Germany3 Department of Neuroanatomy and Molecular Brain Research, Ruhr-University Bochum, Germany4 Department of Internal Medicine lll, Pneumology, Allergology and Sleep Medicine, Ruhr-University Bochum, Germany5 Department of Experimental Pneumology, Ruhr-University Bochum, Germany2005 28 9 2005 6 34 34 3 6 2005 28 9 2005 Copyright © 2005 Hoffjan et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
For allergic disorders, the increasing prevalence over the past decade has been attributed in part to the lack of microbial burden in developed countries ('hygiene hypothesis'). Variation in genes encoding toll-like receptors (TLRs) as the receptor system for the first innate immune response to microbial stimuli has been implicated in various inflammatory diseases. We evaluated here the role of a coding variation, Ser249Pro, in the TLR6 gene in the pathogenesis of asthma, atopic dermatitis (AD) and chronic obstructive pulmonary disease (COPD).
Methods
Genotyping of the Ser249Pro polymorphism in 68 unrelated adult patients and 132 unrelated children with asthma, 185 unrelated patients with COPD, 295 unrelated individuals with AD and 212 healthy control subjects was performed by restriction enzyme digestion.
Results
We found a weak association of the 249Ser allele with childhood asthma (p = 0.03). Yet, significance was lost after Bonferroni correction. No association was evident for AD or COPD.
Conclusion
Variation in TLR6 might play a role in the pathogenesis of childhood asthma.
==== Body
Background
Asthma, atopic dermatitis (AD) and chronic obstructive pulmonary disease (COPD) are common chronic inflammatory diseases with prevalence rates between 5 and 15%, making them major public-health problems worldwide [1,2]. For all three diseases, a multifactorial background has been suggested with genetic as well as environmental factors influencing disease susceptibility [3-5]. There is now strong evidence that exposure to microbial products in early childhood plays an important role in the postnatal maturation of the immune system and that the increase in the prevalence of allergic diseases over the last few decades might be in part due to the decreased microbial burden in industrialized countries (the so-called hygiene hypothesis) [6,7]. The receptor system that constitutes the first, innate immune response to microbial stimuli consists of the family of toll-like receptors (TLRs), highly conserved receptor complexes that recognize pathogen-associated molecular patterns (PAMPs) [8]. Currently, 10 different TLRs are known each of which recognizes a different spectrum of PAMPs. As activators of cytokine production in response to infections, TLRs are believed to be involved in the establishment of T helper (Th)1 immune responses in early life, counter-balancing the Th2-dominated cytokine spectrum at birth [9]. Genetic variation in these receptors could thus influence susceptibility for both Th1- (e.g., autoimmune) and Th2-mediated (e.g., allergic) diseases [9].
While several studies have evaluated the role of variation in TLR4, the receptor for endotoxin, for allergic diseases [10-12], little is known about the role of TLR6 variation so far. TLR6 forms heterodimers with TLR2 for recognition of bacterial lipopeptides [13]. Expression of TLR6 has been demonstrated in human mast cells [14] which play an important role in allergic diseases [15]. Interestingly, activation of mast cells via, both, TLR4 and TLR2/TLR6 resulted in additive effects on cytokine production [16]. Thus, TLR6 appears as interesting candidate gene for allergic disorders. The TLR6 gene is located on chromosome 4p13 and it comprises a single exon encoding a 796 amino-acid polypeptide. Recently, a common coding polymorphism was identified in this gene that leads to an exchange from serine to proline at position 249 in the extracellular domain of the TLR6 protein [17]. In preliminary analyses, the T allele (249Ser) was associated with protection from asthma in African American samples, while the same trend was not significant for European Americans [17]. Although the association reported in this study is intriguing considering the hygiene hypothesis, a replication of this finding for asthma in a different sample has not yet been accomplished. We further speculated that TLR6 coding variation might also play a role in the pathogenesis of COPD since it is – like asthma – a common chronic lung disease characterized by chronic airway inflammation and airway hyper-responsiveness [1]. In addition, there is growing evidence that at least part of the genetic background might be common among COPD and asthma [18]. We therefore studied the TLR6 Ser249Pro polymorphism in four different German cohorts: adult asthmatics, pediatric asthmatics, patients with COPD and patients with AD in comparison with matched German controls.
Methods
Subjects
185 unrelated adult patients with COPD and 68 unrelated adult patients with asthma were recruited while hospitalized at the Bergmannsheil clinics, Ruhr-University Bochum, Germany, and 132 unrelated children with asthma in the Pediatric Pneumology Studycenter (PPS) located at the same place. The COPD diagnosis was based on the classification of the Global Initiative for Chronic Obstructive Lung Disease (GOLD) criteria [19]. Patients included in the study had at least a forced expiratory volume in one second (FEV1) of < 80% of predicted and a FEV1/FVC (forced vital capacity) ratio of < 70% of predicted, classified as moderate to severe COPD (GOLD stages II-IV). Adult asthmatics had a doctor's diagnosis of asthma according to the standards of the American Thoracic Society [20]. The exclusion criteria were α1-antitrypsin deficiency, dyspnea of other origin (including cardiovascular disorders, pneumonia, interstitial lung disease, pleural disease, upper airways obstruction, neuromuscular disease and anaemia) and bronchial carcinoma. For the asthmatic children, the asthma diagnosis was based on the ISAAC questionnaire [21]. Disease severity was assessed by a symptom score recently validated [22]. 294 unrelated patients with AD were recruited by a consultant specialist for AD (QP, Gladbeck, Germany), including 175 children and 119 adults. The AD diagnosis was based on the presence of clinical features, including purities, eczema with age-dependent differences in location, xerosis and chronic or relapsing dermatitis. In addition, all investigated AD patients had a positive family history for atopic diseases.
212 control samples from adults without known allergies, asthma or AD were collected in the same private practice as the AD patients. We specifically chose to use non-allergic adults as controls because for diseases as frequent as asthma and AD, the risk remains very high for asymptomatic children to develop an allergic disease during childhood or even adulthood [23,24]. The control subjects underwent clinical examination in order to exclude symptoms of asthma, AD or COPD, had no self-reported allergies or allergic symptoms and no first degree relatives with known allergic diseases. Lung function tests or IgE measures were not performed for the controls. All patient and control subjects were Caucasians of German origin. Informed consent was obtained from all subjects. The study was approved by the Ethics Committee of the University of Bochum and the Declaration of Helsiniki protocols were followed. DNA was extracted from EDTA anti-coagulated peripheral blood by using a standard salting-out method [25].
Genotyping
Genotyping of the TLR6 Ser249Pro polymorphism was performed by polymerase chain reaction (PCR) with subsequent restriction enzyme digestion. PCR reactions were performed in a total volume of 10 μl, containing 50 ng DNA, 200 mmol of each dNTP, 3 mmol MgCl2, 10 pmol of each primer (forward: GCATTTCCAAGTCGTTTCTATGT; reverse: GCAAAAACCCTTCACCTTGTT), and 0.4 U Taq polymerase (Genecraft, Münster, Germany). Thermal cycling was performed (Biometra, Göttingen, Germany). After two initial cycles at 6°C and 3°C above the annealing temperature, 27 cycles of 95°C (30 sec), 57°C (60 sec) and 72°C (60 sec) were run. The PCR product was digested with Ava II (0.01 U/ng DNA) at 37°C for three hours. The fragments were subsequently separated on 2% agarose gels in 1 × TBE buffer (30 min, 200 V) and visualized using ethidium bromide staining.
Statistical analyses
Genotype and allele frequencies were ascertained by direct counting and subsequently analyzed according to the χ2 method. Deviations from Hardy-Weinberg equilibrium were evaluated using the FINETTI program. P < 0.05 was considered to be significant. Power calculations were performed using the Power and sample size program [26].
Results and discussion
Clinical data concerning the four patient groups are summarized in table 1. Since the nucleotide substitution from C to T at position 744 in the TLR6 gene results in the creation of a new restriction site, the Ser249Pro polymorphism was genotyped by restriction enzyme digestion. Genotype frequencies for case and control groups are shown in table 2. The frequencies in all five groups were in Hardy-Weinberg equilibrium. There was a significant association between the T allele (249Ser) and childhood asthma (p = 0.03, table 3). Yet, significance was lost after Bonferroni correction. For the other three case groups, no significant association was observed.
Although it seems puzzling to identify the opposite allele being associated with childhood asthma as compared to the one reported by Tantisira et al. [17], there are plausible explanations for this phenomenon. First, the association with childhood asthma reported here is only a weak association that did not withstand Bonferroni correction. Yet, we consider the study to be explorative and hypothesis-generating and find the results intriguing in the light of the emerging role of TLRs in allergic diseases [9]. In addition, the reported association refers to a coding variation that might at least theoretically have functional consequences. It remains largely unknown what sort of genetic variants explain inherited variation in complex traits, but recent evidence suggests that common, non-coding genetic variants will explain at least some of the inherited variation in susceptibility to common disease [27]. Thus, testing a single functional variation would be more likely biased into the direction of a negative result rather than of a positive association.
Second, the previous association of the 249Ser allele with protection from asthma was defined in a sample of African American patients and controls. The minor allele frequency of the Ser249Pro polymorphism in this cohort was 0.08 in asthmatics and 0.19 in controls [17] and thus much lower than in our four German cohorts with a minor allele frequency of on average 0.40 (range 0.35 to 0.49). Yet, the allele frequencies we observed in this study are comparable to the ones for European Americans [28]. It appears possible that the Ser249Pro polymorphism might not be disease-causing by itself, but instead be in linkage disequilibrium with the true disease-causing variation. In this case, different alleles of Ser249Pro could be linked to the relevant allele in different populations (e.g., African Americans and Caucasians). Further association studies in these and other populations are needed to answer this question.
Third, environmental factors might play a role in disease susceptibility associated with opposite alleles. One such example has been described for CD14, which associates with TLR4 to form the receptor complex that recognizes lipopolysaccharide (LPS, endotoxin). A single nucleotide polymorphism (SNP) in the CD14 promoter at position -159 (C/T) was identified [29]. The -159T allele was associated with high levels of soluble CD14 and decreased total serum IgE in a cohort of children from Tucson, Arizona [29]. Yet, no association of this SNP with allergy or IgE levels was evident in a large German cohort [30]. In the Hutterites, an isolated population from South Dakota, the -159T allele was instead associated with increased risk for atopy [31]. Vercelli recently postulated an intriguing explanation for this phenomenon: she suggested that the level of endotoxin exposure influences the 'switch' from the Th2-biased cytokine profile at birth to a Th1-biased cytokine profile in early childhood, and that endotoxin levels might interact with the CD14 genotype to confer either risk to or protection from atopic phenotypes later in life [32]. Thus, environmental factors – and even endotoxin load – might also be responsible for the discrepancy between the findings of Tantisira et al. [17] and our results, concerning the role of the Ser249Pro polymorphism in asthma pathogenesis. Since we did not yet measure endotoxin levels, we are unable to explore this question at the moment.
Association studies of polymorphisms in other TLR genes have also revealed somewhat contradictory results. Two missense mutations (Asp299Gly and Thr399Ile) in the gene encoding TLR4 which mediates the biological response to LPS were associated with a decreased response to inhaled endotoxin in humans [33]. Since exposure to LPS in early life has been suggested to exert protective effects on the development of allergic diseases [34-36], the TLR4 gene appeared to be an outstanding functional candidate for the susceptibility to asthma and allergic diseases. Yet only one out of four association studies found a direct association of the TLR4 Asp299Gly polymorphism with asthma [10]. In three other studies, no differences in the overall risk for asthma between carriers of the wild type and mutant genotypes were obvious [11,12,37]. On the other hand, an impact of this polymorphism on the severity of atopy in asthmatics [11] was reported as well as a modified response to endotoxin [12], indicating gene/environment interplay. Similarly, a polymorphism at position -16934 in the TLR2 gene was significantly associated with asthma and atopy in farmers' children but not in children that did not grow up on farms [38], pointing again to a gene/environment interaction. Thus, variation in TLR genes appears to have modifying effects on asthma susceptibility.
The fact that we did not find an association of the Ser249Pro polymorphism with adult asthma might be due to the small sample size of adult asthmatics. Yet, in order to achieve a significance level of 0.05 with a statistical power of 80%, a sample size of over 500 patients would have been required. Replication studies in larger cohorts as well as functional studies are clearly needed. To our knowledge, this is the first evaluation of the TLR6 Ser249Pro polymorphism in patients with atopic dermatitis, and we did not find association with this allergic skin disease, suggesting that the observed association might be restricted to (childhood) asthma, rather than allergic diseases in general. This finding is not totally unexpected since the results of genome-wide screens indicate that loci linked to AD do not overlap substantially with loci linked to asthma and related phenotypes, but rather with loci for psoriasis, another chronic skin disease [4]. Furthermore, we did not find evidence that TLR6 Ser249Pro contributes to susceptibility for COPD.
Conclusion
We found a weak association between the TLR6 Ser249Pro polymorphism and risk for childhood asthma, while no association was evident for AD or COPD. Analysis of the effect of this variation on disease severity and potential gene/environment interactions might be a promising approach since gene/environment interactions have been described for other TLR genes. Further studies of the TLR6 Ser249Pro polymorphism in ethnically well-defined cohorts as well as functional studies of this variation are warranted to evaluate its role in the pathogenesis of allergic diseases.
List of abbreviations
AD atopic dermatitis
COPD chronic obstructive pulmonary disease
FEV1 forced exspiratory volume in the first second
FVC forced vital capacity
LPS lipopolysaccharide
PAMP pathogen-associated molecular pattern
PCR polymerase chain reaction
TLR toll-like receptor
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SH was in charge of the design and coordination of the atopic dermatitis association study, performed the statistical analysis and drafted the manuscript. SS coordinated the asthma and COPD association studies. EPP participated in the design and coordination of the AD study. QP, UA, GR and KRR participated in the recruitement of patients and clinical data collection. GSW, AB and JTE participated in the design and coordination of the whole study and helped to draft the manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This work was supported by the "Forschungsförderung der Ruhr-Universität Bochum Medizinische Fakultät" (FoRUM) grants #F377-03 and #175-99 and by the "Bundesministerium für Bildung und Forschung" (BMBF) grant #01GC 0101/TP6.
We thank Daniela Falkenstein and Natascha Wirkus for technical assistance, Wolfram Klein for helpful discussions and the patients for participating in this study.
Figures and Tables
Table 1 Clinical data of the asthma, AD and COPD cohorts as well as the control group
Adult patients with asthma Children with asthma AD COPD Healthy controls
number of subjects 68 132 294 182 212
gender: f/m 36/32 52/80 180/114 62/120 131/81
age (years) [median(range)] 38 (18–68) 9.5 (2–15) 12 (0.5–72) 68 (33–81) 59 (22–87)
FEV1 (%pred.) (mean ± STD) 73 ± 25 97 ± 18 -- 47 ± 19 --
FEV1/FVC ratio (mean ± STD) 65 ± 16 98 ± 15 -- 49 ± 15 --
pack years [median(range)] 2 (0–44) -- -- 30 (0–200) --
Table 2 Genotype frequencies of the TLR6 Ser249Pro polymorphism in the case and control groups.
TLR6 Ser249Pro Controls Asthma adults Asthma children AD COPD
N = 212 N = 68 N = 132 N = 294 N = 185
Pro/Pro 75 28 32 108 76
Pro/Ser 104 32 72 138 80
Ser/Ser 33 8 28 48 29
p-value - n.s. 0.07 n.s. n.s.
Table 3 Allele frequencies of the TLR6 Ser249Pro polymorphism in asthma children vs. controls.
Asthma children
N = 132 Controls
N = 212
Pro 136 (51.5%) 254 (59.9%)
Ser 128 (48.5%) 170 (40.1%)
p = 0.03
==== Refs
Skrepnek GH Skrepnek SV Epidemiology, clinical and economic burden, and natural history of chronic obstructive pulmonary disease and asthma Am J Manag Care 2004 10 S129 38 15354678
Leung DY Bieber T Atopic dermatitis Lancet 2003 361 151 160 12531593 10.1016/S0140-6736(03)12193-9
Hoffjan S Ober C Present status on the genetic studies of asthma Curr Opin Immunol 2002 14 709 717 12413520 10.1016/S0952-7915(02)00393-X
Bowcock AM Cookson WO The genetics of psoriasis, psoriatic arthritis and atopic dermatitis Hum Mol Genet 2004 13 Spec No 1 R43 55 14996755 10.1093/hmg/ddh094
Molfino NA Genetics of COPD Chest 2004 125 1929 1940 15136409 10.1378/chest.125.5.1929
Martinez FD The coming-of-age of the hygiene hypothesis Respir Res 2001 2 129 132 11686875 10.1186/rr48
Weiss ST Eat dirt--the hygiene hypothesis and allergic diseases N Engl J Med 2002 347 930 931 12239263 10.1056/NEJMe020092
Janssens S Beyaert R Role of Toll-like receptors in pathogen recognition Clin Microbiol Rev 2003 16 637 646 14557290 10.1128/CMR.16.4.637-646.2003
Cook DN Pisetsky DS Schwartz DA Toll-like receptors in the pathogenesis of human disease Nat Immunol 2004 5 975 979 15454920 10.1038/ni1116
Fageras Bottcher M Hmani-Aifa M Lindstrom A Jenmalm MC Mai XM Nilsson L Zdolsek HA Bjorksten B Soderkvist P Vaarala O A TLR4 polymorphism is associated with asthma and reduced lipopolysaccharide-induced interleukin-12(p70) responses in Swedish children J Allergy Clin Immunol 2004 114 561 567 15356557 10.1016/j.jaci.2004.04.050
Yang IA Barton SJ Rorke S Cakebread JA Keith TP Clough JB Holgate ST Holloway JW Toll-like receptor 4 polymorphism and severity of atopy in asthmatics Genes Immun 2004 5 41 45 14735148 10.1038/sj.gene.6364037
Werner M Topp R Wimmer K Richter K Bischof W Wjst M Heinrich J TLR4 gene variants modify endotoxin effects on asthma J Allergy Clin Immunol 2003 112 323 330 12897738 10.1067/mai.2003.1648
Takeda K Akira S Toll-like receptors in innate immunity Int Immunol 2005 17 1 14 15585605 10.1093/intimm/dxh186
McCurdy JD Olynych TJ Maher LH Marshall JS Cutting edge: distinct Toll-like receptor 2 activators selectively induce different classes of mediator production from human mast cells J Immunol 2003 170 1625 1629 12574323
Boyce JA The role of mast cells in asthma Prostaglandins Leukot Essent Fatty Acids 2003 69 195 205 12895603 10.1016/S0952-3278(03)00081-4
Hume DA Underhill DM Sweet MJ Ozinsky AO Liew FY Aderem A Macrophages exposed continuously to lipopolysaccharide and other agonists that act via toll-like receptors exhibit a sustained and additive activation state BMC Immunol 2001 2 11 11686851 10.1186/1471-2172-2-11
Tantisira K Klimecki WT Lazarus R Palmer LJ Raby BA Kwiatkowski DJ Silverman E Vercelli D Martinez FD Weiss ST Toll-like receptor 6 gene (TLR6): single-nucleotide polymorphism frequencies and preliminary association with the diagnosis of asthma Genes Immun 2004 5 343 346 15266299 10.1038/sj.gene.6364096
Meyers DA Larj MJ Lange L Genetics of asthma and COPD. Similar results for different phenotypes Chest 2004 126 105S 110S; discussion 159S-161S 15302770 10.1378/chest.126.2_suppl_1.105S
Hurd S Pauwels R Global Initiative for Chronic Obstructive Lung Diseases (GOLD) Pulm Pharmacol Ther 2002 15 353 355 12220939 10.1006/pupt.2002.0381
Standards for the diagnosis and care of patients with chronic obstructive pulmonary disease (COPD) and asthma. This official statement of the American Thoracic Society was adopted by the ATS Board of Directors, November 1986 Am Rev Respir Dis 1987 136 225 244 3605835
Asher MI Keil U Anderson HR Beasley R Crane J Martinez F Mitchell EA Pearce N Sibbald B Stewart AW International Study of Asthma and Allergies in Childhood (ISAAC): rationale and methods Eur Respir J 1995 8 483 491 7789502 10.1183/09031936.95.08030483
Bufe A Ziegler-Kirbach E Stoeckmann E Heidemann P Gehlhar K Holland-Letz T Braun W Efficacy of sublingual swallow immunotherapy in children with severe grass pollen allergic symptoms: a double-blind placebo-controlled study Allergy 2004 59 498 504 15080830 10.1111/j.1398-9995.2004.00457.x
Bel EH Clinical phenotypes of asthma Curr Opin Pulm Med 2004 10 44 50 14749605 10.1097/00063198-200401000-00008
De Marco R Locatelli F Cerveri I Bugiani M Marinoni A Giammanco G Incidence and remission of asthma: a retrospective study on the natural history of asthma in Italy J Allergy Clin Immunol 2002 110 228 235 12170262 10.1067/mai.2002.125600
Miller SA Dykes DD Polesky HF A simple salting out procedure for extracting DNA from human nucleated cells Nucleic Acids Res 1988 16 1215 3344216
The Vanderbilt Medical Center Power and Sample Size Calculation
Newton-Cheh C Hirschhorn JN Genetic association studies of complex traits: design and analysis issues Mutat Res 2005 573 54 69 15829237
Innate Immunity in Heart, Lung and Blood disease
Baldini M Lohman IC Halonen M Erickson RP Holt PG Martinez FD A Polymorphism* in the 5' flanking region of the CD14 gene is associated with circulating soluble CD14 levels and with total serum immunoglobulin E Am J Respir Cell Mol Biol 1999 20 976 983 10226067
Sengler C Haider A Sommerfeld C Lau S Baldini M Martinez F Wahn U Nickel R Evaluation of the CD14 C-159 T polymorphism in the German Multicenter Allergy Study cohort Clin Exp Allergy 2003 33 166 169 12580907 10.1046/j.1365-2222.2003.01549.x
Ober C Tsalenko A Parry R Cox NJ A second-generation genomewide screen for asthma-susceptibility alleles in a founder population Am J Hum Genet 2000 67 1154 1162 11022011
Vercelli D Learning from discrepancies: CD14 polymorphisms, atopy and the endotoxin switch Clin Exp Allergy 2003 33 153 155 12580903 10.1046/j.1365-2222.2003.01606.x
Arbour NC Lorenz E Schutte BC Zabner J Kline JN Jones M Frees K Watt JL Schwartz DA TLR4 mutations are associated with endotoxin hyporesponsiveness in humans Nat Genet 2000 25 187 191 10835634 10.1038/76048
Braun-Fahrlander C The role of the farm environment and animal contact for the development of asthma and allergies Clin Exp Allergy 2001 31 1799 1803 11737027 10.1046/j.1365-2222.2001.01269.x
Gehring U Bischof W Fahlbusch B Wichmann HE Heinrich J House dust endotoxin and allergic sensitization in children Am J Respir Crit Care Med 2002 166 939 944 12359650 10.1164/rccm.200203-256OC
Gehring U Bischof W Schlenvoigt G Richter K Fahlbusch B Wichmann HE Heinrich J Exposure to house dust endotoxin and allergic sensitization in adults Allergy 2004 59 946 952 15291902 10.1111/j.1398-9995.2004.00551.x
Raby BA Klimecki WT Laprise C Renaud Y Faith J Lemire M Greenwood C Weiland KM Lange C Palmer LJ Lazarus R Vercelli D Kwiatkowski DJ Silverman EK Martinez FD Hudson TJ Weiss ST Polymorphisms in toll-like receptor 4 are not associated with asthma or atopy-related phenotypes Am J Respir Crit Care Med 2002 166 1449 1456 12406828 10.1164/rccm.200207-634OC
Eder W Klimecki W Yu L von Mutius E Riedler J Braun-Fahrlander C Nowak D Martinez FD Toll-like receptor 2 as a major gene for asthma in children of European farmers J Allergy Clin Immunol 2004 113 482 488 15007351 10.1016/j.jaci.2003.12.374
|
16202135
|
PMC1262723
|
CC BY
|
2021-01-04 16:03:35
|
no
|
BMC Med Imaging. 2005 Oct 3; 5:6
|
latin-1
|
BMC Med Imaging
| 2,005 |
10.1186/1471-2342-5-6
|
oa_comm
|
==== Front
BMC MicrobiolBMC Microbiology1471-2180BioMed Central London 1471-2180-5-531620212410.1186/1471-2180-5-53Methodology ArticleHigh-throughput metal susceptibility testing of microbial biofilms Harrison Joe J [email protected] Raymond J [email protected] Howard [email protected] Department of Biological Sciences, University of Calgary, 2500 University Drive N.W., Calgary, Alberta, Canada T2N 1N42 Biofilm Research Group, University of Calgary, 2500 University Drive N.W., Calgary, Alberta, Canada T2N 1N42005 3 10 2005 5 53 53 26 2 2005 3 10 2005 Copyright © 2005 Harrison et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Microbial biofilms exist all over the natural world, a distribution that is paralleled by metal cations and oxyanions. Despite this reality, very few studies have examined how biofilms withstand exposure to these toxic compounds. This article describes a batch culture technique for biofilm and planktonic cell metal susceptibility testing using the MBEC assay. This device is compatible with standard 96-well microtiter plate technology. As part of this method, a two part, metal specific neutralization protocol is summarized. This procedure minimizes residual biological toxicity arising from the carry-over of metals from challenge to recovery media. Neutralization consists of treating cultures with a chemical compound known to react with or to chelate the metal. Treated cultures are plated onto rich agar to allow metal complexes to diffuse into the recovery medium while bacteria remain on top to recover. Two difficulties associated with metal susceptibility testing were the focus of two applications of this technique. First, assays were calibrated to allow comparisons of the susceptibility of different organisms to metals. Second, the effects of exposure time and growth medium composition on the susceptibility of E. coli JM109 biofilms to metals were investigated.
Results
This high-throughput method generated 96-statistically equivalent biofilms in a single device and thus allowed for comparative and combinatorial experiments of media, microbial strains, exposure times and metals. By adjusting growth conditions, it was possible to examine biofilms of different microorganisms that had similar cell densities. In one example, Pseudomonas aeruginosa ATCC 27853 was up to 80 times more resistant to heavy metalloid oxyanions than Escherichia coli TG1. Further, biofilms were up to 133 times more tolerant to tellurite (TeO32-) than corresponding planktonic cultures. Regardless of the growth medium, the tolerance of biofilm and planktonic cell E. coli JM109 to metals was time-dependent.
Conclusion
This method results in accurate, easily reproducible comparisons between the susceptibility of planktonic cells and biofilms to metals. Further, it was possible to make direct comparisons of the ability of different microbial strains to withstand metal toxicity. The data presented here also indicate that exposure time is an important variable in metal susceptibility testing of bacteria.
==== Body
Background
Determination of the minimum inhibitory concentration (MIC), based on antimicrobial activity against planktonic organisms, is the standard assay for susceptibility testing. Biofilms, which present with distinct physiology compared to planktonic cells, are infamous for their ability to withstand a wide range of antimicrobials, including metals [1-4]. Despite the ubiquitous distribution of metals and the predominance of microbial biofilms in the environment and in device-associated infections, very few studies have comparatively examined biofilm susceptibility to metals relative to planktonic cells. The scarcity of data in this regard may be attributable to the existing methods used to grow biofilms, which typically include contamination prone flow systems. Metal susceptibility testing also entails challenges not encountered with antibiotics. This includes complexation of metals with components of growth media, inorganic precipitation, reduction reactions, and carry-over of the metal to the recovery medium.
A recently developed, high-throughput approach to antibiotic and biocide susceptibility testing of microbial biofilms is the Calgary Biofilm Device [5,6] (commercially available as the MBEC-high throughput (HTP) assay, MBEC Bioproducts Inc., Edmonton, Alberta, Canada [7]). This batch culture method of biofilm and planktonic cell susceptibility testing provides three internally consistent, comparative measurements from a single experiment: 1) the planktonic minimum inhibitory concentration (MIC), 2) the planktonic minimum bactericidal concentration (MBC), and 3) the minimum biofilm eradication concentration (MBEC). The MBEC assay is not prone to leakage and contamination since it is manipulated in a Laminar flow hood.
The present study was rooted in two principle aims. The first aim was to develop a method of high-throughput metal susceptibility testing of biofilms using the MBEC assay. As part of this goal, a metal specific neutralizing regime was employed to reduce the biological toxicity of many different metal cations and oxyanions in vitro. This procedure allowed for comparisons between the susceptibility of planktonic cells and biofilms to metals (between different strains and/or microbial species), and provides a significant modification of the procedure originally reported by Ceri et al. for antibiotic susceptibility testing [5,6]. Also presented here is quality control data for the MBEC technique that has not been published elsewhere.
The second aim was to apply this method to examine variables that may influence measurements of metal susceptibility. A common dilemma in comparative studies of different bacterial strains is the ability of each strain to form biofilms. In simple terms, the ratio of bacterial cells (i.e. chemically reactive targets) to metal ions may influence the determination of susceptibility. To address this, biofilm growth of different bacterial species was calibrated to allow relative comparisons of susceptibility between biofilms with similar cell densities. Here, the relative differences in E. coli and P. aeruginosa biofilm susceptibility to the heavy metalloid oxyanions selenite (SeO32-) and tellurite (TeO32-) were examined. These compounds are highly toxic, water soluble pollutants that are spread into the environment in the form of industrial effluent [8,9]. P. aeruginosa was 80 times more resistant to heavy metalloids than E. coli.
Another potential obstacle to accurate measurements from metal susceptibility testing is exposure time. Studies of short duration metal exposure in minimal media have indicated that biofilms are highly tolerant to heavy metals [2,3], whereas long exposures in rich media resulted in complete elimination of the biofilm [3]. Logically, these differences may be based on differences in either exposure time or growth medium. Here, the effect of these variables on the measured tolerance of E. coli biofilms to metal cations was examined. Time-dependent trends in biofilm susceptibility to metals were identified that were not altered by changing the medium in which bacteria were grown or challenged. Exposure time is thus an important consideration in the design of studies centred on biofilm tolerance to metals.
Results
Biofilm growth
Biofilms of Pseudomonas aeruginosa ATCC 27853, E. coli TG1, and E. coli JM109 were grown to an overall mean density of 6.6 ± 0.5, 6.7 ± 0.3, and 6.4 ± 0.4 log10 cfu peg-1 (respectively) in Luria-Bertani medium enriched with vitamin B1 (LB + B1). This corresponded to 9.5 and 24 h of incubation at 35°C for P. aeruginosa and E. coli, respectively. When grown in minimal salts vitamins glucose (MSVG) medium for 24 h at 35°C, biofilms of E. coli JM109 reached a mean cell density of 5.02 ± 0.55 log10 cfu peg-1. Mean and standard deviation calculations for cell densities were based on pooled data from all growth controls performed (i.e. from 36 to 59 replicates each).
Mean viable cell counts and standard deviation (SD) for P. aeruginosa ATCC 27853 and E. coli TG1 on the pegs of each row of the MBEC™-HTP assay are presented in (Fig. 1a and 1c, respectively, 4 to 6 replicates each). The cell counts for each row were pooled and compared using one-way analysis of variance (ANOVA). The cell density of biofilms grown on the different rows of pegs in the MBEC™-HTP assay were statistically equivalent (p = 0.842 for P. aeruginosa; p = 0.274 for E. coli). E. coli JM109 also formed statistically equivalent biofilms across the different rows of pegs. This was similar to E. coli TG1, and thus the data is not presented here.
To allow for a valid comparison of susceptibility data, the biofilm cell counts for P. aeruginosa ATCC 27853 and E. coli TG1 were compared using a Mann-Whitney U-test. The biofilm cell density of these two strains were statistically equivalent (p = 0.209). Under the growth conditions reported, E. coli JM109 formed biofilms with significantly less cell density than the other 2 strains when compared using a Mann-Whitney U-test (p < 0.001).
Microscopy
Biofilms were examined in situ using scanning electron microscopy. Photomicrographs of P. aeruginosa ATCC 27853 and E. coli TG1 are presented in (Fig. 1b and 1d, respectively). SEM pictures show the growth of surface-adherent bacteria in thin layers and mounds on the pegs of the MBEC™ device. These layers were estimated to be up to 10 μm in height in some areas. Biofilms heterogeneously covered the plastic surface. The biofilm growth of E. coli JM109 was similar to strain TG1, except with a slightly more sparse distribution across the peg surface (data not shown).
Susceptibility of E. coli and P. aeruginosa to metalloid oxyanions
In this study and as an example, biofilms of Escherichia coli TG1 and Pseudomonas aeruginosa ATCC 27853 were assayed for susceptibility to selenite (SeO32-) and tellurite (TeO32-). These two organisms formed biofilms with statistically equivalent cell density when grown as described above. This allowed for a direct comparison between the two bacterial strains for relative levels of resistance to these heavy metalloid oxyanions.
For E. coli TG1 and P. aeruginosa ATCC 27853, the mean and standard deviation of MIC, MBC and MBEC values for SeO32- and TeO32- are summarized in Table 1. These calculations were based on 4 to 8 independent replicates each. With 4 h of exposure, biofilms were up to 133 times more tolerant to TeO32- than the corresponding planktonic cultures. With regards to planktonic cells (derived from the surface of the biofilms), P. aeruginosa was 80 times more resistant to tellurite than E. coli.
Time-dependent susceptibility of E. coli biofilms to metal cations
Exposure time may be of pivotal importance as a controlled variable in the design of metal susceptibility assays. To directly address this problem, an array of 11 metal cations was chosen to represent groups 7B to 4A of the periodic table. Using the MBEC assay, the susceptibility of E. coli JM109 to these compounds was tested. E. coli JM109 has been a popular model microorganism for studies of metal resistance in bacteria. Biofilm and planktonic cell susceptibility at 2 and 24 h of exposure in both rich medium (LB + B1) and minimal medium (MSVG) was examined. A larger data set with a greater number of trials at 24 h of exposure in LB + B1 was used in this study to expand on an original report [10]. All other test conditions were previously unexamined. This data is presented in Tables 2 and 3, respectively.
In general, MIC, MBC and MBEC values were greater in rich medium than in minimal medium. This was probably due to the chelation of metal ions by phosphates and organic matter present in the growth medium. However, there were two trends that were invariant with regards to nutrient status: 1) In LB + B1, biofilms were 1.3 to 24 times more tolerant to metal cations than the corresponding planktonic cells when exposed for 2 h. Similarly, E. coli biofilms grown and tested in MSVG were 2.0 to 29 times more tolerant to metal cations with a similar exposure time. In combination with the susceptibility data for SeO32- and TeO32-, this would suggest that with short exposures biofilms are highly tolerant to metal toxicity. 2) By 24 h of exposure, biofilm and planktonic cultures were eradicated at similar concentrations of metal cations in almost every instance. This occurred independently of the growth medium used for bacterial susceptibility testing.
Collectively, these data suggest that biofilm tolerance to metals is time-dependent. We note that (in general) MIC values did not change with exposure time using this method (data not shown).
Discussion
This manuscript describes a high-throughput method for metal susceptibility testing of biofilms using the MBEC assay. This technique has the advantage that metal susceptibility testing may be done using a combinatorial approach. It is possible to employ different mean numbers of bacterial cells, alternate exposure times, various and diverse growth media formulations, as well as a broad range of metals (as well as other antimicrobial compounds). Many bacteria and yeasts, including Staphylococcus aureus [6,10], Mycobacterium spp. [11], Candida spp. (J.J. Harrison, H. Ceri and R.J. Turner, unpublished data), and Burholderia cepacia complex [12] are amenable to biofilm growth using the MBEC assay. Biofilms on the pegs of the MBEC device may be examined in situ using scanning electron microscopy, epifluorescent microscopy [13], and confocal laser-scanning microscopy (CLSM) [14].
As a quality control, a test for equivalent biofilm formation between rows of pegs in the MBEC device was performed. Biofilm cell density was statistically equivalent across the different rows of pegs for E. coli TG1, JM109 and P. aeruginosa ATCC 27853. This quality control needs to be a routine test for studies of biofilm susceptibility using this method. It is important to note that not all bacteria (particularly mucoidal isolates) are amenable to this method of growth (J.J. Harrison, H. Ceri and C. Stremick, unpublished data). If statistically equivalent biofilms are not generated using the MBEC trough format, it is possible to instead place the peg lid in a microtiter plate containing ~150 μl of inoculum in each well. This alternative system may be placed on a gyrorotary shaker to facilitate biofilm formation on the peg lid. In general, biofilms formed using a trough have a 5- to 10-fold greater cell density than those formed using the microtiter plate format (J.J. Harrison, H. Ceri and C. Stremick, unpublished data).
This procedure may be modified to discern biofilm and planktonic viable cell counts. This is included as an amendment with the online supplementary material. Viable cell counting is accomplished by serially diluting aliquots from the recovery and neutralization plates, which are subsequently plated onto agar and enumerated. A characteristic of the MBEC assay is that planktonic cultures are seeded by cells shed from the biofilm. Log-killing of biofilms may be calculated using this method, but as a consequence, log-killing of planktonic cells may not. However, this model does reflect the natural duality of the bacterial life cycle where a recalcitrant nidus of biofilm cells may survive metal exposure to shed planktonic cells back into the surroundings. Antibiotic MIC values obtained from this method are in most cases equivalent to MIC data derived from National Committee of Clinical Laboratory Standards (NCCLS) standard procedures [6]. In this report, MIC and MBC values for TeO32- were consistent with literature values obtained using alternate methods [9,15].
The use of neutralizing agents in metal susceptibility testing allows for more accurate comparisons between the susceptibility of planktonic cells and biofilms to metals. Metal carry over from challenge to recovery media can affect the determination of viable cell counts and bactericidal concentrations [16]. Pertinent to this method, carry over is of particular concern when comparing biofilm and planktonic cell MBEC and MBC values. For example, planktonic cells removed in media containing metals will be inhibited to a different extent during recovery than biofilms removed from the media containing metals. In this protocol, biofilms and planktonic cultures were treated with equal amounts of a neutralizing agent at a concentration with limited toxicity to the bacterial cells. Biofilm and planktonic cells were equally susceptible to metals with long exposure times, suggesting that the neutralizing agents are equivalently effective for the treatment of both bacterial forms. The limitation of chemical neutralization will be acknowledged here. Treating metal exposed biofilms with a chelator increases the number of viable cells recovered from killing kinetics experiments relative to identical experiments that do not use a neutralizing agent (J.J. Harrison, H. Ceri and R.J. Turner, unpublished data). However, it is very likely that the metal-chelator complex is toxic to bacteria [3]. The bottom line of this procedure is that the metal-chelator is less toxic than the free metal ion.
Luria-Bertani medium is appropriate for susceptibility testing of metal oxyanions, as these compounds remain soluble in this medium. In the case of metal cations, precipitation and/or complexation of metals occurs rapidly in rich media. In this instance, minimal media preparations may be more suitable (an excellent array of minimal media preparations have been designed by Teitzel and Parsek [2]). Although the absolute values of MIC, MBC and MBEC determinations were greater in rich than in minimal medium, the tolerance of biofilms to metal toxicity remained time-dependent in both growth conditions.
Reports in the literature have suggested that biofilm tolerance to metals may be time-dependent [3,10,17]. Biofilms of P. aeruginosa are up to 600 times more tolerant to heavy metals than planktonic cells (in minimal media with 2 to 5 h exposure) [2,3]. Using the MBEC assay, it has been noted for P. aeruginosa, Staphylococcus aureus, and E. coli that biofilms and planktonic cells are equally susceptible to metal cations (in rich media with 24 h exposure) [10]. Similar time-dependent phenomena have been previously described for the killing of E. coli JM109 and P. aeruginosa ATCC 27853 biofilms by SeO32- and TeO32- [10,18]. Here this has been revaluated and this study demonstrates that the foundation of this is exposure time. Thus, the amount of time bacteria are exposed to toxic metals is an important controlled variable that contributes to measurements of bacterial susceptibility to metals. This is important both in terms of comparisons of biofilm and planktonic cell tolerance to metals as well as in evaluation of data in the literature.
Conclusion
This high-throughput method is currently being used to elucidate a multifactorial model of metal tolerance in the bacterial biofilm [3,10,14,18,19]. The principle strength of this assay lies in the ability to rapidly screen changes in biofilm susceptibility to metals using diverse permutations of growth conditions, exposure times, and metal compounds. This type of combinatorial experimental approach would not be pragmatic using other methods.
The growth of different microorganisms may be calibrated to compare biofilms of equivalent cell density. This means that it is possible to compare susceptibility data of different strains without the concern that differences in resistance and/or tolerance are due to inequalities in growth of the biofilm populations. Lastly, exposure time influences the observed tolerance of biofilms to metals. Exposure time is thus a key consideration in the design of metal susceptibility assays.
A step-by-step protocol for this method is freely available from the authors [20].
Methods
Bacterial strains and media
Escherichia coli TG1, E. coli JM109 and Pseudomonas aeruginosa ATCC 27853 were stored at -70°C in Cryobanks™ (Prolab Diagnostics) according to the manufacturer's instructions. Bacteria were grown in Luria-Bertani media (pH 7.1, Difco) enriched with 0.001% vitamin B1 (LB + B1) or in the minimal salts vitamins glucose (MSVG, pH 7.1) of Teitzel and Parsek [2]. MSVG was specifically designed for metal susceptibility testing and minimizes the precipitation of metal cations in the growth medium. Sub-cultures and spot plates (see below) were grown on LB + B1 with 1.5% w/v granulated agar. All serial dilutions were performed using 0.9% saline.
Biofilm cultivation
The MBEC assay has two parts: 1) a plastic lid with 96 pegs that fits into 2) a corrugated trough. Biofilms were grown in the MBEC assay according to the method of Ceri and colleagues [5,6], and bacterial cultures and inocula were prepared according to the steps illustrated in Fig. 2(a–d) and as described here. Frozen stocks of bacteria were streaked out on Luria-Bertani agar to obtain a first-subculture. A single colony was picked from the first-subculture and again streaked out on agar obtain a second sub-culture. Colonies were collected from second-subcultures using a sterile cotton swab and suspended in broth medium to a 1.0 McFarland Standard. This suspension was diluted 30-fold in broth, and 22 ml of the 1 in 30 dilution was used to inoculate the MBEC assay. For demonstration purposes in this manuscript, we examined E. coli TG1, E. coli JM109 or P. aeruginosa ATCC 27853. Here, 22 ml of inoculum contained ~107 cfu/ml bacteria and this was transferred into the troughs. Starting bacterial number in the inocula were verified by viable cell counting. The inoculated devices were placed on a rocking table (Bellco Biotechnology) in an incubator at 35°C and 95% relative humidity at 2.5 rocks per minute. The shear force of the rocking motion facilitated the formation of 96 equivalent biofilms on the pegs.
Incubation times in LB + B1 medium were calibrated according to the growth rates of P. aeruginosa ATCC 27853 and E. coli TG1 (i.e. so that they would produce biofilms with an equivalent number of cells). Thus, E. coli and P. aeruginosa were grown for 24 and 9.5 h, respectively. For every MBEC assay, four pegs were broken from the lid (after it had been rinsed, see below), then 'sonciated' in sterile 0.9% saline. This was performed using an Aquasonic water-table sonicator (VWR International, model 250HT) for 5 minutes on the setting 'high' as previously described [5,6]. The disrupted biofilms were serially diluted and plated for viable cell counting. This growth control was used to verify that the appropriate number of bacteria had formed in the biofilm.
A test for equivalent biofilm growth on the rows of pegs in the MBEC assay was performed. This was accomplished by growing biofilms to the desired cell density, rinsing the pegs (i.e. by placing the peg lid into a microtiter plate with 200 μl of 0.9% saline in each well, termed a 'rinse plate'), then by sonicating (as described above) the biofilms into sterile saline (i.e. a fresh rinse plate). The disrupted biofilms were serially diluted then plated onto agar to determine viable cell counts (i.e. cfu peg-1). Plates were incubated for 24 h at 35°C then enumerated. Raw data were first log10-transformed and the mean viable cell counts for the different rows of pegs were compared using one-way analysis of variance (ANOVA).
Stock metal solutions
Silver nitrate (AgNO3), aluminum sulfate (Al2(SO4)3·18H2O), zinc sulfate (ZnSO4·7H2O), stannous chloride (SnCl2·2H2O) and copper sulfate (CuSO4·5H2O) were obtained from Fisher Scientific Company of Fairlawn, NJ. Nickel sulfate (NiSO4·6H2O), mercuric chloride (HgCl2), cobalt chloride (CoCl2·6H2O), lead nitrate (Pb(NO3)2), potassium tellurite (K2TeO3) and sodium selenite (Na2SeO3) were obtained from Sigma Chemical Company of St Louis, MO. Cadmium chloride (CdCl2·5/2H2O) was purchased from Terochem Laboratories of Edmonton, AB, and manganous sulfate (MnSO4H2O) from BDH of Toronto, ON. Reagent grade metal and metalloid compounds were purchased for the purposes of this study to minimize the potential influence of contaminating, residual metals.
All stock metal solutions, with the exception of Sn2+, were made up in double-distilled water at 5 times the highest concentration desired in the challenge plates. These stock solutions were passed through a 0.22 μm syringe filter into sterile glass vials and stored at room temperature. Sn2+ was disolved in 50% ethanol and stored in a sterile polypropylene tube.
Metal susceptibility testing
Susceptibility testing and exposure of biofilms to metals is summarized in Fig. 2(e–l). The peg lid of the MBEC assay fits inside a standard 96-well microtiter plate. This assay may standardized to any standard brand of microplate. The volume of the challenge media in the microplate must be sufficient to submerge the peg past the height of the biofilm produced in the device. Here, we used flat-bottom 96-well microtiter plates (Nunc) with a final volume of 200 μl. Serial two-fold dilutions of metalloid oxyanions were made in LB + B1 broth along the length of these microtiter plates (termed the 'challenge plate'), allowing the first well of each row to serve as a sterility control, and the last well to serve as a growth control. The challenge plates were incubated for the desired exposure time at 35°C and 95% relative humidity. After exposure, the peg lid was removed and rinsed twice with 0.9% saline and the biofilm disrupted by sonciation into LB + B1 broth containing the appropriate neutralizing agent (termed the 'recovery plate', see below). After removal of the peg lid, the challenge plate was covered with a new, sterile lid to protect the planktonic cultures in the challenge plate wells. The planktonic cultures were also treated with the appropriate neutralizing agent (see below).
Aliquots of the neutralized biofilm and planktonic cultures were spot plated onto LB + B1 agar and incubated for 48 h at 35°C. The MIC was determined by reading the optical density of the challenge plate at 650 nm (OD650). MBC and MBEC values were determined by qualitatively scoring the spot plates for bacterial growth. With the exception of Cu2+ and Ni2+ assays, MBEC values were redundantly determined by reading the OD650 of the recovery plates after 48 h.
Neutralizing agents
To differentiate between the bacteriostatic and bactericidal actions of the tested compounds, a two-step neutralizing protocol was designed to reduce the toxicity of residual metals. First, metal cations and oxyanions were treated with a chemical known to chelate or to react with the tested compound. The neutralizing agents currently and previously used in our laboratories have been summarized in Table 4. Second, neutralized cultures were plated onto a rich agar medium. Ideally, this latter step facilitates additional complexation of metals with components of the medium (such as phosphates, sulfates, and amines), and also allows diffusion of the metals into the agar. Thus, exposed bacteria are left on top of the agar medium to recover where there is a reduced concentration of biologically available metal.
As examples encountered in this study, metalloid oxyanions were reacted with 5 mM reduced glutathione (GSH, Sigma Chemical). GSH is used by the bacterial cell as a reduction-oxidation buffer to reductively eliminate a diverse array of inorganic oxidants [21,22], including selenite [23] and tellurite [24], and this is the basis for its use as a neutralizing agent here. Similarily, GSH was used to counter the effects of Zn2+, Co2+, Pb2+, Hg2+, and Cd2+ toxicity. Many metals are postulated to exert toxicity through oxidative stress on the thiol groups of proteins [25,26] and thus addition of GSH or L-cysteine may partially counteract this mechanism [16]. Sodium diethyldithiocarbate (Na2DDTC) was used to chelate Ni2+ and Cu2+, which rapidly formed metal precipitates with this organic chelator [27]. Citrate was used to coordinate Ag+ [10]. Tin was complexed using the amino acid glycine [28]. We have previously reported that 5-sulfosalicylic acid may be used as a chelator of Al3+ and Mn2+ [10]. As an improvement to this technique, we suggest that crushed asprin may be used as an alternative. Salicylate derivatives can be toxic to bacteria. The less soluble acetylated form may be employed with a wider range of bacterial strains (J.J. Harrison, H. Ceri, and R.J. Turner, unpublished data).
Stock solutions of citrate (0.5 M, Sigma), Na2DDTC (0.25 M, ICN), glutathione (0.25 M, Sigma), acetylsalicylic acid (~0.01 M, West-Can Pharmaceuticals, available at Calgary Coop), glycine (0.25 M, Bio-Rad Laboratories), and L-cysteine (0.25 M, Sigma) were prepared in double-distilled water and sterile filtered. With the exception of acetylsalicylic acid (ASA), all of these stocks were stored at -20°C until use. ASA was stored at room temperature. Neutralizing agents for biofilm cultures were added directly to LB+B1 broth used in the recovery plates. Neutralizing agents for the planktonic cultures were prepared at 5 times the desired neutralizing concentration in 0.9% saline. Aliquots (10 μl) of the diluted stock solutions were then added to the wells of a sterile 96-well plate (the neutralizing plate) to which 40 μl from each well of the challenge plate were added. The final concentration of neutralizing agent used to treat the planktonic cultures was thus equal to that used to treat biofilm cultures.
Scanning electron microscopy
Biofilms were examined using scanning electron microscopy (SEM) as previously described [4,6,10]. Briefly, pegs were broken from the lid of the MBEC device, rinsed once in 0.9% saline, then fixed for 16 h at 4°C in 5% glutaraldehyde dissolved in 0.1 M cacodylate buffer (pH 7.2). The next day, pegs were rinsed with 0.1 M cacodylate buffer, dehydrated with 95% ethanol, and then air dried for 30 h before mounting. SEM was performed using a Hitachi model 450 scanning electron microscopy according to the method of Morck and colleagues [29].
Statistical tests
Mean, standard deviation and one-way analysis of variance (ANOVA) calculations were performed using GraphPad InStat 3.0 (GraphPad Software Inc., San Diego, CA).
List of abbreviations
CLSM = confocal laser-scanning microscopy, MBC = minimum bactericidal concentration, MBEC = minimum biofilm eradication concentration, MIC = minimum inhibitory concentration, Na2DDTC = sodium diethyldithiocarbamate
Authors' contributions
JJH carried out the studies of metal susceptibility, designed the experimental protocols, performed statistical analysis of the data, and drafted the manuscript. HC and RJT conceived of the experiments, participated in their design, and helped to draft the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
HTP Metal Testing Protocol.doc - A step-by-step protocol. This is a copy of the step-by-step protocol to be made available free of charge on the world wide web through the authors' web site or BMC Microbiology whichever is preferred.
Click here for file
Acknowledgements
This research has been funded by operating grants from the Natural Sciences and Engineering Research Council of Canada (NSERC) to Drs. Raymond J. Turner and Howard Ceri. NSERC has also provided an Industrial Postgraduate Scholarship (IPS1) to Joe Harrison. Thanks to Carol A. Stremick, Nicole J. Roper and Erin A. Badry for expert technical assistance. Dr. Howard Ceri is Director of Business Development for MBEC BioProducts Inc. No corporate funding was provided for the research described in this manuscript.
Figures and Tables
Figure 1 Growth of bacterial biofilms in the MBEC assay. (A) Mean cell density of Pseudomonas aeruginosa ATCC 27853 biofilms on the pegs in different rows of the MBEC assay. Each value is expressed as the mean and standard deviation of 4 to 6 trials. There is no significant difference between cell density of biofilms in the different rows (p = 0.842 using one-way ANOVA). (B) SEM photomicrograph of a P. aeruginosa biofilm on the peg surface. (C) Mean cell density of Escherichia coli TG1 on the pegs in different rows of the MBEC assay. Each value is expressed as the mean and standard deviation of 4 to 6 trials. There is no significant difference between cell density of biofilms in the different rows (p = 0.274 using one-way ANOVA). (D) SEM photomicrograph of an E. coli biofilm on the peg surface. The bar represents 5 μm.
Figure 2 An overview of the high-throughput protocol for metal susceptibility testing using the MBEC assay. (A) Frozen stocks of bacteria were streaked out on the appropriate agar medium to obtain a first- and a subsequent second-subculture. (B) Colonies were collected from second-subcultures and suspended in broth medium to a 1.0 McFarland Standard. (C) This suspension was diluted 30-fold in broth, and the 1 in 30 dilution was used to inoculate the MBEC assay. (D) The inoculated device was placed on a rocking table in an incubator. (E) Serial dilutions of metal cations and oxyanions were set up along length of a microtiter plate along (the challenge plate). (F) The biofilms were rinsed to remove loosely adherent planktonic bacteria. (G) The first peg from each row was removed. These pegs were used to verify growth of the biofilms on the pegs. The peg lid was then inserted into the challenge plate. (H) During exposure, metals diffuse into the biofilm while planktonic cells are shed from the surface of the biofilm. Sloughed cells serve as the inoculum for planktonic MIC and MBC determinations. (I) The exposed biofilms were rinsed twice and the peg lid was inserted into fresh recovery medium containing the appropriate neutralizing agent (the recovery plate). The biofilms were disrupted into the recovery medium by sonciation on a water table sonicator. (J) Aliquots of planktonic cultures were transferred from the challenge plate to a microtiter plate containing the appropriate neutralizing agents (the neutralizing plate). (K) An aliquot from the recovery and neutralizing plates were spotted onto rich agar media. (L) MIC values are determined by reading the optical density at 650 nm (OD650) of the challenge plate after the desired period of incubation using a microtiter plate reader. Spot plates were qualitatively scored for growth to obtain MBC and MBEC values. MBEC values were redundantly determined by determining the A650 of the recovery plates after incubation.
Table 1 Comparative susceptibility of bacterial biofilms to metalloid oxyanions with 4 hours of exposure
Strain Metal MIC (mM) MBC (mM) MBEC (mM) Fold tolerance1
E. coli TG1 SeO32- > 5.8 > 5.8 > 5.8 na
TeO32- 0.02 ± 0.01 0.03 ± 0.02 > 2.0 ≥ 133
P. aeruginosa ATCC 27853 SeO32- > 187 > 187 > 187 na
TeO32- 1.6 ± 0.8 6.6 ± 3.0 > 16 2.4
na indicates a measurement that is not applicable
1the fold tolerance, given the sensitivity of the assay on a log2 scale, is equal to the ratio of the means of MBEC:MBC
Table 2 Susceptibility of Escherichia coli JM109 to metal cations with 2 or 24 h of exposure in rich (LB + B1) medium
Periodic group Metal ion Exposure time (h) MIC (mM) MBC (mM) MBEC (mM) Fold Tolerance1
7B Mn2+ 2 37 ± 0 > 149 > 149 na
24 198 ± 86 223 ± 86 1.1
8B Ni2+ 2 7.5 ± 2.1 > 140 >140 na
24 17 ± 0 29 ± 11 1.7
1B Cu2+ 2 4.5 ± 1.4 16 ± 0 16 ± 0 1.0
24 16 ± 0 16 ± 0 1.0
Ag+ 2 0.06 ± 0.02 0.06 ± 0.02 1.6 ± 0.7 24
24 0.06 ± 0.02 0.04 ± 0 0.7
2B Zn2+ 2 4.2 ± 2.4 125 ± 0 > 125 ≥ 2.0
24 31 ± 0 31 ± 0 1.0
Cd2+ 2 1.1 ± 0.2 55 ± 21 73 ± 0 1.3
24 2.3 ± 0 3.0 ± 1.3 1.3
Hg2+ 2 0.04 ± 0.02 0.14 ± 0.04 0.32 ± 0.22 2.3
24 0.04 ± 0 0.04 ± 0 1.0
3A Al3+ 2 nd > 304 > 304 na
24 19 ± 0 19 ± 0 1.0
4A Sn2+ 2 nd > 17 > 17 na
24 17 ± 0 17 ± 0 1.0
A portion of the data in this table (at the 24 h timepoint) represents a greater number of trials and reanalysis of the data originally reported in an earlier study [10].
na indicates a calculation that is not applicable
nd indicates an MIC that could not be determined due to metal precipitation
bold indicates the fold tolerance at 24 h of exposure
1the fold tolerance, given the sensitivity of the assay on a log2 scale, is equal to the ratio of the means of MBEC:MBC100
Table 3 Susceptibility of Escherichia coli JM109 to metal cations with 2 or 24 h of exposure in minimal (MSVG) medium
Periodic group Metal ion Exposure time (h) MIC (mM) MBC (mM) MBEC (mM) Fold Tolerance1
7B Mn2+ 2 33 ± 9 > 149 > 149 na
24 93 ± 37 84 ± 47 0.90
8B Co2+ 2 ≤ 0.27 26 ± 10 > 139 ≥ 11
24 1.0 ± 3 1.0 ± 3 1.0
Ni2+ 2 ≤ 0.27 > 139 > 139 na
24 0.48 ± 0.13 0.54 ± 0.39 1.1
1B Cu2+ 2 ≤ 0.25 ≤ 0.25 ≤ 0.25 na
24 ≤ 0.25 ≤ 0.25 na
Ag+ 2 ≤ 0.04 ≤ 0.04 0.59 ± 0 ≥ 15
24 ≤ 0.04 ≤ 0.04 1.0
2B Zn2+ 2 ≤ 0.25 11 ± 6 > 125 ≥ 23
24 0.55 ± 0.31 0.49 ± 0.35 0.89
Cd2+ 2 ≤ 0.14 5.1 ± 2.9 > 73 ≥ 29
24 0.49 ± 0.14 1.4 ± 1.1 2.9
Hg2+ 2 ≤ 0.02 ≤ 0.02 0.05 ± 0.02 ≥ 2.5
24 ≤ 0.02 ≤ 0.02 1.0
3A Al3+ 2 nd 42 ± 40 > 304 ≥ 14
24 1.5 ± 0.6 4.0 ± 1.4 2.7
4A Sn2+ 2 2.2 ± 0 8.6 ± 0 8.6 ± 0 1.0
24 4.3 ± 0 7.5 ± 2.2 1.7
Pb2+ 2 nd 20 ± 0 40 ± 0 2.0
24 4.9 ± 0 1.2 ± 0 0.25
na indicates a measurement that is not applicable
bold indicates the fold tolerance at 24 h of exposure
1the fold tolerance, given the sensitivity of the assay on a log2 scale, is equal to the ratio of the means of MBEC:MBC100
Table 4 Potential neutralizing agents for the microbiological application of inactivating metals cations and oxyanions*
Metal(loid) Neutralizing agent Maximum concentration1 Reference(s)
All oxyanions2 Glutathione 10 mM [10, 18, 24]
Al3+, Mn2+ Crushed acetylsalicylic acid (ASA) ~1–2 mM3 [30, 31]
Hg2+, Cd2+ Glutathione 10 mM [10]
L-cysteine 10 mM [16]
Cu2+, Ni2+ Diethyldithiocarbamic acid3 (DDTC) 2.5 mM (E. coli) up to 5 mM (P. aeruginosa) [27]
Sn2+ Glycine 10 mM [28]
Ag+ Sodium citrate 10 mM [10]
Zn2+, Co2+, Pb2+ Glutathione 10 mM [3]
*This is the first part of a two-part strategy to reduce the in vitro toxicity of metals (see text for details)
1 The maximum concentration tested and employed in studies by our laboratories using the high-throughput metal susceptibility testing method presented in this paper.
2 Tested heavy metal and metalloid oxyanions from our laboratories include TeO32-, TeO42-, SeO32-, CrO42-, AsO43-, AsO2-, WO42- and MoO42-.
3 Application is limited by the low solubility of salicylic acid and its acetylated derivatives in water.
4 The maximum concentration listed is inhibitory to bacterial growth in broth culture. Recovery broth media must be spot plated onto agar to allow bacterial growth and determination of accurate MBC and MBEC values.
==== Refs
Hall-Stoodley L Costerton JW Stoodley P Bacterial biofilms: From the natural environment to infectious diseases Nature Reviews Microbiology 2004 2 95 108 15040259 10.1038/nrmicro821
Teitzel GM Parsek MR Heavy metal resistance of biofilm and planktonic Pseudomonas aeruginosa Applied and Environmental Microbiology 2003 69 2313 2320 12676715 10.1128/AEM.69.4.2313-2320.2003
Harrison JJ Turner RJ Ceri H Persister cells, the biofilm matrix and tolerance to metal cations in biofilm and planktonic Pseudomonas aeruginosa Environmental Microbiology 2005 7 981 994 15946294 10.1111/j.1462-2920.2005.00777.x
Olson ME Ceri H Morck DW Buret AG Read RR Biofilm bacteria: formation and comparative susceptibility to antibiotics Canadian Journal of Veterinary Research 2002 66 86 92 11989739
Ceri H Olson ME Morck DW Storey D Read RR Buret AG Olson B The MBEC assay system: Multiple equivalent biofilms for antibiotic and biocide susceptibility testing Methods in Enzymology 2001 337 377 384 11398443
Ceri H Olson ME Stremick C Read RR Morck DW Buret AG The Calgary Biofilm Device: New technology for rapid determination of antibiotic susceptibilities in bacterial biofilms Journal of Clinical Microbiology 1999 37 1771 1776 10325322
MBEC Bioproducts Incorporated
Chasteen TG Bentley R Biomethylation of selenium and tellurium: microorganisms and plants Chemical Reviews 2003 103 1 25 12517179 10.1021/cr010210+
Turner RJ Tellurite toxicity and resistance in Gram-negative bacteria Recent Research Developments in Microbiology 2001 5 69 77
Harrison JJ Ceri H Stremick C Turner RJ Biofilm susceptibility to metal toxicity Environmental Microbiology 2004 6 1220 1227 15560820 10.1111/j.1462-2920.2004.00656.x
Bardouniotis E Ceri H Olson ME Biofilm formation and biocide susceptibility testing of Mycobacterium fortuitum and Mycobacterium marium Current Microbiology 2003 46 28 32 12432460 10.1007/s00284-002-3796-4
Tomlin KL Coll OP Ceri H Interspecies biofilms of Pseudomonas aeruginosa and Burkholderia cepacia Canadian Journal of Microbiology 2001 47 949 954 11718549 10.1139/cjm-47-10-949
Tomlin KL Clark SR Ceri H Green and red fluorescent protein vectors for use in biofilm studies of the intrinsically resistant Burkholderia cepacia complex Journal of Microbiological Methods 2004 57 95 106 15003692 10.1016/j.mimet.2003.12.007
Harrison JJ Ceri H Badry EA Roper NJ Tomlin KL Turner RJ Effects of the twin-arginine translocase on the structure and antimicrobial susceptibility of Escherichia coli biofilms Canadian Journal of Microbiology 2005 submitted work
Taylor DE Bacterial tellurite resistance Trends in Microbiology 1999 7 111 115 10203839 10.1016/S0966-842X(99)01454-7
Russel AD Ahonkhai I Rogers DT Microbiological applications of the inactivation of antibiotics and other antimicrobial agents Journal of Applied Bacteriology 1979 46 207 245 378921
Cortizo MC Lorenzo de Mele MF Cytotoxicity of copper ions released from metal: variation with the exposure period and concentration gradients Biological Trace Element Research 2004 102 129 142 15621934 10.1385/BTER:102:1-3:129
Harrison JJ Ceri H Stremick C Turner RJ Differences in biofilm and planktonic cell mediated reduction of metalloid oxyanions FEMS Microbiology Letters 2004 235 357 362 15183885 10.1016/j.femsle.2004.05.005
Harrison JJ Turner RJ Ceri H Metal tolerance in bacterial biofilms Chapter 2: Recent Research Developments in Microbiology 2005 9 33 55
University of Calgary, Personal Web Page of Raymond J. Turner
Aslund F Zheng M Beckwith J Storlz G Regulation of the OxyR transcription factor by hydrogen peroxide and the cellular thiol-disulfide status Proceedings of the National Academy of Sciences of the United States of America 1999 96 6161 6165 10339558 10.1073/pnas.96.11.6161
Ding H Demple B Thiol-mediated disassembly and reassembly of [2Fe-2S] clusters in the redox-regulated transcription factor SoxR Biochemistry 1998 37 17280 17286 9860842 10.1021/bi980532g
Turner RJ Weiner J Taylor DE Selenium metabolism in Escherichia coli BioMetals 1998 11 223 227 9850565 10.1023/A:1009290213301
Turner RJ Aharonowitz Y Weiner J Taylor DE Glutathione is a target of tellurite toxicity and is protected by tellurite resistance determinants in Escherichia coli Canadian Journal of Microbiology 2001 47 33 40 15049447 10.1139/cjm-47-1-33
Stohs SJ Bagchi D Oxidative mechanisms in the toxicity of metal ions Free Radical Biology and Medicine 1995 18 321 336 7744317 10.1016/0891-5849(94)00159-H
Turner RJ Weiner J Taylor DE Tellurite mediated thiol oxidation in Escherichia coli Microbiology 1999 145 2549 2557 10517608
Gottofrey J Borg K Jasmin S Effect of postassium ethylxanthate and sodium diethyldithiocarbamate on the accumulation and deposition of nickel in the brown trout (Salmo trutta) Pharmacology and Toxicology 1988 63 46 51 2840649
Diurdjevic P Djokic D Protein interactions with bivalent tin. 1. Hydrolysis and complexation of tin(II) with glycine Journal of Inorganic Biochemistry 1996 62 17 29 8936420 10.1016/0162-0134(95)00085-2
Morck DW Lam K McKay SG Olson ME Prosser B Ellis BD Cleeland R Costerton JW Comparative evaluation of fleroxacin, ampicillin, trimethoprim-sulfamethoxazole, and gentamicin as treatments of catheter associated urinary tract infections in a rabbit model. American Journal of Medicine 1994 94 23S 30S
Missy P Lanhers MC Grignon Y Joyeux M Burnel D In vitro and in vivo studies on chelation of manganese Human and Experimental Toxicology 2000 19 448 456 11125715 10.1191/096032700682694260
Graff L Muller G Burnel D In vitro and in vivo evaluation of potential aluminum chelators Veterinary and Human Toxicology 1995 88 271 292
|
16202124
|
PMC1262724
|
CC BY
|
2021-01-04 16:03:40
|
no
|
BMC Microbiol. 2005 Oct 3; 5:53
|
utf-8
|
BMC Microbiol
| 2,005 |
10.1186/1471-2180-5-53
|
oa_comm
|
==== Front
BMC MicrobiolBMC Microbiology1471-2180BioMed Central London 1471-2180-5-531620212410.1186/1471-2180-5-53Methodology ArticleHigh-throughput metal susceptibility testing of microbial biofilms Harrison Joe J [email protected] Raymond J [email protected] Howard [email protected] Department of Biological Sciences, University of Calgary, 2500 University Drive N.W., Calgary, Alberta, Canada T2N 1N42 Biofilm Research Group, University of Calgary, 2500 University Drive N.W., Calgary, Alberta, Canada T2N 1N42005 3 10 2005 5 53 53 26 2 2005 3 10 2005 Copyright © 2005 Harrison et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Microbial biofilms exist all over the natural world, a distribution that is paralleled by metal cations and oxyanions. Despite this reality, very few studies have examined how biofilms withstand exposure to these toxic compounds. This article describes a batch culture technique for biofilm and planktonic cell metal susceptibility testing using the MBEC assay. This device is compatible with standard 96-well microtiter plate technology. As part of this method, a two part, metal specific neutralization protocol is summarized. This procedure minimizes residual biological toxicity arising from the carry-over of metals from challenge to recovery media. Neutralization consists of treating cultures with a chemical compound known to react with or to chelate the metal. Treated cultures are plated onto rich agar to allow metal complexes to diffuse into the recovery medium while bacteria remain on top to recover. Two difficulties associated with metal susceptibility testing were the focus of two applications of this technique. First, assays were calibrated to allow comparisons of the susceptibility of different organisms to metals. Second, the effects of exposure time and growth medium composition on the susceptibility of E. coli JM109 biofilms to metals were investigated.
Results
This high-throughput method generated 96-statistically equivalent biofilms in a single device and thus allowed for comparative and combinatorial experiments of media, microbial strains, exposure times and metals. By adjusting growth conditions, it was possible to examine biofilms of different microorganisms that had similar cell densities. In one example, Pseudomonas aeruginosa ATCC 27853 was up to 80 times more resistant to heavy metalloid oxyanions than Escherichia coli TG1. Further, biofilms were up to 133 times more tolerant to tellurite (TeO32-) than corresponding planktonic cultures. Regardless of the growth medium, the tolerance of biofilm and planktonic cell E. coli JM109 to metals was time-dependent.
Conclusion
This method results in accurate, easily reproducible comparisons between the susceptibility of planktonic cells and biofilms to metals. Further, it was possible to make direct comparisons of the ability of different microbial strains to withstand metal toxicity. The data presented here also indicate that exposure time is an important variable in metal susceptibility testing of bacteria.
==== Body
Background
Determination of the minimum inhibitory concentration (MIC), based on antimicrobial activity against planktonic organisms, is the standard assay for susceptibility testing. Biofilms, which present with distinct physiology compared to planktonic cells, are infamous for their ability to withstand a wide range of antimicrobials, including metals [1-4]. Despite the ubiquitous distribution of metals and the predominance of microbial biofilms in the environment and in device-associated infections, very few studies have comparatively examined biofilm susceptibility to metals relative to planktonic cells. The scarcity of data in this regard may be attributable to the existing methods used to grow biofilms, which typically include contamination prone flow systems. Metal susceptibility testing also entails challenges not encountered with antibiotics. This includes complexation of metals with components of growth media, inorganic precipitation, reduction reactions, and carry-over of the metal to the recovery medium.
A recently developed, high-throughput approach to antibiotic and biocide susceptibility testing of microbial biofilms is the Calgary Biofilm Device [5,6] (commercially available as the MBEC-high throughput (HTP) assay, MBEC Bioproducts Inc., Edmonton, Alberta, Canada [7]). This batch culture method of biofilm and planktonic cell susceptibility testing provides three internally consistent, comparative measurements from a single experiment: 1) the planktonic minimum inhibitory concentration (MIC), 2) the planktonic minimum bactericidal concentration (MBC), and 3) the minimum biofilm eradication concentration (MBEC). The MBEC assay is not prone to leakage and contamination since it is manipulated in a Laminar flow hood.
The present study was rooted in two principle aims. The first aim was to develop a method of high-throughput metal susceptibility testing of biofilms using the MBEC assay. As part of this goal, a metal specific neutralizing regime was employed to reduce the biological toxicity of many different metal cations and oxyanions in vitro. This procedure allowed for comparisons between the susceptibility of planktonic cells and biofilms to metals (between different strains and/or microbial species), and provides a significant modification of the procedure originally reported by Ceri et al. for antibiotic susceptibility testing [5,6]. Also presented here is quality control data for the MBEC technique that has not been published elsewhere.
The second aim was to apply this method to examine variables that may influence measurements of metal susceptibility. A common dilemma in comparative studies of different bacterial strains is the ability of each strain to form biofilms. In simple terms, the ratio of bacterial cells (i.e. chemically reactive targets) to metal ions may influence the determination of susceptibility. To address this, biofilm growth of different bacterial species was calibrated to allow relative comparisons of susceptibility between biofilms with similar cell densities. Here, the relative differences in E. coli and P. aeruginosa biofilm susceptibility to the heavy metalloid oxyanions selenite (SeO32-) and tellurite (TeO32-) were examined. These compounds are highly toxic, water soluble pollutants that are spread into the environment in the form of industrial effluent [8,9]. P. aeruginosa was 80 times more resistant to heavy metalloids than E. coli.
Another potential obstacle to accurate measurements from metal susceptibility testing is exposure time. Studies of short duration metal exposure in minimal media have indicated that biofilms are highly tolerant to heavy metals [2,3], whereas long exposures in rich media resulted in complete elimination of the biofilm [3]. Logically, these differences may be based on differences in either exposure time or growth medium. Here, the effect of these variables on the measured tolerance of E. coli biofilms to metal cations was examined. Time-dependent trends in biofilm susceptibility to metals were identified that were not altered by changing the medium in which bacteria were grown or challenged. Exposure time is thus an important consideration in the design of studies centred on biofilm tolerance to metals.
Results
Biofilm growth
Biofilms of Pseudomonas aeruginosa ATCC 27853, E. coli TG1, and E. coli JM109 were grown to an overall mean density of 6.6 ± 0.5, 6.7 ± 0.3, and 6.4 ± 0.4 log10 cfu peg-1 (respectively) in Luria-Bertani medium enriched with vitamin B1 (LB + B1). This corresponded to 9.5 and 24 h of incubation at 35°C for P. aeruginosa and E. coli, respectively. When grown in minimal salts vitamins glucose (MSVG) medium for 24 h at 35°C, biofilms of E. coli JM109 reached a mean cell density of 5.02 ± 0.55 log10 cfu peg-1. Mean and standard deviation calculations for cell densities were based on pooled data from all growth controls performed (i.e. from 36 to 59 replicates each).
Mean viable cell counts and standard deviation (SD) for P. aeruginosa ATCC 27853 and E. coli TG1 on the pegs of each row of the MBEC™-HTP assay are presented in (Fig. 1a and 1c, respectively, 4 to 6 replicates each). The cell counts for each row were pooled and compared using one-way analysis of variance (ANOVA). The cell density of biofilms grown on the different rows of pegs in the MBEC™-HTP assay were statistically equivalent (p = 0.842 for P. aeruginosa; p = 0.274 for E. coli). E. coli JM109 also formed statistically equivalent biofilms across the different rows of pegs. This was similar to E. coli TG1, and thus the data is not presented here.
To allow for a valid comparison of susceptibility data, the biofilm cell counts for P. aeruginosa ATCC 27853 and E. coli TG1 were compared using a Mann-Whitney U-test. The biofilm cell density of these two strains were statistically equivalent (p = 0.209). Under the growth conditions reported, E. coli JM109 formed biofilms with significantly less cell density than the other 2 strains when compared using a Mann-Whitney U-test (p < 0.001).
Microscopy
Biofilms were examined in situ using scanning electron microscopy. Photomicrographs of P. aeruginosa ATCC 27853 and E. coli TG1 are presented in (Fig. 1b and 1d, respectively). SEM pictures show the growth of surface-adherent bacteria in thin layers and mounds on the pegs of the MBEC™ device. These layers were estimated to be up to 10 μm in height in some areas. Biofilms heterogeneously covered the plastic surface. The biofilm growth of E. coli JM109 was similar to strain TG1, except with a slightly more sparse distribution across the peg surface (data not shown).
Susceptibility of E. coli and P. aeruginosa to metalloid oxyanions
In this study and as an example, biofilms of Escherichia coli TG1 and Pseudomonas aeruginosa ATCC 27853 were assayed for susceptibility to selenite (SeO32-) and tellurite (TeO32-). These two organisms formed biofilms with statistically equivalent cell density when grown as described above. This allowed for a direct comparison between the two bacterial strains for relative levels of resistance to these heavy metalloid oxyanions.
For E. coli TG1 and P. aeruginosa ATCC 27853, the mean and standard deviation of MIC, MBC and MBEC values for SeO32- and TeO32- are summarized in Table 1. These calculations were based on 4 to 8 independent replicates each. With 4 h of exposure, biofilms were up to 133 times more tolerant to TeO32- than the corresponding planktonic cultures. With regards to planktonic cells (derived from the surface of the biofilms), P. aeruginosa was 80 times more resistant to tellurite than E. coli.
Time-dependent susceptibility of E. coli biofilms to metal cations
Exposure time may be of pivotal importance as a controlled variable in the design of metal susceptibility assays. To directly address this problem, an array of 11 metal cations was chosen to represent groups 7B to 4A of the periodic table. Using the MBEC assay, the susceptibility of E. coli JM109 to these compounds was tested. E. coli JM109 has been a popular model microorganism for studies of metal resistance in bacteria. Biofilm and planktonic cell susceptibility at 2 and 24 h of exposure in both rich medium (LB + B1) and minimal medium (MSVG) was examined. A larger data set with a greater number of trials at 24 h of exposure in LB + B1 was used in this study to expand on an original report [10]. All other test conditions were previously unexamined. This data is presented in Tables 2 and 3, respectively.
In general, MIC, MBC and MBEC values were greater in rich medium than in minimal medium. This was probably due to the chelation of metal ions by phosphates and organic matter present in the growth medium. However, there were two trends that were invariant with regards to nutrient status: 1) In LB + B1, biofilms were 1.3 to 24 times more tolerant to metal cations than the corresponding planktonic cells when exposed for 2 h. Similarly, E. coli biofilms grown and tested in MSVG were 2.0 to 29 times more tolerant to metal cations with a similar exposure time. In combination with the susceptibility data for SeO32- and TeO32-, this would suggest that with short exposures biofilms are highly tolerant to metal toxicity. 2) By 24 h of exposure, biofilm and planktonic cultures were eradicated at similar concentrations of metal cations in almost every instance. This occurred independently of the growth medium used for bacterial susceptibility testing.
Collectively, these data suggest that biofilm tolerance to metals is time-dependent. We note that (in general) MIC values did not change with exposure time using this method (data not shown).
Discussion
This manuscript describes a high-throughput method for metal susceptibility testing of biofilms using the MBEC assay. This technique has the advantage that metal susceptibility testing may be done using a combinatorial approach. It is possible to employ different mean numbers of bacterial cells, alternate exposure times, various and diverse growth media formulations, as well as a broad range of metals (as well as other antimicrobial compounds). Many bacteria and yeasts, including Staphylococcus aureus [6,10], Mycobacterium spp. [11], Candida spp. (J.J. Harrison, H. Ceri and R.J. Turner, unpublished data), and Burholderia cepacia complex [12] are amenable to biofilm growth using the MBEC assay. Biofilms on the pegs of the MBEC device may be examined in situ using scanning electron microscopy, epifluorescent microscopy [13], and confocal laser-scanning microscopy (CLSM) [14].
As a quality control, a test for equivalent biofilm formation between rows of pegs in the MBEC device was performed. Biofilm cell density was statistically equivalent across the different rows of pegs for E. coli TG1, JM109 and P. aeruginosa ATCC 27853. This quality control needs to be a routine test for studies of biofilm susceptibility using this method. It is important to note that not all bacteria (particularly mucoidal isolates) are amenable to this method of growth (J.J. Harrison, H. Ceri and C. Stremick, unpublished data). If statistically equivalent biofilms are not generated using the MBEC trough format, it is possible to instead place the peg lid in a microtiter plate containing ~150 μl of inoculum in each well. This alternative system may be placed on a gyrorotary shaker to facilitate biofilm formation on the peg lid. In general, biofilms formed using a trough have a 5- to 10-fold greater cell density than those formed using the microtiter plate format (J.J. Harrison, H. Ceri and C. Stremick, unpublished data).
This procedure may be modified to discern biofilm and planktonic viable cell counts. This is included as an amendment with the online supplementary material. Viable cell counting is accomplished by serially diluting aliquots from the recovery and neutralization plates, which are subsequently plated onto agar and enumerated. A characteristic of the MBEC assay is that planktonic cultures are seeded by cells shed from the biofilm. Log-killing of biofilms may be calculated using this method, but as a consequence, log-killing of planktonic cells may not. However, this model does reflect the natural duality of the bacterial life cycle where a recalcitrant nidus of biofilm cells may survive metal exposure to shed planktonic cells back into the surroundings. Antibiotic MIC values obtained from this method are in most cases equivalent to MIC data derived from National Committee of Clinical Laboratory Standards (NCCLS) standard procedures [6]. In this report, MIC and MBC values for TeO32- were consistent with literature values obtained using alternate methods [9,15].
The use of neutralizing agents in metal susceptibility testing allows for more accurate comparisons between the susceptibility of planktonic cells and biofilms to metals. Metal carry over from challenge to recovery media can affect the determination of viable cell counts and bactericidal concentrations [16]. Pertinent to this method, carry over is of particular concern when comparing biofilm and planktonic cell MBEC and MBC values. For example, planktonic cells removed in media containing metals will be inhibited to a different extent during recovery than biofilms removed from the media containing metals. In this protocol, biofilms and planktonic cultures were treated with equal amounts of a neutralizing agent at a concentration with limited toxicity to the bacterial cells. Biofilm and planktonic cells were equally susceptible to metals with long exposure times, suggesting that the neutralizing agents are equivalently effective for the treatment of both bacterial forms. The limitation of chemical neutralization will be acknowledged here. Treating metal exposed biofilms with a chelator increases the number of viable cells recovered from killing kinetics experiments relative to identical experiments that do not use a neutralizing agent (J.J. Harrison, H. Ceri and R.J. Turner, unpublished data). However, it is very likely that the metal-chelator complex is toxic to bacteria [3]. The bottom line of this procedure is that the metal-chelator is less toxic than the free metal ion.
Luria-Bertani medium is appropriate for susceptibility testing of metal oxyanions, as these compounds remain soluble in this medium. In the case of metal cations, precipitation and/or complexation of metals occurs rapidly in rich media. In this instance, minimal media preparations may be more suitable (an excellent array of minimal media preparations have been designed by Teitzel and Parsek [2]). Although the absolute values of MIC, MBC and MBEC determinations were greater in rich than in minimal medium, the tolerance of biofilms to metal toxicity remained time-dependent in both growth conditions.
Reports in the literature have suggested that biofilm tolerance to metals may be time-dependent [3,10,17]. Biofilms of P. aeruginosa are up to 600 times more tolerant to heavy metals than planktonic cells (in minimal media with 2 to 5 h exposure) [2,3]. Using the MBEC assay, it has been noted for P. aeruginosa, Staphylococcus aureus, and E. coli that biofilms and planktonic cells are equally susceptible to metal cations (in rich media with 24 h exposure) [10]. Similar time-dependent phenomena have been previously described for the killing of E. coli JM109 and P. aeruginosa ATCC 27853 biofilms by SeO32- and TeO32- [10,18]. Here this has been revaluated and this study demonstrates that the foundation of this is exposure time. Thus, the amount of time bacteria are exposed to toxic metals is an important controlled variable that contributes to measurements of bacterial susceptibility to metals. This is important both in terms of comparisons of biofilm and planktonic cell tolerance to metals as well as in evaluation of data in the literature.
Conclusion
This high-throughput method is currently being used to elucidate a multifactorial model of metal tolerance in the bacterial biofilm [3,10,14,18,19]. The principle strength of this assay lies in the ability to rapidly screen changes in biofilm susceptibility to metals using diverse permutations of growth conditions, exposure times, and metal compounds. This type of combinatorial experimental approach would not be pragmatic using other methods.
The growth of different microorganisms may be calibrated to compare biofilms of equivalent cell density. This means that it is possible to compare susceptibility data of different strains without the concern that differences in resistance and/or tolerance are due to inequalities in growth of the biofilm populations. Lastly, exposure time influences the observed tolerance of biofilms to metals. Exposure time is thus a key consideration in the design of metal susceptibility assays.
A step-by-step protocol for this method is freely available from the authors [20].
Methods
Bacterial strains and media
Escherichia coli TG1, E. coli JM109 and Pseudomonas aeruginosa ATCC 27853 were stored at -70°C in Cryobanks™ (Prolab Diagnostics) according to the manufacturer's instructions. Bacteria were grown in Luria-Bertani media (pH 7.1, Difco) enriched with 0.001% vitamin B1 (LB + B1) or in the minimal salts vitamins glucose (MSVG, pH 7.1) of Teitzel and Parsek [2]. MSVG was specifically designed for metal susceptibility testing and minimizes the precipitation of metal cations in the growth medium. Sub-cultures and spot plates (see below) were grown on LB + B1 with 1.5% w/v granulated agar. All serial dilutions were performed using 0.9% saline.
Biofilm cultivation
The MBEC assay has two parts: 1) a plastic lid with 96 pegs that fits into 2) a corrugated trough. Biofilms were grown in the MBEC assay according to the method of Ceri and colleagues [5,6], and bacterial cultures and inocula were prepared according to the steps illustrated in Fig. 2(a–d) and as described here. Frozen stocks of bacteria were streaked out on Luria-Bertani agar to obtain a first-subculture. A single colony was picked from the first-subculture and again streaked out on agar obtain a second sub-culture. Colonies were collected from second-subcultures using a sterile cotton swab and suspended in broth medium to a 1.0 McFarland Standard. This suspension was diluted 30-fold in broth, and 22 ml of the 1 in 30 dilution was used to inoculate the MBEC assay. For demonstration purposes in this manuscript, we examined E. coli TG1, E. coli JM109 or P. aeruginosa ATCC 27853. Here, 22 ml of inoculum contained ~107 cfu/ml bacteria and this was transferred into the troughs. Starting bacterial number in the inocula were verified by viable cell counting. The inoculated devices were placed on a rocking table (Bellco Biotechnology) in an incubator at 35°C and 95% relative humidity at 2.5 rocks per minute. The shear force of the rocking motion facilitated the formation of 96 equivalent biofilms on the pegs.
Incubation times in LB + B1 medium were calibrated according to the growth rates of P. aeruginosa ATCC 27853 and E. coli TG1 (i.e. so that they would produce biofilms with an equivalent number of cells). Thus, E. coli and P. aeruginosa were grown for 24 and 9.5 h, respectively. For every MBEC assay, four pegs were broken from the lid (after it had been rinsed, see below), then 'sonciated' in sterile 0.9% saline. This was performed using an Aquasonic water-table sonicator (VWR International, model 250HT) for 5 minutes on the setting 'high' as previously described [5,6]. The disrupted biofilms were serially diluted and plated for viable cell counting. This growth control was used to verify that the appropriate number of bacteria had formed in the biofilm.
A test for equivalent biofilm growth on the rows of pegs in the MBEC assay was performed. This was accomplished by growing biofilms to the desired cell density, rinsing the pegs (i.e. by placing the peg lid into a microtiter plate with 200 μl of 0.9% saline in each well, termed a 'rinse plate'), then by sonicating (as described above) the biofilms into sterile saline (i.e. a fresh rinse plate). The disrupted biofilms were serially diluted then plated onto agar to determine viable cell counts (i.e. cfu peg-1). Plates were incubated for 24 h at 35°C then enumerated. Raw data were first log10-transformed and the mean viable cell counts for the different rows of pegs were compared using one-way analysis of variance (ANOVA).
Stock metal solutions
Silver nitrate (AgNO3), aluminum sulfate (Al2(SO4)3·18H2O), zinc sulfate (ZnSO4·7H2O), stannous chloride (SnCl2·2H2O) and copper sulfate (CuSO4·5H2O) were obtained from Fisher Scientific Company of Fairlawn, NJ. Nickel sulfate (NiSO4·6H2O), mercuric chloride (HgCl2), cobalt chloride (CoCl2·6H2O), lead nitrate (Pb(NO3)2), potassium tellurite (K2TeO3) and sodium selenite (Na2SeO3) were obtained from Sigma Chemical Company of St Louis, MO. Cadmium chloride (CdCl2·5/2H2O) was purchased from Terochem Laboratories of Edmonton, AB, and manganous sulfate (MnSO4H2O) from BDH of Toronto, ON. Reagent grade metal and metalloid compounds were purchased for the purposes of this study to minimize the potential influence of contaminating, residual metals.
All stock metal solutions, with the exception of Sn2+, were made up in double-distilled water at 5 times the highest concentration desired in the challenge plates. These stock solutions were passed through a 0.22 μm syringe filter into sterile glass vials and stored at room temperature. Sn2+ was disolved in 50% ethanol and stored in a sterile polypropylene tube.
Metal susceptibility testing
Susceptibility testing and exposure of biofilms to metals is summarized in Fig. 2(e–l). The peg lid of the MBEC assay fits inside a standard 96-well microtiter plate. This assay may standardized to any standard brand of microplate. The volume of the challenge media in the microplate must be sufficient to submerge the peg past the height of the biofilm produced in the device. Here, we used flat-bottom 96-well microtiter plates (Nunc) with a final volume of 200 μl. Serial two-fold dilutions of metalloid oxyanions were made in LB + B1 broth along the length of these microtiter plates (termed the 'challenge plate'), allowing the first well of each row to serve as a sterility control, and the last well to serve as a growth control. The challenge plates were incubated for the desired exposure time at 35°C and 95% relative humidity. After exposure, the peg lid was removed and rinsed twice with 0.9% saline and the biofilm disrupted by sonciation into LB + B1 broth containing the appropriate neutralizing agent (termed the 'recovery plate', see below). After removal of the peg lid, the challenge plate was covered with a new, sterile lid to protect the planktonic cultures in the challenge plate wells. The planktonic cultures were also treated with the appropriate neutralizing agent (see below).
Aliquots of the neutralized biofilm and planktonic cultures were spot plated onto LB + B1 agar and incubated for 48 h at 35°C. The MIC was determined by reading the optical density of the challenge plate at 650 nm (OD650). MBC and MBEC values were determined by qualitatively scoring the spot plates for bacterial growth. With the exception of Cu2+ and Ni2+ assays, MBEC values were redundantly determined by reading the OD650 of the recovery plates after 48 h.
Neutralizing agents
To differentiate between the bacteriostatic and bactericidal actions of the tested compounds, a two-step neutralizing protocol was designed to reduce the toxicity of residual metals. First, metal cations and oxyanions were treated with a chemical known to chelate or to react with the tested compound. The neutralizing agents currently and previously used in our laboratories have been summarized in Table 4. Second, neutralized cultures were plated onto a rich agar medium. Ideally, this latter step facilitates additional complexation of metals with components of the medium (such as phosphates, sulfates, and amines), and also allows diffusion of the metals into the agar. Thus, exposed bacteria are left on top of the agar medium to recover where there is a reduced concentration of biologically available metal.
As examples encountered in this study, metalloid oxyanions were reacted with 5 mM reduced glutathione (GSH, Sigma Chemical). GSH is used by the bacterial cell as a reduction-oxidation buffer to reductively eliminate a diverse array of inorganic oxidants [21,22], including selenite [23] and tellurite [24], and this is the basis for its use as a neutralizing agent here. Similarily, GSH was used to counter the effects of Zn2+, Co2+, Pb2+, Hg2+, and Cd2+ toxicity. Many metals are postulated to exert toxicity through oxidative stress on the thiol groups of proteins [25,26] and thus addition of GSH or L-cysteine may partially counteract this mechanism [16]. Sodium diethyldithiocarbate (Na2DDTC) was used to chelate Ni2+ and Cu2+, which rapidly formed metal precipitates with this organic chelator [27]. Citrate was used to coordinate Ag+ [10]. Tin was complexed using the amino acid glycine [28]. We have previously reported that 5-sulfosalicylic acid may be used as a chelator of Al3+ and Mn2+ [10]. As an improvement to this technique, we suggest that crushed asprin may be used as an alternative. Salicylate derivatives can be toxic to bacteria. The less soluble acetylated form may be employed with a wider range of bacterial strains (J.J. Harrison, H. Ceri, and R.J. Turner, unpublished data).
Stock solutions of citrate (0.5 M, Sigma), Na2DDTC (0.25 M, ICN), glutathione (0.25 M, Sigma), acetylsalicylic acid (~0.01 M, West-Can Pharmaceuticals, available at Calgary Coop), glycine (0.25 M, Bio-Rad Laboratories), and L-cysteine (0.25 M, Sigma) were prepared in double-distilled water and sterile filtered. With the exception of acetylsalicylic acid (ASA), all of these stocks were stored at -20°C until use. ASA was stored at room temperature. Neutralizing agents for biofilm cultures were added directly to LB+B1 broth used in the recovery plates. Neutralizing agents for the planktonic cultures were prepared at 5 times the desired neutralizing concentration in 0.9% saline. Aliquots (10 μl) of the diluted stock solutions were then added to the wells of a sterile 96-well plate (the neutralizing plate) to which 40 μl from each well of the challenge plate were added. The final concentration of neutralizing agent used to treat the planktonic cultures was thus equal to that used to treat biofilm cultures.
Scanning electron microscopy
Biofilms were examined using scanning electron microscopy (SEM) as previously described [4,6,10]. Briefly, pegs were broken from the lid of the MBEC device, rinsed once in 0.9% saline, then fixed for 16 h at 4°C in 5% glutaraldehyde dissolved in 0.1 M cacodylate buffer (pH 7.2). The next day, pegs were rinsed with 0.1 M cacodylate buffer, dehydrated with 95% ethanol, and then air dried for 30 h before mounting. SEM was performed using a Hitachi model 450 scanning electron microscopy according to the method of Morck and colleagues [29].
Statistical tests
Mean, standard deviation and one-way analysis of variance (ANOVA) calculations were performed using GraphPad InStat 3.0 (GraphPad Software Inc., San Diego, CA).
List of abbreviations
CLSM = confocal laser-scanning microscopy, MBC = minimum bactericidal concentration, MBEC = minimum biofilm eradication concentration, MIC = minimum inhibitory concentration, Na2DDTC = sodium diethyldithiocarbamate
Authors' contributions
JJH carried out the studies of metal susceptibility, designed the experimental protocols, performed statistical analysis of the data, and drafted the manuscript. HC and RJT conceived of the experiments, participated in their design, and helped to draft the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
HTP Metal Testing Protocol.doc - A step-by-step protocol. This is a copy of the step-by-step protocol to be made available free of charge on the world wide web through the authors' web site or BMC Microbiology whichever is preferred.
Click here for file
Acknowledgements
This research has been funded by operating grants from the Natural Sciences and Engineering Research Council of Canada (NSERC) to Drs. Raymond J. Turner and Howard Ceri. NSERC has also provided an Industrial Postgraduate Scholarship (IPS1) to Joe Harrison. Thanks to Carol A. Stremick, Nicole J. Roper and Erin A. Badry for expert technical assistance. Dr. Howard Ceri is Director of Business Development for MBEC BioProducts Inc. No corporate funding was provided for the research described in this manuscript.
Figures and Tables
Figure 1 Growth of bacterial biofilms in the MBEC assay. (A) Mean cell density of Pseudomonas aeruginosa ATCC 27853 biofilms on the pegs in different rows of the MBEC assay. Each value is expressed as the mean and standard deviation of 4 to 6 trials. There is no significant difference between cell density of biofilms in the different rows (p = 0.842 using one-way ANOVA). (B) SEM photomicrograph of a P. aeruginosa biofilm on the peg surface. (C) Mean cell density of Escherichia coli TG1 on the pegs in different rows of the MBEC assay. Each value is expressed as the mean and standard deviation of 4 to 6 trials. There is no significant difference between cell density of biofilms in the different rows (p = 0.274 using one-way ANOVA). (D) SEM photomicrograph of an E. coli biofilm on the peg surface. The bar represents 5 μm.
Figure 2 An overview of the high-throughput protocol for metal susceptibility testing using the MBEC assay. (A) Frozen stocks of bacteria were streaked out on the appropriate agar medium to obtain a first- and a subsequent second-subculture. (B) Colonies were collected from second-subcultures and suspended in broth medium to a 1.0 McFarland Standard. (C) This suspension was diluted 30-fold in broth, and the 1 in 30 dilution was used to inoculate the MBEC assay. (D) The inoculated device was placed on a rocking table in an incubator. (E) Serial dilutions of metal cations and oxyanions were set up along length of a microtiter plate along (the challenge plate). (F) The biofilms were rinsed to remove loosely adherent planktonic bacteria. (G) The first peg from each row was removed. These pegs were used to verify growth of the biofilms on the pegs. The peg lid was then inserted into the challenge plate. (H) During exposure, metals diffuse into the biofilm while planktonic cells are shed from the surface of the biofilm. Sloughed cells serve as the inoculum for planktonic MIC and MBC determinations. (I) The exposed biofilms were rinsed twice and the peg lid was inserted into fresh recovery medium containing the appropriate neutralizing agent (the recovery plate). The biofilms were disrupted into the recovery medium by sonciation on a water table sonicator. (J) Aliquots of planktonic cultures were transferred from the challenge plate to a microtiter plate containing the appropriate neutralizing agents (the neutralizing plate). (K) An aliquot from the recovery and neutralizing plates were spotted onto rich agar media. (L) MIC values are determined by reading the optical density at 650 nm (OD650) of the challenge plate after the desired period of incubation using a microtiter plate reader. Spot plates were qualitatively scored for growth to obtain MBC and MBEC values. MBEC values were redundantly determined by determining the A650 of the recovery plates after incubation.
Table 1 Comparative susceptibility of bacterial biofilms to metalloid oxyanions with 4 hours of exposure
Strain Metal MIC (mM) MBC (mM) MBEC (mM) Fold tolerance1
E. coli TG1 SeO32- > 5.8 > 5.8 > 5.8 na
TeO32- 0.02 ± 0.01 0.03 ± 0.02 > 2.0 ≥ 133
P. aeruginosa ATCC 27853 SeO32- > 187 > 187 > 187 na
TeO32- 1.6 ± 0.8 6.6 ± 3.0 > 16 2.4
na indicates a measurement that is not applicable
1the fold tolerance, given the sensitivity of the assay on a log2 scale, is equal to the ratio of the means of MBEC:MBC
Table 2 Susceptibility of Escherichia coli JM109 to metal cations with 2 or 24 h of exposure in rich (LB + B1) medium
Periodic group Metal ion Exposure time (h) MIC (mM) MBC (mM) MBEC (mM) Fold Tolerance1
7B Mn2+ 2 37 ± 0 > 149 > 149 na
24 198 ± 86 223 ± 86 1.1
8B Ni2+ 2 7.5 ± 2.1 > 140 >140 na
24 17 ± 0 29 ± 11 1.7
1B Cu2+ 2 4.5 ± 1.4 16 ± 0 16 ± 0 1.0
24 16 ± 0 16 ± 0 1.0
Ag+ 2 0.06 ± 0.02 0.06 ± 0.02 1.6 ± 0.7 24
24 0.06 ± 0.02 0.04 ± 0 0.7
2B Zn2+ 2 4.2 ± 2.4 125 ± 0 > 125 ≥ 2.0
24 31 ± 0 31 ± 0 1.0
Cd2+ 2 1.1 ± 0.2 55 ± 21 73 ± 0 1.3
24 2.3 ± 0 3.0 ± 1.3 1.3
Hg2+ 2 0.04 ± 0.02 0.14 ± 0.04 0.32 ± 0.22 2.3
24 0.04 ± 0 0.04 ± 0 1.0
3A Al3+ 2 nd > 304 > 304 na
24 19 ± 0 19 ± 0 1.0
4A Sn2+ 2 nd > 17 > 17 na
24 17 ± 0 17 ± 0 1.0
A portion of the data in this table (at the 24 h timepoint) represents a greater number of trials and reanalysis of the data originally reported in an earlier study [10].
na indicates a calculation that is not applicable
nd indicates an MIC that could not be determined due to metal precipitation
bold indicates the fold tolerance at 24 h of exposure
1the fold tolerance, given the sensitivity of the assay on a log2 scale, is equal to the ratio of the means of MBEC:MBC100
Table 3 Susceptibility of Escherichia coli JM109 to metal cations with 2 or 24 h of exposure in minimal (MSVG) medium
Periodic group Metal ion Exposure time (h) MIC (mM) MBC (mM) MBEC (mM) Fold Tolerance1
7B Mn2+ 2 33 ± 9 > 149 > 149 na
24 93 ± 37 84 ± 47 0.90
8B Co2+ 2 ≤ 0.27 26 ± 10 > 139 ≥ 11
24 1.0 ± 3 1.0 ± 3 1.0
Ni2+ 2 ≤ 0.27 > 139 > 139 na
24 0.48 ± 0.13 0.54 ± 0.39 1.1
1B Cu2+ 2 ≤ 0.25 ≤ 0.25 ≤ 0.25 na
24 ≤ 0.25 ≤ 0.25 na
Ag+ 2 ≤ 0.04 ≤ 0.04 0.59 ± 0 ≥ 15
24 ≤ 0.04 ≤ 0.04 1.0
2B Zn2+ 2 ≤ 0.25 11 ± 6 > 125 ≥ 23
24 0.55 ± 0.31 0.49 ± 0.35 0.89
Cd2+ 2 ≤ 0.14 5.1 ± 2.9 > 73 ≥ 29
24 0.49 ± 0.14 1.4 ± 1.1 2.9
Hg2+ 2 ≤ 0.02 ≤ 0.02 0.05 ± 0.02 ≥ 2.5
24 ≤ 0.02 ≤ 0.02 1.0
3A Al3+ 2 nd 42 ± 40 > 304 ≥ 14
24 1.5 ± 0.6 4.0 ± 1.4 2.7
4A Sn2+ 2 2.2 ± 0 8.6 ± 0 8.6 ± 0 1.0
24 4.3 ± 0 7.5 ± 2.2 1.7
Pb2+ 2 nd 20 ± 0 40 ± 0 2.0
24 4.9 ± 0 1.2 ± 0 0.25
na indicates a measurement that is not applicable
bold indicates the fold tolerance at 24 h of exposure
1the fold tolerance, given the sensitivity of the assay on a log2 scale, is equal to the ratio of the means of MBEC:MBC100
Table 4 Potential neutralizing agents for the microbiological application of inactivating metals cations and oxyanions*
Metal(loid) Neutralizing agent Maximum concentration1 Reference(s)
All oxyanions2 Glutathione 10 mM [10, 18, 24]
Al3+, Mn2+ Crushed acetylsalicylic acid (ASA) ~1–2 mM3 [30, 31]
Hg2+, Cd2+ Glutathione 10 mM [10]
L-cysteine 10 mM [16]
Cu2+, Ni2+ Diethyldithiocarbamic acid3 (DDTC) 2.5 mM (E. coli) up to 5 mM (P. aeruginosa) [27]
Sn2+ Glycine 10 mM [28]
Ag+ Sodium citrate 10 mM [10]
Zn2+, Co2+, Pb2+ Glutathione 10 mM [3]
*This is the first part of a two-part strategy to reduce the in vitro toxicity of metals (see text for details)
1 The maximum concentration tested and employed in studies by our laboratories using the high-throughput metal susceptibility testing method presented in this paper.
2 Tested heavy metal and metalloid oxyanions from our laboratories include TeO32-, TeO42-, SeO32-, CrO42-, AsO43-, AsO2-, WO42- and MoO42-.
3 Application is limited by the low solubility of salicylic acid and its acetylated derivatives in water.
4 The maximum concentration listed is inhibitory to bacterial growth in broth culture. Recovery broth media must be spot plated onto agar to allow bacterial growth and determination of accurate MBC and MBEC values.
==== Refs
Hall-Stoodley L Costerton JW Stoodley P Bacterial biofilms: From the natural environment to infectious diseases Nature Reviews Microbiology 2004 2 95 108 15040259 10.1038/nrmicro821
Teitzel GM Parsek MR Heavy metal resistance of biofilm and planktonic Pseudomonas aeruginosa Applied and Environmental Microbiology 2003 69 2313 2320 12676715 10.1128/AEM.69.4.2313-2320.2003
Harrison JJ Turner RJ Ceri H Persister cells, the biofilm matrix and tolerance to metal cations in biofilm and planktonic Pseudomonas aeruginosa Environmental Microbiology 2005 7 981 994 15946294 10.1111/j.1462-2920.2005.00777.x
Olson ME Ceri H Morck DW Buret AG Read RR Biofilm bacteria: formation and comparative susceptibility to antibiotics Canadian Journal of Veterinary Research 2002 66 86 92 11989739
Ceri H Olson ME Morck DW Storey D Read RR Buret AG Olson B The MBEC assay system: Multiple equivalent biofilms for antibiotic and biocide susceptibility testing Methods in Enzymology 2001 337 377 384 11398443
Ceri H Olson ME Stremick C Read RR Morck DW Buret AG The Calgary Biofilm Device: New technology for rapid determination of antibiotic susceptibilities in bacterial biofilms Journal of Clinical Microbiology 1999 37 1771 1776 10325322
MBEC Bioproducts Incorporated
Chasteen TG Bentley R Biomethylation of selenium and tellurium: microorganisms and plants Chemical Reviews 2003 103 1 25 12517179 10.1021/cr010210+
Turner RJ Tellurite toxicity and resistance in Gram-negative bacteria Recent Research Developments in Microbiology 2001 5 69 77
Harrison JJ Ceri H Stremick C Turner RJ Biofilm susceptibility to metal toxicity Environmental Microbiology 2004 6 1220 1227 15560820 10.1111/j.1462-2920.2004.00656.x
Bardouniotis E Ceri H Olson ME Biofilm formation and biocide susceptibility testing of Mycobacterium fortuitum and Mycobacterium marium Current Microbiology 2003 46 28 32 12432460 10.1007/s00284-002-3796-4
Tomlin KL Coll OP Ceri H Interspecies biofilms of Pseudomonas aeruginosa and Burkholderia cepacia Canadian Journal of Microbiology 2001 47 949 954 11718549 10.1139/cjm-47-10-949
Tomlin KL Clark SR Ceri H Green and red fluorescent protein vectors for use in biofilm studies of the intrinsically resistant Burkholderia cepacia complex Journal of Microbiological Methods 2004 57 95 106 15003692 10.1016/j.mimet.2003.12.007
Harrison JJ Ceri H Badry EA Roper NJ Tomlin KL Turner RJ Effects of the twin-arginine translocase on the structure and antimicrobial susceptibility of Escherichia coli biofilms Canadian Journal of Microbiology 2005 submitted work
Taylor DE Bacterial tellurite resistance Trends in Microbiology 1999 7 111 115 10203839 10.1016/S0966-842X(99)01454-7
Russel AD Ahonkhai I Rogers DT Microbiological applications of the inactivation of antibiotics and other antimicrobial agents Journal of Applied Bacteriology 1979 46 207 245 378921
Cortizo MC Lorenzo de Mele MF Cytotoxicity of copper ions released from metal: variation with the exposure period and concentration gradients Biological Trace Element Research 2004 102 129 142 15621934 10.1385/BTER:102:1-3:129
Harrison JJ Ceri H Stremick C Turner RJ Differences in biofilm and planktonic cell mediated reduction of metalloid oxyanions FEMS Microbiology Letters 2004 235 357 362 15183885 10.1016/j.femsle.2004.05.005
Harrison JJ Turner RJ Ceri H Metal tolerance in bacterial biofilms Chapter 2: Recent Research Developments in Microbiology 2005 9 33 55
University of Calgary, Personal Web Page of Raymond J. Turner
Aslund F Zheng M Beckwith J Storlz G Regulation of the OxyR transcription factor by hydrogen peroxide and the cellular thiol-disulfide status Proceedings of the National Academy of Sciences of the United States of America 1999 96 6161 6165 10339558 10.1073/pnas.96.11.6161
Ding H Demple B Thiol-mediated disassembly and reassembly of [2Fe-2S] clusters in the redox-regulated transcription factor SoxR Biochemistry 1998 37 17280 17286 9860842 10.1021/bi980532g
Turner RJ Weiner J Taylor DE Selenium metabolism in Escherichia coli BioMetals 1998 11 223 227 9850565 10.1023/A:1009290213301
Turner RJ Aharonowitz Y Weiner J Taylor DE Glutathione is a target of tellurite toxicity and is protected by tellurite resistance determinants in Escherichia coli Canadian Journal of Microbiology 2001 47 33 40 15049447 10.1139/cjm-47-1-33
Stohs SJ Bagchi D Oxidative mechanisms in the toxicity of metal ions Free Radical Biology and Medicine 1995 18 321 336 7744317 10.1016/0891-5849(94)00159-H
Turner RJ Weiner J Taylor DE Tellurite mediated thiol oxidation in Escherichia coli Microbiology 1999 145 2549 2557 10517608
Gottofrey J Borg K Jasmin S Effect of postassium ethylxanthate and sodium diethyldithiocarbamate on the accumulation and deposition of nickel in the brown trout (Salmo trutta) Pharmacology and Toxicology 1988 63 46 51 2840649
Diurdjevic P Djokic D Protein interactions with bivalent tin. 1. Hydrolysis and complexation of tin(II) with glycine Journal of Inorganic Biochemistry 1996 62 17 29 8936420 10.1016/0162-0134(95)00085-2
Morck DW Lam K McKay SG Olson ME Prosser B Ellis BD Cleeland R Costerton JW Comparative evaluation of fleroxacin, ampicillin, trimethoprim-sulfamethoxazole, and gentamicin as treatments of catheter associated urinary tract infections in a rabbit model. American Journal of Medicine 1994 94 23S 30S
Missy P Lanhers MC Grignon Y Joyeux M Burnel D In vitro and in vivo studies on chelation of manganese Human and Experimental Toxicology 2000 19 448 456 11125715 10.1191/096032700682694260
Graff L Muller G Burnel D In vitro and in vivo evaluation of potential aluminum chelators Veterinary and Human Toxicology 1995 88 271 292
|
16202131
|
PMC1262725
|
CC BY
|
2021-01-04 16:38:05
|
no
|
BMC Microbiol. 2005 Oct 3; 5:54
|
latin-1
|
BMC Microbiol
| 2,005 |
10.1186/1471-2180-5-54
|
oa_comm
|
==== Front
BMC MicrobiolBMC Microbiology1471-2180BioMed Central London 1471-2180-5-561620214410.1186/1471-2180-5-56Research ArticleA gonococcal homologue of meningococcal γ-glutamyl transpeptidase gene is a new type of bacterial pseudogene that is transcriptionally active but phenotypically silent Takahashi Hideyuki [email protected] Haruo [email protected] Department of Bacteriology, National Institute of Infectious Diseases, Tokyo, Japan2005 4 10 2005 5 56 56 13 6 2005 4 10 2005 Copyright © 2005 Takahashi and Watanabe; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
It has been speculated that the γ-glutamyl transpeptidase (ggt) gene is present only in Neisseria meningitidis and not among related species such as Neisseria gonorrhoeae and Neisseria lactamica, because N. meningitidis is the only bacterium with GGT activity. However, nucleotide sequences highly homologous to the meningococcal ggt gene were found in the genomes of N. gonorrhoeae isolates.
Results
The gonococcal homologue (ggt gonococcal homologue; ggh) was analyzed. The nucleotide sequence of the ggh gene was approximately 95 % identical to that of the meningococcal ggt gene. An open reading frame in the ggh gene was disrupted by an ochre mutation and frameshift mutations induced by a 7-base deletion, but the amino acid sequences deduced from the artificially corrected ggh nucleotide sequences were approximately 97 % identical to that of the meningococcal ggt gene. The analyses of the sequences flanking the ggt and ggh genes revealed that both genes were localized in a common DNA region containing the fbp-ggt (or ggh)-glyA-opcA-dedA-abcZ gene cluster. The expression of the ggh RNA could be detected by dot blot, RT-PCR and primer extension analyses. Moreover, the truncated form of ggh-translational product was also found in some of the gonococcal isolates.
Conclusion
This study has shown that the gonococcal ggh gene is a pseudogene of the meningococcal ggt gene, which can also be designated as Ψggt. The gonococcal ggh (Ψggt) gene is the first identified bacterial pseudogene that is transcriptionally active but phenotypically silent.
==== Body
Background
Two members of the gram-negative diplococci, Neisseria meningitidis and Neisseria gonorrhoeae, are particularly associated with pathological infections. N. meningitidis is specialized for the mucosa of the nasopharynx and causes meningitis and septicemia. N. gonorrhoeae is adapted for the mucosa of the urogenital tract and causes gonorrhoea and pelvic inflammatory diseases. Both species colonize only humans and share a great deal of relatedness at the nucleotide level [1]. This high degree of relatedness is reflected in the many common genetic, biochemical and antigenic features of the two bacteria.
γ-Glutamyl transpeptidase (also called γ-glutamyl aminopeptidase) (EC2.3.2.2; GGT) catalyzes the hydrolysis of γ-glutamyl compounds, and is found in a variety of bacteria such as Escherichia coli [2] and Helicobacter pylori [3,4]. To distinguish N. meningitidis from N. gonorrhoeae, GGT activity is used as one of the identification markers for N. meningitidis because N. meningitidis is positive for this activity but N. gonorrhoeae and related species, e.g., Neisseria lactamica, are not [5]. In fact, the detection of GGT activity is applied for the identification of N. meningitidis in the Gonochek II enzymatic identification system (E-Y Laboratories Inc., U.S.A.) [6-9]. From these empirical facts, it was believed that the gene encoding for GGT should exist only in N. meningitidis, but this has not been proven yet [3].
Recent remarkable progress in the sequencing of various genomes has led to the detection of nucleotide sequences that appear to be phenotypically silent, termed pseudogenes. The pseudogenes are defined as DNA sequences of formerly functional genes rendered nonfunctional by mutations and usually identified by their disrupted open reading frames (ORFs). Pseudogenes have been identified in a variety of eukaryotes, including insects [10], plants [11], and particularly vertebrates [10,12], but are relatively few in the bacterial genomes. Notable exceptions are intracellular bacterial parasites such as Rickettsia prowazekii and Mycobacterium leprae [13], which seem to have lost many genes due to obtaining nutritional supplies from the host cells. Cryptic genes such as the cel operon in E. coli [14-16] and the flagellar master operon in the genus Shigella [17-19] seem to be a kind of pseudogenes, but are different from pseudogenes because the cryptic genes completely retain intact ORFs, which can be occasionally activated by rare genetic events such as mutation, recombination, insertion of elements. As a whole, compared to the pseudogenes in eukaryotes, relatively few pseudogenes have been reported in bacterial genomes [20].
In this study, a gonococcal ggh gene, which is highly homologous to the meningococcal ggt gene, was found to be pseudogene. Sequence analyses of the flanking regions of both the ggt and ggh genes suggest that both genes were derived from a gene in a common ancestor, and subsequently diversified.
Results
The gonococcal ggh gene was highly homologous to the meningococcal ggt gene
Since GGT activity was detected only in N. meningitidis among the related species, it was speculated that the corresponding gene also existed only in N. meningitidis. However, by BLAST search, the nucleotide sequences highly homologous to the meningococcal ggt gene were found in the genome of N. gonorrhoeae FA1090 [GenBank:NC_002946]. The overall nucleotide sequence of the meningococcal ggt homologue was approximately 95 % identical to that of the meningococcal ggt gene (data not shown and additional file). Eleven N. gonorrhoeae clinical strains were analyzed by PCR, and the corresponding DNA fragments were amplified in all of these strains (Figure 1B), indicating that this gene was generally present in N. gonorrhoeae. To analyze whether ggt homologues existed in the genomes of the other neisserial strains, Southern blotting was performed (Figure 1C). DNA fragments that hybridized with the meningococcal ggt gene were found in the meningococcal and gonococcal genomes (Figure 1C lanes 1–3) but not in the other neisserial genomes (Figure 1C lanes 4–11). These results suggested that the meningococcal ggt homologue was present only in N. gonorrhoeae among the neisserial species examined. The putative gene in N. gonorrhoeae was named ggh (ggt gonococcal homologue).
Variations of the nucleotide sequences of ggh genes among clinical isolates
To characterize the gonococcal ggh gene, the ggh genes were amplified from the chromosomal DNA of 11 N. gonorrhoeae strains and sequenced. The nucleotide sequences of the ggt genes from 7 N. meningitidis strains and of the ggh genes from the 11 N. gonorrhoeae strains were aligned, and the distance matrix calculated from these data was displayed as a phylogenetic tree (Figure 2A). The results revealed that the nucleotide sequences of the gonococcal ggh genes were more divergent than those of the meningococcal ggt genes (Figure 2A). Alignment of the nucleotide sequences of the 11 gonococcal ggh genes also showed that the mutations in the ggh gene consisted of the following four polymorphisms: 1) a 6-base insertion (named Type I), 2) an ochre mutation (Type II), 3) a 7-base deletion (Type III), 4) a 46-base insertion (Type IV) (Figure 2B and Table 2). In addition, all of the 11 ggh genes had one-nucleotide substitutions compared to the ggt genes in the same 25 sites (Figure 2B and additional file), with only 2 exceptions: a one-nucleotide variation in NIID109 (48th base A to G) and another in NIID105 (213th base G to A) (Table 2). The one-base substitutions in the common sites of the ggh genes strongly suggested that reconstruction of the ggh gene would have occurred at an early stage after speciation (See Discussion).
Putative amino acid sequences in hypothetical coding region of the ggh genes
Due to the ochre (Type II) and 7-base deletion (Type III) mutations, the ORF in each ggh gene was completely disrupted by the formation of 8 or 20 stop codons (Figure 3A). In fact, none of the gonococcal isolates showed any GGT activity (data not shown and [5,9]), indicating that there was no expression of functional GGT-like protein in N. gonorrhoeae. All of these results showed that the gonococcal ggh gene was a pseudogene of the functional ggt gene in N. meningitidis. On the other hand, if the two types of mutations (Types II and III) in the ggh genes were artificially corrected, the hypothetical amino acid sequences were approximately 97 % identical to those of the meningococcal ggt genes and were highly conserved among the gonococcal ggh genes (Figure 3B). This result also supported the idea that the ggh and ggt genes were derived from a common ancestral gene and that the translational inactivation of the gonococcal ggh gene was solely due to the ochre (Type II) and the frame-shift mutations caused by the 7-base deletion (Type III).
The genetic organization of the ggt- and ggh-flanking regions in the genomes of N. meningitidis and N. gonorrhoeae
By using the information in the database for N. meningitidis strain MC58 [21], N. gonorrhoeae strain FA1090 and N. lactamica (the neisserial species most closely related to the above two species) ST-640 strain, the flanking regions of the meningococcal ggt and the gonococcal ggh genes were further analyzed. The ggt and ggh genes were both localized in the identical gene cluster of fbp-ggt (or ggh)-glyA-opcA-dedA-abcZ in the genomes of N. meningitidis and N. gonorrhoeae, respectively (Figure 4). The fbp-glyA-dedA-abcZ locus was also found in the genome of N. lactamica but lacked the ggt and opcA homologues (Figure 4). This highly conserved genetic organization implied that a DNA island containing an original ggt gene was first incorporated into an ancestor's genome of the above three species and subsequently diversified after the speciation (see Discussion).
The expression of the ggh gene in N. gonorrhoeae
The hitherto identified bacterial pseudogenes are not expressed transcriptionally or translationally [20,22-24]. To examine the ggh transcriptional expression, dot blot analysis was first performed and RNA that hybridized with the ggt probe was detected in the total RNAs of 4 N. gonorrhoeae strains (Figure 5A). The transcriptional expression was also confirmed by RT-PCR, and the products were amplified with all 3 sets of primers from total RNA of all 4 N. gonorrhoeae strains tested (Figure 5B). These results strongly suggested that the full-length ggh gene transcript was expressed in N. gonorrhoeae. Primer extension analysis further revealed that the gonococcal ggh RNA was transcribed from the same starting point as the meningococcal ggt mRNA (Figure 5C and 5D). All of these results indicated that the gonococcal ggh gene was transcriptionally active though it was a pseudogene.
To further study the translational expression of the truncated GGT-like protein in N. gonorrhoeae, Western blotting was performed with anti-meningococcal GGT rabbit antiserum [25]. When the same amounts of the whole cell extracts were analyzed (Figure 6B), approximately 15-kDa bands were detected in the extracts of NIID103 and NIID106 N. gonorrhoeae strains (Figure 6A lanes 5 and 7). Because the 15-kDa protein was not observed in the Δggh background of NIID103 and NIID106 (Figure 6A lanes 6 and 8), the 15-kDa protein was thought to be the ggh gene product whose translation was terminated at the 145th codon (Figure 3B). However, the 15-kDa protein was not found in the extracts of the ATCC49226 and NIID54 N. gonorrhoeae strains (Figure 6A, lanes 3 and 4) and seemed not to be expressed in any of the other gonococcal strains (see Discussion).
Discussion
In this study, it was shown that the gonococcal ggh gene is a member of bacterial pseudogenes, and is transcribed but not properly translated so that active ggh protein product is not produced. 11Fh-mtTFA [26], OsMu4-2 [27], NA88-A [28], Makorin1-p1 [29], Dnm3a2 [30] and pseudoNOS [31] genes are known to be transcriptionally active eukaryotic pseudogenes. In neisseriae, the gonococcal porA and ΨopcB genes have been reported as neisserial pseudogenes [1,32,33]. The porA pseudogene contains mutations in the promoter and the coding regions, and is not translated [24]. While some hypothetical bacterial pseudogenes with repetitive runs of A and T are speculated to be potentially expressed by transcriptional slippage [34], the expression, including the transcription, has not been proven yet. To our knowledge, the gonococcal ggh gene is the first identified bacterial pseudogene that is transcriptionally active.
The 15-kDa derivative of the putative ggh protein product is detected in the NIID103 and NIID106 N. gonorrhoeae strains but not in the ATCC49226 and NIID54 N. gonorrhoeae strains (Figure 6A). Since the predicted amino acid sequences of the putative 15-kDa proteins seem to be similar among the 4 gonococcal strains, the reason why the 15-kDa protein was not detected in ATCC49226 and NIID54 is not clear. The 15-kDa protein might be degraded in ATCC49226 and NIID54 backgrounds but not in NIID103 and NIID106 backgrounds. It seems unlikely that the 15-kDa protein encoded by the ggh gene has an essential function for N. gonorrhoeae because the 15-kDa protein is not always detected in any of gonococcal strains (Figure 6A).
Why does the gonococcal ggh gene still retain the transcriptional activity? There are some examples in which RNAs transcribed from pseudogenes have some biological functions: antisense RNA expressed from the pseudoNOS gene hybridizes with nNOS (nitric oxide synthase) mRNA, resulting in the suppression of the nNOS gene expression in the neurons of the snail Lymnaea stagnalis [31]. RNA of Makorin1-1p, a pseudogene of Makorin1, regulates the Makorin1 mRNA stability, which is important for the correct formation of the kidneys and bone in mice [29]. Some eukaryotic genes may be duplicated and one of the plural genes may be subsequently reconstructed due to its redundancy, resulting in a pseudogene. However, since a bacterial pseudogene generally does not have a functional counterpart (wild-type gene) in a single organism, the ggh RNA does not seem to have the same kind of biological function as the pseudoNOS and Makorin1-1p genes. In fact, we could not find any prominent phenotype for a Δggh gonococcal mutant (unpublished data). However, we cannot exclude the possibility that the ggh RNA has some biological function(s) in other milieus such as the urogenital tract and further analyses will be required to address this possibility.
The ggt and ggh genes are located in the fbp-ggt (ggh)-glyA-opcA-dedA-abcZ common gene cluster in the genomes of both N. meningitidis and N. gonorrhoeae (Figure 4) [23]. The genome of N. lactamica lacks ggt and opcA homologues in the fbp-glyA-dedA-abcZ locus (Figure 4 and [23]). It would not be the result of chance that the two ggt (or ggh) and opcA genes are located in the same respective sites of the fbp-glyA-dedA-abcZ gene locus of both N. meningitidis and N. gonorrhoeae but not in that of N. lactamica. Moreover, it is also unlikely that a nonfunctional ggh gene was horizontally transferred into the gonococcal fbp-glyA-opcA-dedA-abcZ gene cluster since such a pseudogene could not have been sustained due to the lack of selection [23]. Therefore, it seems more probable that the fbp-ggt-glyA-opcA-dedA-abcZ gene cluster was present in an ancestor of the three neisserial species, and has been subsequently diversified independently among the three species, as shown for the opcA gene [23]. During the diversification, the meningococcal ggt gene has been maintained in an active state while the gonococcal ggh gene has been reconstructed by insertion, deletion and substitutions, resulting in the translational inactivation. In N. lactamica, the ggt and opcA homologues might have been lost because of their dispensability for N. lactamica (see below).
It is also interesting that the ggh gene has not been fully deleted from the gonococcal genome. The kinds and sites of mutations in the ggh genes are relatively few and highly conserved, respectively, among the gonococcal isolates (Figure 2B, Table 2 and additional file). It is also noted that, while in general the RNA polymerase-binding sites and SD regions of pseudogenes are highly degraded, there are also a few exceptions in species such as Y. pestis that could have emerged in recent evolutionary times [35]. Since the ribosome-binding sites (Shine-Dalgarno regions) of the gonococcal ggh gene are identical to those of the meningococcal ggt gene and the RNA polymerase-binding sites are almost completely conserved (with a one-nucleotide difference) (Figure 5D), it is speculated that the reconstruction of the ggh gene may have occurred in relatively recent evolutionary times. From the evolutional viewpoint, the drastic deletion of the approximately 2-kb DNA region containing the ggh gene in N. gonorrhoeae may not have been likely to occur in such a short period. Alternatively, deletion of the 2-kb DNA region may not be more advantageous for gonococcal evolution than reconstruction involving short deletions, insertions and substitutions.
The maintenance of a functional ggt gene in N. meningitidis would have some advantages for its survival. N. meningitidis causes meningitis, which is due to the meningococcal invasion into the human central nervous system, including cerebrospinal fluid (CSF) [36-38]. It has been shown that meningococcal GGT has a physiological function of acquiring cysteine from environmental γ-glutamyl-cysteinyl peptides under cysteine-limited environments such as the CSF [39]. Almost all (98.8 %) meningococcal isolates from humans are positive for GGT activity [40]. All of these results suggest that the GGT activity is important for N. meningitidis but not for N. gonorrhoeae. The dispensability of GGT activity for N. gonorrhoeae seems to be consistent with the fact that a cysteine-limited milieu such as the CSF in humans is not a natural gonococcal habitat. However, it is not very likely that the milieu of CSF exerts selective pressure for an active ggt gene because human CSF is not a relevant milieu for human-to-human spread of meningococcus [41]. We believe that the meningococcal GGT must have some unknown essential function(s) for N. meningitidis and further studies will elucidate the function(s).
Conclusion
Our data on the ggh gene indicate that the ggh gene in N. gonorrhoeae is a pseudogene of the functional ggt gene in N. meningitidis. To our knowledge, the ggh gene is the first reported bacterial pseudogene that is transcriptionally active but phenotypically silent. Our findings may also contribute to understanding the speciation of N. meningitidis and N. gonorrhoeae.
Methods
Bacterial strains and growth conditions
Seven N. meningitidis strains (H44/76, H114/90, 2996, NIID68, NIID76, NIID413 and NIID414) were described in our previous reports [42,43]. Ten N. gonorrhoeae strains (NIID54, NIID102, NIID103, NIID104, NIID105, NIID106, NIID107, NIID108, NIID109 and NIID111), one N. mucosa strain (NIID16) and one N. sicca strain (NIID17) were clinical isolates donated by T. Kuroki and Y. Watanabe. All of the clinical strains were mutually independent: they were isolated in different periods and from different persons who lived in different areas of Japan. The following 7 neisserial strains were obtained from ATCC (species / ATCC no.): N. gonorrhoeae / ATCC49226; N. lactamica / ATCC23970; N. flavescence / ATCC13120; N. denitrificans / ATCC14686; N. elongata / ATCC25295; N. canis / ATCC14687; N. cinerea / ATCC14685. All of the strains were stored by the gelatin disc method [42] and cultivated on GC agar (Becton-Dickinson) supplemented with 1 % IsovitaleX enrichment (Becton-Dickinson) at 37°C in 5 % CO2, or in GC broth (1.5 % proteose-peptone, 0.5 % NaCl, 0.05 % soluble starch, 0.1 % K2HPO4, 0.4 % KH2PO4, 1 % IsoVitaleX, 5 mM NaHCO3 10 mM MgCl2) at 37°C with shaking.
Isolation of chromosomal DNA, PCR, Southern blotting and dot blotting
Isolation of chromosomal DNA, PCR, Southern blotting and dot blotting were performed as described in our previous report [43].
Nucleotide sequence determination and analyses
The ggt and ggh genes were amplified with a set of primers (ggt-3 and ggt-4) by PCR, and the resulting products were purified using High Pure PCR Product Purification Kit (Roche) as templates for sequencing. The sequencing was performed as described previously [25]. Primers used for sequencing are listed in Table 1. Raw data from the ABI sequencer were assembled with the program DNASIS ver. 3.2 (HITACHI, Japan). The sequence alignment was performed with GENETYX-MAC ver.11 (GENETYX, Japan). Phylogenetic analyses were performed by constructing a distance matrix of nucleotide mismatches using the web site of the Belozersky Institute at Moscow State University [44] and visualized by Split decomposition analysis with the program SPLITSTREE, version 3.2 [45].
Construction of Δggh::spc N. gonorrhoeae mutants
HT1195 (NIID103 Δggh::spc) and HT1196 (NIID106 Δggh::spc), in which a spectinomycin resistance gene (spc) was inserted into the ggh gene, were constructed as follows: A 2-kb fragment containing the ggh gene of NIID103 or NIID106 was amplified by PCR and cloned in the SmaI site of pUC18 (Takara Bio) to construct pHT412 or pHT413, respectively. A blunted 1-kb fragment containing the spc gene [25] was inserted into the EcoRV sites of pHT412 and pHT413, respectively. The EcoRV sites are located at 277 bp and 1642 bp (NIID103 [DDBJ:AB175025]) and at 271 bp and 1590 bp (NIID106 [DDBJ:AB175027]) downstream from the transcriptional start point of each ggh gene, respectively (see Figure 5D). The resulting plasmids were named pHT414 and pHT415, respectively. Five hundred nanograms of the plasmids linearized by digestion with EcoRI were transformed into NIID103 and NIID106, respectively, as described previously [43]. Spectinomycin-resistant clones were selected on GC agar plates containing 75 μg/ml spectinomycin. The resulting mutants were isolated as Δggh::spc NIID103 (HT1195), and Δggh::spc NIID106 (HT1196), respectively. The allelic exchange was confirmed by PCR and Southern blotting.
RT-PCR
Bacteria grown on GC agar plates were suspended in 20 ml of GC broth to an OD600 of 0.1 and continuously cultured to mid-log phase (OD600 of ~ 0.6) at 37°C with shaking. The total RNA was isolated from the harvested bacteria as previously described [46] with an additional treatment with DNase I. RT-PCR was performed using one step RT-PCR Kit Ver. 1.1 (Takara Bio, Japan) with approximately 2 μg of total RNA according to the manufacturer's instructions. The products were visualized by electrophoresis in a 2 % agarose gel followed by ethidium bromide staining.
Primer extension analysis
Fifty micrograms of total RNA and 5 pmol of biotin-labeled primer (ggt-ext-2) were hybridized in 20 μl of buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA, 250 mM KCl). The hybridized RNA-DNA probe was treated with 35 units of AMV reverse transcriptase (RTase) XL (Takara Bio) in a reaction mixture (250 μM dNTPs, 1 × AMV reverse transcriptase XL buffer) at 37°C for 30 min. The ethanol-precipitated DNA product was dissolved in 20 μl of formamide dye (80 % formamide, 10 mM NaOH, 1 mM EDTA, 0.025 % bromophenol blue, 0.025 % xylene cyanol). Sequencing of the ggt and ggh genes with the ggt-ext-2 primer was performed by using ΔTh polymerase sequencing high -cycle- (TOYOBO, Japan). Aliquots of the reaction products were analyzed by electrophoresis on 8 % acrylamide-7 M urea gel followed by capillary blotting to Hybond-N+ (Amersham). The bands were visualized with Imaging high (TOYOBO) according to the manufacturer's protocol.
SDS-PAGE and Western blotting
SDS-PAGE and Western blotting were performed as described previously [43] using 1 × 103-fold diluted anti-GGT polyclonal rabbit antiserum [25] and 2 × 103-fold diluted horseradish peroxidase-conjugated secondary antibody (Amersham)
Abbreviations
GGT, γ-glutamyl transpeptidase; IS, insertional sequence; ORF, open reading frame; RTase, reverse transcriptase
Authors' contributions
HT carried out all of the studies including molecular genetic studies, sequence determination, sequence analyses and drafting the manuscript. HW made a critical reading of the manuscript and final approval of the version to be published.
Supplementary Material
Additional File 1
Alignment of the nucleotide sequences within the ggt and ggh genes of N. meningitidis H44/76 [DDBJ:AB089320]; N. meningitidis H119/90 [DDBJ: AB211221]; N. gonorrhoeae ATCC49226 [DDBJ:AB175023]; N. gonorrhoeae NIID103 [DDBJ:AB175025] and N. gonorrhoeae NIID106 [DDBJ:AB175029] strains, respectively. Sequence identity is represented as *, polymorphism within the sequences of the 5 strains is indicated by the appropriate letter, and the absence of a base is shown with a hyphen (-).
Click here for file
Acknowledgements
We thank T. Kuroki and Y. Watanabe for donating neisserial clinical isolates. The data from the gonococcal genome sequencing project at the University of Oklahoma [47] and the genome sequence of Neisseria lactamica at the Sanger Institute are gratefully acknowledged. The N. lactamica genome sequence was generated by the Sanger Institute Pathogen Sequencing Unit [48]. This work was supported by grants from the Ministry of Health, Labor and Welfare of Japan and from the Ministry of Education, Culture, Sports, Science and Technology of Japan (Grant no. 13770142 and 16790265).
Figures and Tables
Figure 1 The presence of a meningococcal ggt gene homologue in N. gonorrhoeae. A. A schematic diagram showing the position of the set of primers used for the detection of ggt gene homologues. The black bar shows the region of the DNA probe used for Southern blotting in panel C. B. Amplification of gonococcal ggh gene by PCR. The genomic DNAs of neisserial species used for PCR were as follows: lane 1, H44/76 (N. meningitidis ggt+); lane 2, NIID113 (N. meningitidis ggt::IS) [49]; lane 3–13, ATCC49226, NIID54, NIID102, NIID103, NIID104, NIID105, NIID106, NIID107, NIID108, NIID109, NIID111 (N. gonorrhoeae). C. Southern blotting using the meningococcal ggt gene as a probe. Two micrograms of purified chromosomal DNA digested with ClaI were subjected to this analysis. Lane 1, H44/76 (N. meningitidis ggt+); lane 2, NIID113 (N. meningitidis ggt::IS) [49]; lane 3, NIID54 (N. gonorrhoeae); lane 4, ATCC23970 (N. lactamica); lane 5, ATCC13120 (N. flavescens); lane 6, ATCC14686 (N. denitrificans); lane 7, ATCC25295 (N. elongata); lane 8, ATCC14687 (N. canis); lane 9, ATCC14685 (N. cinerea); lane 10, NIID16 (N. mucosa); lane 11, NIID17 (N. sicca).
Figure 2 A. Split graph showing the relationships among ggt genes in 7 meningococcal strains and ggh genes in 11 gonococcal strains. The sequence data have been submitted to the DDBJ/EMBL/GenBank Databases under the following Accession Numbers:N. meningitidis strains H44/76 [DDBJ:AB089320], H114/90 [DDBJ:AB177989], 2996 [DDBJ:AB177990], NIID68 [DDBJ:AB177991], NIID76 [DDBJ:AB177992], NIID413 [DDBJ:AB177993] and NIID414 [DDBJ:AB177994]; N. gonorrhoeae strains ATCC49226 [DDBJ:AB175023], NIID54 [DDBJ:AB175024], NIID102 [DDBJ:AB193248], NIID103 [DDBJ:AB175025], NIID104 [DDBJ:AB175026], NIID105 [DDBJ:AB193249], NIID106 [DDBJ:AB175027], NIID107 [DDBJ:AB175028], NIID108 [DDBJ:AB193250], NIID109 [DDBJ:AB194328], NIID111 [DDBJ:AB193251]. The scale bar represents uncorrected distances, and a fit parameter is also shown. B. Alignment of the nucleotide sequences containing 4 kinds of differences in the ggt and ggh genes of N. meningitidis strains H44/76 and H114/90; N. gonorrhoeae strains ATCC49226, NIID54, NIID103 and NIID106, respectively. Sequence identity is represented as *, polymorphism among the sequences of the 6 strains is indicated by the appropriate letter, and the absence of a base is shown by a hyphen (-). The nucleotide substitution for an ochre (Type II) mutation is shown in bold. Boxes at the Type I and Type III mutations indicate the tandem repeat. The tetranucleotide repeat in type I is also shown as a gray box. The newly deduced start codon of the meningococcal GGT is shown in underlined bold.
Figure 3 A. Positions of stop codons in the hypothetical ORF in the ggh gene. Black squares indicate the positions of stop codons and white squares indicate the ochre (Type II) mutation shown in Figure 2B. * indicates the position of the 7-bp deletion (Type III) mutation in the ggh gene. B. Putative translated products from the corrected nucleotide sequences of the ggh genes. The Type IV insertional mutation shown in Figure 2B is removed and the site of Type III deletion is replaced by the letter B. The site of the ochre (Type II) mutation is shown as X. The identical amino acid sequences between N. meningitidis and N. gonorrhoeae are shown in a black box and amino acid sequences common to only N. gonorrhoeae are shown in a gray box.
Figure 4 Genetic organization of the genes around the ggt and ggh genes in N. meningitidis strain MC58 [21], N. gonorrhoeae strain FA1090 [47] and N. lactamica strain ST-640 [48]. Black arrows and the open bar indicate coding and non-coding genes, respectively.
Figure 5 Transcriptional expression of the ggh genes. A. Dot blot analysis using the meningococcal ggt gene as a probe. One to 16 micrograms of RNAs isolated from H44/76, HT1089 (H44/76 Δggt::spc) [39], ATCC49226, NIID54, NIID103 and NIID106 were subjected to this analysis. B. RT-PCR to detect the transcripts of the gonococcal ggh genes. The schematic figure in the box depicts the position of the primers used in this experiment (see Table 1). RT-PCR was performed without reverse transcriptase (RTase) (lanes 1 to 5) or with RTase (lanes 6 to 10). Lanes 1 and 6, H44/76 (N. meningitidis); lanes 2 and 7, ATCC49226 (N. gonorrhoeae); lanes 3 and 8, NIID54 (N. gonorrhoeae); lanes 4 and 9, NIID103 (N. gonorrhoeae); lanes 5 and 10, NIID106 (N. gonorrhoeae). The marker in the left-most lane is φ X174 DNA digested with HaeIII. Primer sets used for RT-PCR are shown on the left side, and the corresponding PCR products are indicated by arrows on the right side. C. Primer extension analysis to detect the transcriptional start point of the ggt and ggh genes. Total RNA extracted from H44/76 (N. meningitidis), ATCC49226 and NIID106 (N. gonorrhoeae) was used for the primer extension with AMV reverse transcriptase XL and biotin-labeled oligonucleotide ggt-ext-2. The arrow on the right side indicates the transcriptional start site. D. Alignment of the nucleotide sequences of the upstream regions of the ggt and ggh genes. The sequence data have been deposited in the DDBJ/EMBL/GenBank Databases under the following Accession Numbers:N. meningitidis strains H44/76 [DDBJ:AB193252], H114/90 [DDBJ:AB193253], N. gonorrhoeae strains ATCC49226 [DDBJ:AB193254], NIID54 [DDBJ:AB193255], NIID103 [DDBJ:AB193296], NIID106 [DDBJ:AB193256]. An identical nucleotide is represented as *. The transcriptional start site is shown in bold as +1. The putative -35, -10 elements and Shine-Dalgarno sequence (SD) are depicted in the box, and the ideal -35 and -10 nucleotide sequences are shown above the boxes. The previously predicted start codon (ATG) [25] and newly predicted start codon (GTG) of the meningococcal {\it ggt} gene are underlined. The amino acid sequence deduced from the putative start codon GTG (shown in bold) in the meningococcal {\it ggt} gene is also shown under the corresponding nucleotide sequences.
Figure 6 Western blotting with anti-meningococcal GGT rabbit antiserum [25] (A) and Coomassie Brilliant Blue staining (B) of the whole cell extracts after SDS-PAGE. Bacterial whole cell extracts equivalent to 0.025 OD600 were analyzed. Lane 1, N. meningitidis strain H44/76; lane 2, HT1089 (H44/76 Δggt::spc); lane 3, ATCC49226 (N. gonorrhoeae); lane 4, NIID54 (N. gonorrhoeae); lane 5, NIID103 (N. gonorrhoeae); lane 6, HT1195 (NIID103 Δggh::spc); lane 7, NIID106 (N. gonorrhoeae); lane 8, HT1196 (NIID106 Δggh::spc). Black arrows show the bands corresponding to the processed small and large subunits of meningococcal GGT and the gray arrow indicates the 15-kDa band corresponding to the truncated protein product of the gonococcal ggh gene. M stands for molecular weight marker.
Table 1 Oligonucleotides used in this study
Primers for the sequencing of ggt and ggh genes, and RT-PCR (for ggt-9 and ggt-10)
Oligonucleotide name Position in sequence Length (bp) Sequence (5'-3') Reference
ggt-3 *1265–1286 21 GACTGCTGATGACATTAGCGG [49]
ggt-4 *3250–3228 22 GATTACTCACAATTTCCCCCTA [49]
ggt-5 *1791–1811 20 CGATGCGTGCGACGCCGGAA [25]
ggt-6 *2676–2654 23 ATAGCACATTGCCCGCCTTATCC [25]
ggt-7 *2241–2262 22 CAAGATTTATCTGATTATCAAG [25]
ggt-9 *2779–2800 21 GGGCAAACAGGTCGCCAATCG [25]
ggt-10 *2089–2068 21 TGTAGCGGCACACCATTCGGC [25]
ggt-18 *1554–1534 21 CGGTCAGTCCCGTTGCATGTT [25]
Primers for RT-PCR
ggt-29 *1452–1475 24 GGATGTCAAGTCATCCATGCCAAT This study
ggt-20 *1682–1659 24 TGTCGTCTGCACCGCCACCATCGC This study
ggt-31 *1878–1901 24 GGTACGCCTGCTATCCCTAAACTG This study
ggt-22 *3079–3056 24 CGCACATCAGTCTTATAGCCCAAA This study
Primer for primer extension
primer-ext-2 *1492–1467 26 GTATTAACCTTACCTTGATTGGCATG (Biotin-labeled at the 5'-terminus) This study
*Numbers of positions indicate the position from the 5'-nucleotide of the ggt locus in N. meningitidis strain H44/76 [DDBJ:AB175033].
Table 2 Mutation types found in the ggh genes of 11 N. gonorrhoeae strains
Mutation type*
Strains I (+6 bp) II (ochre) III (Δ7 bp) IV (+46 bp)
ATCC49226
NIID102
NIID104 - + + +
NIID108
NIID111
NIID109 (48th A to G)
NIID103 + + + +
NIID54 + + + -
NIID106
NIID107 - + + -
NIID105 (213th G to A)
"+" or "-" denotes the presence or absence of the mutation, respectively.
*Mutation type corresponds to the categories in Figure 2B.
==== Refs
Nassif X Interaction mechanisms of encapsulated meningococci with eucaryotic cells: what does this tell us about the crossing of the blood-brain barrier by Neisseria meningitidis? Curr Opin Microbiol 1999 2 71 77 10047554 10.1016/S1369-5274(99)80012-5
Suzuki H Hashimoto W Kumagai H Glutathione metabolism in Escherichia coli J Mol Catal B 1999 6 175 184 10.1016/S1381-1177(98)00116-7
Chevalier C Thiberge JM Ferrero RL Labigne A Essential role of Helicobacter pylori γ glutamyltranspeptidase for the colonization of the gastric mucosa of mice Mol Microbiol 1999 31 1359 1372 10200957 10.1046/j.1365-2958.1999.01271.x
McGovern KJ Blanchard TG Gutierrez JA Czinn SJ Krakowka S Youngman P γ Glutamyltransferase is a Helicobacter pylori virulence factor but is not essential for colonization Infect Immun 2001 69 4168 4173 11349094 10.1128/IAI.69.6.4168-4173.2001
D'Amato RF Eriquez LA Tomfohrde KM Singerman E Rapid identification of Neisseria gonorrhoeae and Neisseria meningitidis by using enzymatic profiles J Clin Microbiol 1978 7 77 81 203607
Brown JD Thomas KR Rapid enzyme system for the identification of pathogenic Neisseria spp J Clin Microbiol 1985 21 857 858 2860121
Janda WM Ulanday MG Bohnhoff M LeBeau LJ Evaluation of the RIM-N, Gonochek II, and Phadebact systems for the identification of pathogenic Neisseria spp. and Branhamella catarrhalis J Clin Microbiol 1985 21 734 737 3889044
Philip A Garton GC Comparative evaluation of five commercial systems for the rapid identification of pathogenic Neisseria species J Clin Microbiol 1985 22 101 104 3926813
Welborn PP Uyeda CT Ellison-Birang N Evaluation of Gonochek-II as a rapid identification system for pathogenic Neisseria species J Clin Microbiol 1984 20 680 683 6436315
Ramos-Onsins S Aguade M Molecular evolution of the Cecropin multigene family in Drosophila. functional genes vs. pseudogenes Genetics 1998 150 157 171 9725836
Loguercio LL Wilkins TA Structural analysis of a hmg-coA-reductase pseudogene: insights into evolutionary processes affecting the hmgr gene family in allotetraploid cotton (Gossypium hirsutum L.) Curr Genet 1998 34 241 249 9799357 10.1007/s002940050393
Mighell AJ Smith NR Robinson PA Markham AF Vertebrate pseudogenes FEBS Lett 2000 468 109 114 10692568 10.1016/S0014-5793(00)01199-6
Lawrence JG Hendrix RW Casjens S Where are the pseudogenes in bacterial genomes? Trends Microbiol 2001 9 535 540 11825713 10.1016/S0966-842X(01)02198-9
Hall BG Yokoyama S Calhoun DH Role of cryptic genes in microbial evolution Mol Biol Evol 1983 1 109 124 6400646
Schnetz K Rak B Regulation of the bgl operon of Escherichia coli by transcriptional antitermination EMBO J 1988 7 3271 3277 2846278
Parker LL Hall BG Characterization and nucleotide sequence of the cryptic cel operon of Escherichia coli K12 Genetics 1990 124 455 471 2179047
Al Mamun AA Tominaga A Enomoto M Detection and characterization of the flagellar master operon in the four Shigella subgroups J Bacteriol 1996 178 3722 3726 8682772
Al Mamun AA Tominaga A Enomoto M Cloning and characterization of the region III flagellar operons of the four Shigella subgroups: genetic defects that cause loss of flagella of Shigella boydii and Shigella sonnei J Bacteriol 1997 179 4493 4500 9226258
Tominaga A Mahmoud MA Mukaihara T Enomoto M Molecular characterization of intact, but cryptic, flagellin genes in the genus Shigella Mol Microbiol 1994 12 277 285 8057852
Andersson JO Andersson SG Pseudogenes, junk DNA, and the dynamics of Rickettsia genomes Mol Biol Evol 2001 18 829 839 11319266
Tettelin H Saunders NJ Heidelberg J Jeffries AC Nelson KE Eisen JA Ketchum KA Hood DW Peden JF Dodson RJ Nelson WC Gwinn ML DeBoy R Peterson JD Hickey EK Haft DH Salzberg SL White O Fleischmann RD Dougherty BA Mason T Ciecko A Parksey DS Blair E Cittone H Clark EB Cotton MD Utterback TR Khouri H Qin H Vamathevan J Gill J Scarlato V Masignani V Pizza M Grandi G Sun L Smith HO Fraser CM Moxon ER Rappuoli R Venter JC Complete genome sequence of Neisseria meningitidis serogroup B strain MC58 Science 2000 287 1809 1815 10710307 10.1126/science.287.5459.1809
Andersson JO Andersson SG Genome degradation is an ongoing process in Rickettsia Mol Biol Evol 1999 16 1178 1191 10486973
Zhu P Morelli G Achtman M The opcA and ψopcB regions in Neisseria: genes, pseudogenes, deletions, insertion elements and DNA islands Mol Microbiol 1999 33 635 650 10417653 10.1046/j.1365-2958.1999.01514.x
Feavers IM Maiden MC A gonococcal porA pseudogene: implications for understanding the evolution and pathogenicity of Neisseria gonorrhoeae Mol Microbiol 1998 30 647 656 9822829 10.1046/j.1365-2958.1998.01101.x
Takahashi H Watanabe H Post-translational processing of Neisseria meningitidis γ-glutamyl aminopeptidase and its association with inner membrane facing to the cytoplasmic space FEMS Microbiol Lett 2004 234 27 35 15109716 10.1016/j.femsle.2004.03.003
Reyes A Mezzina M Gadaleta G Human mitochondrial transcription factor A (mtTFA): gene structure and characterization of related pseudogenes Gene 2002 291 223 232 12095695 10.1016/S0378-1119(02)00600-5
Asakura N Nakamura C Ishii T Kasai Y Yoshida S A transcriptionally active maize MuDR-like transposable element in rice and its relatives Mol Genet Genomics 2002 268 321 330 12436254 10.1007/s00438-002-0737-7
Moreau-Aubry A Le Guiner S Labarriere N Gesnel MC Jotereau F Breathnach R A processed pseudogene codes for a new antigen recognized by a CD8(+) T cell clone on melanoma J Exp Med 2000 191 1617 1624 10790436 10.1084/jem.191.9.1617
Hirotsune S Yoshida N Chen A Garrett L Sugiyama F Takahashi S Yagami K Wynshaw-Boris A Yoshiki A An expressed pseudogene regulates the messenger-RNA stability of its homologous coding gene Nature 2003 423 91 96 12721631 10.1038/nature01535
Lees-Murdock DJ McLoughlin GA McDaid JR Quinn LM O'Doherty A Hiripi L Hack CJ Walsh CP Identification of 11 pseudogenes in the DNA methyltransferase gene family in rodents and humans and implications for the functional loci Genomics 2004 84 193 204 15203217 10.1016/j.ygeno.2004.02.004
Korneev SA Park JH O'Shea M Neuronal expression of neural nitric oxide synthase (nNOS) protein is suppressed by an antisense RNA transcribed from an NOS pseudogene J Neurosci 1999 19 7711 7720 10479675
Dehio C Gray-Owen SD Meyer TF The role of neisserial Opa proteins in interactions with host cells Trends Microbiol 1998 6 489 495 10036728 10.1016/S0966-842X(98)01365-1
Nassif X Pujol C Morand P Eugene E Interactions of pathogenic Neisseria with host cells. Is it possible to assemble the puzzle? Mol Microbiol 1999 32 1124 1132 10383754 10.1046/j.1365-2958.1999.01416.x
Baranov PV Hammer AW Zhou J Gesteland RF Atkins JF Transcriptional slippage in bacteria: distribution in sequenced genomes and utilization in IS element gene expression Genome Biol 2005 6 R25 15774026 10.1186/gb-2005-6-3-r25
Mira A Pushker R The Silencing of Pseudogenes Mol Biol Evol 2005 PIMD 16014873
Hart CA Rogers TR Meningococcal disease J Med Microbiol 1993 39 3 25 8100858
van Deuren M Brandtzaeg P van der Meer JW Update on meningococcal disease with emphasis on pathogenesis and clinical management Clin Microbiol Rev 2000 13 144 166 10627495
Tunkel AR Scheld WM Pathogenesis and pathophysiology of bacterial meningitis Clin Microbiol Rev 1993 6 118 136 8472245
Takahashi H Hirose K Watanabe H Necessity of meningococcal γ-glutamyl aminopeptidase for the Neisseria meningitidis growth in rat cerebrospinal fluid (CSF) and CSF mimicking medium J Bacteriol 2004 186 244 247 14679245 10.1128/JB.186.1.244-247.2004
Takahashi H Kuroki T Watanabe Y Dejsirilert S Saengsuk L Yamai S Watanabe H Reliability of the detection of meningococcal γ-glutamyl transpeptidase as an identification marker for Neisseria meningitidis Microbiol Immunol 2004 48 485 487 15215623
Takahashi H Kuroki T Watanabe Y Tanaka H Inouye H Yamai S Watanabe H Characterization of Neisseria meningitidis isolates collected from 1974 to 2003 in Japan by multilocus sequence typing J Med Microbiol 2004 53 657 662 15184538 10.1099/jmm.0.45541-0
Takahashi H Watanabe H A broad-host-range vector of incompatibility group Q can work as a plasmid vector in Neisseria meningitidis: a new genetical tool Microbiology 2002 148 229 236 11782515
phylogenic tree prediction
Huson DH SplitsTree: analyzing and visualizing evolutionary data Bioinformatics 1998 14 68 73 9520503 10.1093/bioinformatics/14.1.68
Takahashi H Inada T Postma P Aiba H CRP down-regulates adenylate cyclase activity by reducing the level of phosphorylated IIA(Glc), the glucose-specific phosphotransferase protein, in Escherichia coli Mol Gen Genet 1998 259 317 326 9749675 10.1007/s004380050818
Neisseria gonorrhoeae genome sequencing strain FA1090
The Sanger Institute: Neisseria lactamica
Takahashi H Tanaka H Inouye H Kuroki T Watanabe Y Yamai S Watanabe H Isolation from a healthy carrier and characterization of a Neisseria meningitidis strain that is deficient in γ-glutamyl aminopeptidase activity J Clin Microbiol 2002 40 3035 3037 12149372 10.1128/JCM.40.8.3035-3037.2002
|
16202144
|
PMC1262726
|
CC BY
|
2021-01-04 16:03:40
|
no
|
BMC Microbiol. 2005 Oct 4; 5:56
|
utf-8
|
BMC Microbiol
| 2,005 |
10.1186/1471-2180-5-56
|
oa_comm
|
==== Front
BMC OphthalmolBMC Ophthalmology1471-2415BioMed Central London 1471-2415-5-231620217310.1186/1471-2415-5-23Case ReportCitalopram associated with acute angle-closure glaucoma: case report Croos Robert [email protected] Srinivasa [email protected] Sabit [email protected] Jane Da Roza [email protected] Department of Psychiatry, Prospect Park Hospital, Reading, RG30 4EJ, UK2 Department of Ophthalmology, Royal Berkshire Hospital, Reading, RG1 5AN, UK3 Department of Psychiatry, Marlborough House Regional Secure Unit, Milton Keynes Hospital Site, Eaglestone, Standing Way, Milton Keynes, MK6 5NG, UK2005 4 10 2005 5 23 23 3 2 2005 4 10 2005 Copyright © 2005 Croos et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Acute angle-closure glaucoma is a rare complication in patients receiving anti-depressant treatment. In the following case, we report the development of acute angle closure glaucoma in a patient who overdosed on Citalopram, an antidepressant, and discuss the possible etiological mechanisms for the condition.
Case presentation
We report a 54 year old, Caucasian lady, with depression and alcohol dependence syndrome, who developed acute angle-closure glaucoma after an overdose of Citalopram, along with alcohol. She was treated with medications and had bilateral Yag laser iridotomies to correct the glaucoma and has made complete recovery. In this case, the underlying cause for glaucoma appears to be related to the ingestion of Citalopram.
Conclusion
The patho-physiological basis for acute angle closure glaucoma in relation to antidepressant medications remains unclear. The authors suggest Citalopram may have a direct action on the Iris or Ciliary body muscle through serotonergic or anti-cholinergic mechanisms or both. This case highlights the importance of the awareness of the underlying risks, which may predispose an individual to develop acute angle-closure glaucoma, and reminds the clinicians the significance of history taking and examination of the eye before and after starting anti-depressants. This area needs to be further researched.
==== Body
Background
Depression is the most common psychological disorder in the world. The prevalence of unipolar depression is estimated to be between 3% and 13%, with as much as 20% of the world adult population experiencing at least some depressive symptoms at any given time [1]. Lifetime incidence of depression is estimated to be 20% to 55%. Approximately, 70% of moderately to severely depressed patients respond to anti-depressant therapy [2]. SSRI are increasingly the first line choice of anti-depressant because of their tolerable side-effect profile and low rate of lethality if taken in an overdose [3]. All SSRI are equally effective in treatment for depression [4].
Citalopram is an antidepressant of the selective serotonin reuptake inhibitor (SSRI) class. They act by producing a gradual increase in postsynaptic levels of serotonin (5-hydroxytryptamine, 5-HT) via desensitization of the feedback systems that controls the rate-limiting enzyme in 5-HT synthesis [5].
Serotonin (5-HT) receptors have been shown to be present in human eyes [6]. Furthermore, it is reported that Serotonin (5-HT) receptors are present at a higher concentration in mammalian ciliary body and cornea than in non-mammalian species [7]. Experimental studies have shown that topical application of serotonin increases the Intra-ocular pressure (IOP) in rabbit's eyes, and that 5-caboxamidotryptamine, a 5-HT 1a receptor agonist, is even more effective than 5-HT itself in elevating IOP [8]. Similarly, in a study of 20 consecutive depressed patients, following of a single dose of 20 mgs Fluoxetine it was shown to increase IOP by 4 mmHg [9]. In another study, Ketanserin, a compound with serotonergic blocking properties, reduced IOP in both animals and humans stressing the role of exerted by 5-HT on IOP [10].
Glaucoma is defined as a heterogeneous group of diseases that have in common a characteristic optic neuropathy and visual defects, for which elevated IOP is the primary risk factor [11]. There are approximately 67 million persons, worldwide, who suffer from glaucomatous disease of the eye [12]. These figures may not include the drug-induced glaucoma's because the precise information on the incidence of glaucoma as a result of local or systemic therapies is uncertain [13]. Angle-closure glaucoma is a disease with acute onset that occurs in 1 of 1000 Caucasians, about 1 in 100 Asians (especially mongoloids) and Hispanics, and 2–4 of 100 Inuit's (Eskimos) [13].
Risk factors for angle-closure glaucoma are narrow angle of anterior chamber, shallow anterior chamber depth, hyperopic, small eyes, positive family history of angle closure, elderly, female sex and use of medications that cause papillary dilatation and excitatory situations [14]. Drugs that cause or exacerbate angle-closure glaucoma include several classes of drugs including adrenergic agonists, cholinergics, anti-cholinergics, sulpha-based drugs, selective serotonin reuptake inhibitors, tricyclic and tetra cyclic antidepressants, anticoagulants and HI and H2 receptor antagonists, especially in people predisposed with narrow angles of anterior chamber. In some instances, bilateral involvement and blindness have occurred [11].
The patho-physiological mechanism of SSRI induced acute angle-closure glaucoma remains unclear, even though anti-cholinergic adverse effects or increased levels of serotonin, which cause partial papillary dilatation, have been implicated [11].
Case presentation
We describe the case of a 54 year old, non-smoker, Caucasian woman, a computer programmer, who was admitted to the General Hospital in June 2003, following an episode of overdose with Citalopram and alcohol. At the time of her admission, she gave a history of depression and suicidal ideation for six months. She was not known to the local psychiatric service.
The patient was discovered by her twin sister soon after the overdose. Initially, she had disclosed to the medical doctor admitting her that she had taken approximately 14 tablets of 20 mgs of Citalopram along with 2 bottles of red wine. However, later on she informed us that she might have taken up to 30 tablets of 20 mgs Citalopram. The actual amount ingested remains unclear.
Soon after her admission, she complained of painful left eye with blurred vision, and was seen by the Ophthalmologist, who found that our patient had an intra-ocular pressure of 23 mmHg in the right eye and 60 mmHg of mercury in the left eye (Normal IOP-10–20 mm Hg), with left corneal edema, and fixed dilated pupils. She was not hypermetripic and had averaged sized eyes. She was noted to have shallow anterior chambers (central and peripheral depth not available) in both of her eyes. A diagnosis of left angle-closure glaucoma was made and medications were commenced to reduce the elevated IOP. Further investigations including routine blood investigations revealed no abnormalities. Her pulse rate and blood pressure were normal. She had a blood alcohol level of 85 mgs/dl (Less than 10 indicates safe levels and 50–100 indicates toxic levels). Her blood test did not reveal any detectable levels paracetamol or salicylates.
With regard to her background history, there was previous episode of overdose with paracetamol tablets in 1978. There was no history suggestive of physical illness, and specifically, no history of previous eye problems. There was no family history of eye related conditions. In December 2002, her general practitioner had diagnosed her to be suffering from depression with harmful misuse of alcohol, and commenced her on Citalopram 20 mgs daily. Later, she informed us that she had not taken any of this prescribed medication prior to her overdose and had only been taking Estrogens, given for hormone replacement (Prempak-C).
Approximately 48 hours after the overdose, on examination of her eyes, she was found to have left subhyaloid and retinal hemorrhages. After 72 hours, the visual acuity in the right eye was 6/9 and hand movements in the left eye. The intraocular pressures were reasonably controlled and she had bilateral Yag laser iridotomies. Subsequently, she was discharged to the local psychiatric unit as she continued to express suicidal thoughts and was low in mood. Apart from receiving treatment for her eye problems at the general hospital, she was commenced on a reducing regime of chlordiazepoxide for her alcohol dependence. On discharge from the general hospital, she was advised to continue on Timolol, pilocarpine and dexamethasone eye drops for further 14 days.
She remained free of anti-depressants and her mood improved gradually in the absence of alcohol. After 14 days, her visual acuity was noted to be 6/18 in the left eye, which improved to 6/12 on pinhole. Her ocular condition was noted to be stable. Her ocular pressures were 14 mmHg in the right eye and 15 mmHg in the left eye, but still had retinal hemorrhages on the left eye, with a clearing vitreous hemorrhage.
She was followed-up by the Ophthalmologists and discharged from the psychiatric hospital after 4 weeks with further community support.
In August 2004, her right eye visual acuity was noted to be 6/9, which improved to 6/6 with pinhole and her left eye vision was 6/24, which did not improve with pinhole. She had a left afferent papillary deficit. The pre-retinal and retinal hemorrhages had cleared, but she still had some residual vitreous hemorrhage. Her right eye Intra-ocular pressure was 18 mmHg and left was 19 mmHg without any treatment. She is reviewed by the local eye clinic every six months.
Discussion
Tricyclic antidepressants, such as, amitriptyline and imipramine have been associated with acute angle closure glaucoma [[15] &[16]]. Although there are eight reports of glaucoma and amitriptyline, there is only one reported case of association of amitriptyline with acute angle closure glaucoma and this had taken place following an episode of overdose [15]. In another study, four patients with narrow angles suffered from acute angle closure glaucoma after routinely prescribed doses of imipramine [16]. Bilateral acute angle closure glaucoma has been reported in a hypermetripic patient on venlafaxine and chlorpromazine, suggesting, perhaps, this occurred by the hepatic inhibition of chlorpromazine metabolism of venlafaxine, increasing anticholinnergic activity, or by a direct effect of venlafaxine on the eye, unrelated to mydriasis [17]. The manufacturers of venlafaxine (Wyeth Laboratories) report glaucoma as a rare adverse event. There has been one previous report of increased intraocular pressure in two patients with known narrow angle glaucoma who began taking venlafaxine [18].
Bilateral angle closure glaucoma and visual loss was precipitated by maprotiline and alprazolam in a 71 years old lady with a history of depression, who had previously complained of ocular pain and blurred vision [19]. Fluoxetine, Paroxetine and fluoxamine have been associated with angle closure glaucoma [[11] &[20]]. Voluntary reporting of suspected adverse events with Fluoxetine has identified a total of 63 cases of glaucoma in an estimated patient population of 21 million in 1998 [21]. The manufacturers of Paroxetine are aware of four cases of acute angle closure glaucoma, and one of raised IOP in a UK patient population of over a million in 1998 [21]. The manufacturers of Citalopram are aware of 15 reports of glaucoma, but causality has not been assigned in these cases and there is no published literature concerning glaucoma as a recognized adverse effect after overdose with Citalopram (Lundbeck Ltd, personal communication).
To our knowledge, this is the first report of acute, unilateral, angle closure glaucoma in a patient following an overdose of Citalopram antidepressant. In our case, there was a clear temporal link between Citalopram overdose and the development of acute angle closure glaucoma. It may be that the onset of glaucoma may be due to the rapid rise in blood concentration of the drug after the overdose (serum concentration of Citalopram was not measured in our case). It is also possible that our patient was predisposed to develop glaucoma due some other unknown inherent factor. Although, it would have been ideal to reintroduce our patient to Citalopram in order to demonstrate causality, especially as our patient had surgical intervention in both of her eyes, we decided this was not in the best interest our patient as the risk of another episode of glaucoma would be unacceptable.
If there is a causal relationship, it could be proposed that Citalopram may directly act on the iris or ciliary body muscle through serotonergic or cholinergic mechanisms or both.
We think our case demonstrates that the need to prescribe Citalopram and other SSRI with caution, especially in older female patients, anatomically predisposed individuals, glaucoma patients and those with a family history of glaucoma. We would like to suggest that careful consideration should be given to include history taking and fundoscopic examination before and after starting SSRI in depressed patients.
We think this area merits further investigation and colleagues should continue to report cases of glaucoma or raised IOP to relevant Drugs safety committees.
Abbreviations
SSRI – Selective serotonin reuptake inhibitor
5-HT-5-Hydroxytriptyamine
IOP – Intra ocular pressure
Competing interests
RC – None
SH – None
JDRD – None
ST-Has been sponsored to attend meetings by Astra Zeneca, Sanofi-synthelabo, Eli-Lilly, Pfizer, Novartis and Wyeth pharmaceuticals.
Authors' contributions
RC-participated in information gathering, literature search, data analysis, drafting and co-ordination of the case report. SH-Participated by managing the patients ophthalmologic problems and helped to draft the manuscript. JDRD-Involvement in the psychiatric management, co-ordination of the work and co-wrote the Manuscript. ST-conceived the idea, participated in information gathering, literature search, data analysis, and psychiatric management of the case and drafting the final manuscript. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
Written consent was obtained from the patient for the publication of the patient's details.
==== Refs
Kessler RC McGonagle KA Zhao S Nelson CB Hughes M Eshelman S Wittchen H Kendler KS Lifetime and 12 month prevalence of DSM-3R psychiatric disorders in the united states Archives of General Psychiatry 1994 51 8 19 8279933
Lonquist J Sutonen H Syvalahti E Appleberg B Koskinen T Maninikko T Mehtones OP Narala M Shiro S Auvinen J Antidepressant efficacy and quality of life in Depression: A double blind study with moclobomide and Fluoxetine Acta Psychiatrica Scandinavia 1994 89 363 9
Stokes PE Fluoxetine: A five-year review clinical therapeutics 1993 15 216 43 8519034
Nemeroff CB Paroxetine: An overview of the efficacy and safety of a new selective reuptake inhibitor in the treatment of depression Journal of clinical Psychopharmacology 1993 13 10S 17S 8106649
Birley M Moret C Neurological mechanisms involved in antidepressant therapies Clin Neuropharmacol 1993 16 387 400 8221701
Barrett NL Osbourne NN The presence of serotonin (5-HT1) receptors negatively coupled to adenylate cyclase in rabbit and human iris-Ciliary processes Exp Eye Res 1993 57 209 16 8405187 10.1006/exer.1993.1116
Osbourne NN Tobin AB Serotonin accumulating cells in the iris-Ciliary body and cornea of various species Exp Eye Res 1987 44 731 46 3653269
Osbourne NN Serotonin and melatonin in the iris/ ciliary processes and their involvement in intraocular presence Acta Neurobiol Exp 1994 54 57 64
Costagliola C Mastropasqua L Streado L Fluoxetine oral administration increases intraocular pressure Br J Opthalmol 1996 80 678
Costagliola C Inliano G Rinaldi M Effect of topical of Ketanserin administration on intra-ocular pressure Br J Ophthalmol 1993 77 344 8 8318480
Tripathi RC Tripathi BJ Haggerty C Drug – induced glaucoma's: mechanisms and management Drugs safety 2003 26 749 767
Quigley HA Number of people with glaucoma worldwide Br J Opthalmol 1996 80 389 393
Tielsch JM The epidemiology and control of open angle glaucoma. A population based perspective Annu Rev Public Health 1996 17 121 36 8724220
Salmon JF Predisposing factors for chronic angle-closure glaucoma Prog Retin Eye Res 1999 18 121 32 9920501 10.1016/S1350-9462(98)00007-X
Lowe RF Amitriptyline and glaucoma Med J Aust 1966 2 509 510 5923732
Ritch R Krupin T Henry C Kurata F Oral imipramine and acute angle closure glaucoma Arch Ophthalmol 1994 112 67 68 8285896
Bradley NG Sanbrook MC Malourf AJ Agarwal SA Venlafaxine and bilateral acute closure glaucoma Med J Aust 2002 4;176 241
Eke T Bates AK Acute angle closure glaucoma associated with Paroxetine (letter) BMJ 1997 314 1387 9161312
Kadoi C Hayasaka S Tsukamoto E Matsumoto M Hayasaka Y Nagaki Y Bilateral angle closure glaucoma and visual loss precipitated by antidepressant and ant anxiety agents in a patient with depression Opthalmologica 2000 214 360 361 10.1159/000027521
Ahmad S Fluoxetine and glaucoma (letter) Drug intell clin Pharm 1991 25 436
Eke T Carr S Acute glaucoma, chronic glaucoma and serotonergic drugs Br J Ophthalmol 1998 82 976 979 9828790
|
16202173
|
PMC1262727
|
CC BY
|
2021-01-04 16:03:50
|
no
|
BMC Ophthalmol. 2005 Oct 4; 5:23
|
utf-8
|
BMC Ophthalmol
| 2,005 |
10.1186/1471-2415-5-23
|
oa_comm
|
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-1001620215910.1186/1471-2458-5-100Research ArticleBirth outcomes in Colorado's undocumented immigrant population Reed Mary M [email protected] John M [email protected] Caroline [email protected] Catherine [email protected] Alexandra [email protected] Department of Neurology, University of Colorado at Denver and Health Sciences Center, Denver, CO, USA2 Department of Family Medicine, University of Colorado at Denver and Health Sciences Center, Denver, CO, USA3 The Joffit Group, Inc., Denver, CO, USA4 American Indian and Alaska Native Program, Department of Psychiatry, University of Colorado at Denver and Health Sciences Center, Denver, CO, USA2005 4 10 2005 5 100 100 14 3 2005 4 10 2005 Copyright © 2005 Reed et al; licensee BioMed Central Ltd.2005Reed et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The birth outcomes of undocumented women have not been systematically studied on a large scale. The growing number of undocumented women giving birth in the United States has important implications for clinical care and public health policy. The objective of this study was to describe birth outcomes of undocumented immigrants in Colorado.
Methods
Retrospective descriptive study of singleton births to 5961 undocumented women using birth certificate data for 1998–1999.
Results
Undocumented mothers were younger, less educated, and more likely to be single. They had higher rates of anemia, were less likely to gain enough weight, and less likely to receive early prenatal care. They were much less likely to use alcohol or tobacco. Undocumented women had a lower rate of low birth weight (5.3% v 6.5%, P < .001) or preterm infants (12.9% v 14.5%; p = .001). Undocumented women experienced higher rates of labor complications including excessive bleeding (2.3% v 0.8%, p < .001) and fetal distress (8.7% v 3.6%, p < .001).
Conclusion
Undocumented women have lower rates of preterm delivery and low birth weight infants, but higher rates of pregnancy related risk factors. Higher prevalence of some risk factors which are amenable to medical intervention reveals the need for improved prenatal care in this group.
==== Body
Background
Studies of health status and health care in the undocumented immigrant population in the United States have been limited in both number and scope due to the transient and clandestine nature of the population, as well as the lack and inaccessibility of their medical records [1]. There are an estimated 7 million undocumented immigrants in the United States, growing 5% each year [2]. Birth certificate data is recorded for undocumented women who give birth in the United States; however, information regarding the mother's immigration status is not collected. Therefore, it is not possible to describe this population using birth certificate data alone. Previous studies of pregnancy and birth in this population have been limited to chart review in a single hospital setting,[3,4] analysis of proxy populations, like migrant women enrolled in Federal nutritional programs,[5,6] or by identifying the mothers by their country of origin[7].
Although the Personal Responsibility and Work Opportunity Act of 1996 bans undocumented immigrants form receiving most public benefits, they remain eligible for emergency medical services in life-threatening situations. Under the provisions of Emergency Medicaid, undocumented women qualify for coverage of labor and delivery services, provided that they meet all other eligibility requirements (including income). However, payor information is also not available on the birth certificate. Other researchers have successfully used billing data, specifically Medicaid claims, to identify births to women whose delivery services were covered by public funds [8].
Studies of Latin women of Mexican decent have demonstrated that immigrant groups tend to have healthier birth outcomes than would be expected from their socioeconomic profile, and that they are less likely to experience an adverse outcome with their first pregnancy than they are with subsequent pregnancies [9-11]. Studies of maternal characteristics and birth outcomes in migrant workers have shown that these women delay seeking prenatal care and gain significantly less weight during pregnancy than their non-migrant counterparts [6]. Recent studies have demonstrated an increase in incidences of neural tube defects in infants born to certain populations of Hispanic women who live along the Texas-Mexico border [12,13]. There has been no large systematic analysis of birth outcomes for undocumented immigrants.
The purpose of this study was to identify the birth outcomes for all undocumented women in Colorado for a 2-year period. Using Medicaid claims data linked to birth certificate data allowed anonymous identification of the infant birth record and description of maternal risks and birth outcomes in this population.
Methods
This study linked Medicaid data to birth certificate data to isolate the cohort of undocumented women who gave birth in a Colorado hospital in 1998 and 1999. Emergency Medicaid is provided to undocumented non-citizens who are financially eligible for Medicaid. Emergency Medicaid only provides medical coverage for medical emergencies, which includes labor and delivery [14,15]. Emergency Medicaid does not provide coverage for routine prenatal care or out-patient post partum care. Colorado Medicaid adds the letter "J" to the identification number of all enrollees with Emergency Medicaid (EMJ). We obtained Medicaid records for all EMJ labor and delivery claims. Using a combination of non-unique identifiers, we merged the EMJ database with the Colorado Birth Record for the two-year period, 1998 to 1999. We used a previously described match-merge method[16] within SAS statistical software to merge the data sets on the basis of non-unique identifiers available in both datasets (infant's date of birth, date of hospitalization, maternal county of residence, maternal month and year of birth, and maternal country of birth). This method provides a robust dataset without identifying information on any individual patient. This study was reviewed and granted exempt status by the Colorado Multiple Institutional Review Board (Protocol 02-1048).
Sample
All EMJ claims were abstracted from the total Medicaid claims database. All EMJ claims not related to delivery were eliminated (such as claims for preterm labor). The EMJ records were merged with the Colorado Department of Public Health and Environment's Birth Records birth certificate records for 1998–1999 using SAS statistical software's SQL procedure. The birth certificate record includes data on singleton and twin births. All twin and multiple birth certificate records were eliminated from analysis to avoid bias in low birth weight findings. Eighty-six percent of the EMJ delivery claims successfully matched to a single birth certificate. Approximately 1% of EMJ records matched multiple birth certificate records and 13% of EMJ records matched no birth certificate record. Of the 118,904 live singleton birth records, 5961 (5.01%) were matched to EMJ claims. Mexico was given as the mother's country of origin on ninety-three percent of the matched birth certificates.
Measure and outcomes
Information obtained from the Colorado Vital Statistics Birth Record for this cohort was used to characterize known maternal risk factors before and during pregnancy, and birth outcome measures associated with higher risks of infant mortality or morbidity. Maternal risk factors predating pregnancy include maternal age, education level, marital status, and parity. Measures of maternal risks during pregnancy include adequate weight gain, number of prenatal visits, month of pregnancy when prenatal care began, and smoking or alcohol use.
Primary outcome measures included low birth weight, and gestational age. Secondary outcome measures included complications of labor and delivery, method of delivery, maternal medical risks, abnormal conditions of the newborn, and congenital anomalies of the child.
Analysis
Chi-square tests were used for analysis of dichotomous maternal characteristics and pregnancy outcomes between undocumented women and all other women. The Student t-test statistic was used to analyze continuous variables of mother's age, infant's birth weight, and infant's gestational age. The Mantel-Haenzel test of association was used to determine if there was an association between low birth weight or preterm birth and undocumented status, while individually controlling for smoking status, weight gain during pregnancy, and age of the mother, the available variables known to have a significant impact on birth weight.
Results
Maternal characteristics of the undocumented women were significantly different from the population as a whole (Table 1). Undocumented women were more likely to be between 17 and 35 years of age (93% v 87%, p < .001), less likely to have finished high school, and more likely to be unmarried than their counterparts in the general population.
Table 1 Maternal and pregnancy characteristics, Colorado singleton births, 1998–1999
Maternal characteristics Undocumented immigrants N = 5961 (%) All other women N = 112,943 (%) p-value
Demographics
Age <.001
<17 3.1 2.2
17–35 93.4 87.0
>35 3.5 10.8
Education: high school graduate 29.5 82.7 <.001
Not married 34.6 25.3 <.001
Parity 0.02
0 42.7 43.4
1–2 47.5 47.9
>2 9.8 8.7
Behavioral characteristics
Smoking 1.9 11.1 <.001
Alcohol Use 0.3 1.3 <.001
Weight Gain <20 lbs. 23.5 12.7 <.001
Number of Prenatal Visits <.001
<9 47.3 20.1
9–15 49.6 69.3
>15 3.1 9.6
Trimester Care Began <.001
No prenatal care 1.7 1.0
1 52.0 83.3
2 34.3 12.9
3 12.0 2.8
Medical risk factors
Anemia 7.7 2.2 <.001
Lung disease 0.8 0.5 .004
Gestational diabetes 2.7 1.7 <.001
Hypertension 2.9 3.2 0.23
Oligohydramnios 3.8 1.1 <.001
Previous >4000 gms birth 3.1 0.7 <.001
Previous preterm birth 2.1 1.3 <.001
No risk factors 51.5 76.0 <.001
Behavioral characteristics were also significantly different for the two groups. The undocumented women were far less likely to smoke (1.9 % v 11.1 %, p < .001). However, the undocumented women were less likely to have gained an adequate amount of weight during their pregnancy (23.5 % v 12.7 %, p < .001), and were less likely to receive an adequate number of prenatal care visits (47.3% v 20.1%, p < .001).
Undocumented women were less likely to have a primary (first time) C-section than other women in the population, and more likely to have a vaginal birth after C-section (Table 2). Undocumented women also experienced more complications of labor and delivery. They were significantly more likely to have meconium staining, excessive bleeding, precipitous labor, breech presentation, cord prolapse, and fetal distress (p < 0.01).
Table 2 Labor and delivery methods and complications; Colorado singleton births, 1998–1999
Undocumented immigrants N = 5961 (%) All other women N = 112,943 (%) p-value
Method of delivery
Vaginal delivery 80.1 81.7 0.17
Vaginal after C-Section 3.8 2.4 <.001
Primary C-Section 10.1 10.9 0.05
Repeat C-Section 5.1 4.9 0.56
Forceps/vacuum 9.7 8.6 <.01
Complications of delivery
Meconium staining 11.2 4.3 <.001
Excessive bleeding 2.3 0.8 <.00
Premature rupture 1.9 2.3 0.06
Precipitous labor 2.4 1.8 <.01
Malpresentation 3.5 3.0 0.04
Cord prolapse 0.7 0.3 <.001
Fetal distress 8.7 3.6 <.001
No complications 60.1 73.5 <.001
The difference in mean birth weights was not clinically significant (Table 3). Undocumented women were significantly less likely to deliver a low birth weight infant (6.5% general population v 5.3% for undocumented women, p < 0. 001). The mean gestational age was slightly higher for the infants of undocumented women (39.1 weeks v. 38.9 weeks, p < .001). This difference is likely not clinically significant. However, the rate of preterm births was significantly lower among the undocumented group (12.9% v 14.5%, p < .001).
Table 3 Birth outcomes, Colorado singleton births, 1998–1999
Newborn characteristics Undocumented immigrants N = 5961 All other women N = 112,943
Mean birth weight 3268 gms 3250 gms 0.01
Very low birth weight (<1500 gms) 1.1% 1.0% 0.70
Low birth weight (<2500 gms) 5.3% 6.5% <.001
Mean gestational age 39.1 weeks 38.9 weeks <.001
Pre-term births (<37 weeks) 12.9% 14.5% 0.001
1-minute Apgar 7.68 7.73 <.01
5-minute Apgar 8.8 8.9 <.01
1-minute Apgar <5 (%) 7.8% 6.0% <.001
5-minute Apgar <5 (%) 1.0% 0.7% <.01
Abnormal conditions of the newborn 10.0% 7.8% <.001
Congenital defects of the newborn 1.0% 1.2% 0.41
All abnormal conditions of the newborn (infant anemia, birth injury, fetal alcohol syndrome, hyaline membrane disease, seizures, and requirements for assisted ventilation) were collapsed into one category due to small numbers in individual cells. However, undocumented women showed significantly higher percentages than the general population in this combined category (10.0% v. 7.8%, p < .001). We did not find an increased rate of neural tube defects among infants of undocumented women in Colorado.
The Mantel-Haenzel test was used to assess the association of low birth weight and preterm birth with undocumented status, while individually controlling for the effects of smoking status, maternal age, or inadequate weight gain. The Breslow-Day test of the homogeneity of the odds ratio was not significant for any of the three analyses, and the common relative risk was estimated from the common odds ratio given by the Mantel-Haenzel test. Undocumented women were less likely to deliver a low birth weight infant even after controlling for smoking status (RR = 0.88, CI 0.79–0.99), weight gain (RR = 0.71, CI 0.63–0.80), and age (RR = 0.81, CI 0.72–0.90). Undocumented women were less likely to have preterm delivery even after controlling for smoking status (RR = 0.91, C.I. 0.85–0.98), weight gain (RR = 0.81, C.I. 0.75–0.87), and age (RR = 0.89, C.I. 0.84–0.96).
Discussion
In this large statewide cohort of undocumented women we found lower rates of preterm delivery and low birth weight than the general population. Undocumented women had higher rates of maternal medical risks and less prenatal care. They had lower rates of tobacco and alcohol use. This is the first study to evaluate the birth outcomes of such a large, statewide cohort of undocumented women.
A 1992 study of maternal care coordination for migrant women in North Carolina cites several studies that document a high incidence of infant mortality and low-birth weight, as well as delay in seeking prenatal care [4]. The study was limited to one migrant health care center in the state, and the study population was comprised of only 599 female farm workers over a 5 year period. A Centers for Disease Control report on pregnancy related behaviors among migrant farm workers enrolled in a Special supplemental Nutrition Program for Women Infants and Children found that more migrants delayed seeking prenatal care, and gained less weight than non-migrant women [5]. However, prevalence was similar for low birth weight and preterm birth in the two populations. Although this was a four-state study, it was limited to women who were enrolled in the Nutrition program (N = 4840).
Studies of Latin women of Mexican decent have revealed at least two persistent differences in birth outcomes from the general population. In what has become known as the "Epidemiological Paradox", studies have repeatedly shown that some immigrant groups tend to have healthier birth outcomes than would be expected given their socioeconomic profiles.
However, children of immigrants do not seem to enjoy the same positive reproductive outcomes as their parents. This "acculturation effect" suggests that as immigrants become more Americanized, their risk of delivering a low-birth weight infant increases [9-11]. Populations of Mexican decent who reside along the Texas-Mexican border have been shown to have higher than average occurrences of neural tube defects. Furthermore, folic acid supplement in the Mexican-born Hispanic population along the Texas-Mexico border showed only a modest risk reduction in the incidence of neural tube defects [12,13].
Studies that have focused exclusively on the undocumented immigrant women have been limited to chart reviews at a single hospital, or survey data collected after using census data to identify areas likely to have a high immigrant population. One such study reported that fewer than 10% of illegal immigrants ever enroll in Medicaid, an indication of their reluctance to use government programs due to their fear of deportation [7].
Only about half of the undocumented women began prenatal care in their first trimester, as compared to almost 85 percent of the general population. The fact that they are residing in the United States illegally creates an obvious barrier to access to care. Because they often work at low paying jobs, they are frequently subject to substandard living conditions, and delay seeking professional medical advice regarding nutrition or prenatal care. Our study confirms these previous findings. The proportion of undocumented women who gained less than the Institute of Medicine's recommended weight was nearly twice that of the general population of pregnant women [17].
The effect of prenatal care has been measured largely in terms of its influence on low birth weight and preterm birth rates [18]. Our study showed significant differences in the proportions of infants born to undocumented women who experienced abnormal conditions at birth, as well as maternal complications during delivery and medical risk factors for the mothers. Better monitoring of both fetus and mother during pregnancy may reduce the risk of many adverse outcomes in addition to low birth weight and preterm birth. For example, medical assessment of the pregnant woman can detect risk factors such as gestational diabetes, anemia, and acute and chronic lung disease that may be amenable to intervention. Cohort studies of diabetic women have shown that control of the diabetes before conception can reduce the risk of congenital anomalies, which is a much as three times higher than for infants of non-diabetic mothers. The cause of anemia in pregnancy is almost always iron deficiency, and treatment with iron supplementation is generally effective.
One contributing factor to the finding that undocumented women were at reduced risk for delivering infants of low birth weight may be that they are much less likely to smoke or use alcohol. This advantage may be partially outweighed by the likelihood that they are less likely to gain the recommended amount of weight during their pregnancy, and that they postpone seeking prenatal care until later in their pregnancy. Our analysis shows that the undocumented women's reduced risk for delivering a low birth weight or preterm infant persisted after controlling either for maternal age or smoking history. Therefore, the reduced risk for low birth weight and preterm births for this group must be due to some factor (or factors) other than age, tobacco use, and inadequate weight gain.
The lack of prenatal care results in missed opportunities to monitor and prepare for labor and delivery, prepare for potential complications like malpresentation and placenta previa, or detect other pregnancy complication such as fetal anomalies and amniotic fluid abnormalities. Many complications of pregnancy and delivery may be avoided simply by assessing a woman's reproductive history [19]. While the outcome of low birth weight was lower among undocumented women, improved access to prenatal care could address the numerous other risk factors related to poor birth outcomes and might lead to even better outcomes in this population.
The major limitation to this study is the technique of merging distinctly different databases without unique identifiers. There were a significant number of EMJ files that did not match to a birth record and some that matched to more than one birth record. However, there is no clinical reason why matched records might systematically differ from unmatched records. The unmatched and multiple matched records were not included in the analysis. Birth certificate records are often incomplete, particularly in the areas of maternal complications and birth defects. However, birth certificate records are widely available, very accurate in terms of delivery method and birth weight and represent the standard measure for research on birth outcomes. The strength of this study was the creation and analysis of a database of a large cohort of women and infants that has previously gone unstudied. This method could be used in other states to evaluate the birth outcomes among undocumented women and to evaluate the impact of interventions to improve access and pregnancy management. The matching technique we used has been previously used and validated, and our matching rate was well within the range of previous studies [8,16].
Conclusion
This study demonstrated that undocumented women are unique in terms of pregnancy risks among Colorado women, and their infants have characteristics that differ from the general population of newborns in the state. Undocumented women have more favorable birth outcomes despite receiving less prenatal care. Lower smoking rates may explain a large proportion of this difference. Focused care to improve access to early pregnancy care and diagnosis and treatment of common medical conditions such as anemia, gestational diabetes, and lung disease, may further help improve birth outcomes in this group.
Competing interests
The author(s) declare there are no competing interests.
Authors' contributions
John Westfall, MD, MPH, supervised all aspects of the study and assisted with the analysis, interpretation, and review of the manuscript. Catherine Battaglia RN, MS, PhD, conceived of the study and obtained the initial Medicaid billing dataset. Alexandra Fichenscher, MPH, assisted in the design and implementation of the study. Mary Reed, MPH, synthesized the analysis, assisted in the interpretation, and led the writing. Caroline Bublitz, MS performed all statistical analysis and assisted in the interpretation of the findings.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
The authors wish to thank Dr. Dennis Lezotte, Department of Preventive Medicine and Biometrics, University of Colorado at Denver and the Health Sciences Center, for his invaluable guidance in the analysis, and Dr. Dann Milne, University of Colorado at Denver and the Health Sciences Center, and formerly of the Colorado Department of Health Care Policy and Finance, for his assistance and insight on Medicaid administration and policy.
==== Refs
Rust G Health Status of Migrant Farmworkers: A Literature Review and Commentary Am J Public Health 1990 80 1213 1216 2205116
Office of Policy and Planning "Estimates of the Unauthorized Immigrant Population Residing in the United States: 1990 to 2000." Statistics Jan 2003 US Citizenship and Immigration Service Accessed 11/4/03
Siddharthan K Alalasundaram S Undocumented aliens and uncompensated care: Whose responsibility? Am J Public Health 1993 83 410 412 8031329
Lu Mc Lin YG Pretto NM Garite TJ Elimination of public funding of prenatal care for undocumented immigrants in California: a cost/benefit analysis American J or Obstetrics and Gynecology 2000 182 233 9
Larson K McGuire J Watkins E Mountain K Maternal Care Coordination for Migrant Farmworkers Women: Program Structure and Evaluation of Effects on Use of Prenatal Care and Birth Outcome J Rural Health 1992 8 128 133 10119763
CDC Pregnancy-related behaviors among migrant farm workers-Four states, 1989–1993 MMWR Weekly 1997 46 283 286
Berk ML Schur CL Chavez LR Frankel M Health Care use among undocumented Latino Immigrants Health Affairs 2000 51 64 10916960 10.1377/hlthaff.19.4.51
Buescher PA Method of linking Medicaid records to birth certificates may affect infant outcome statistics Am J Public Health 1999 89 564 56 10191803
Fuentes-Afflick E Hessol NA Perez-Stable EJ Maternal birthplace, ethnicity, and low birth weight in California Arch Pediatr Adolesc Med 1998 152 1105 12 9811289
Bueckens P Notzon F Kotelchuck M Wilcox A Why do Mexican Americans give birth to few low-birth-weight infants? Am J Epidemiol 2000 152 347 51 10968379 10.1093/aje/152.4.347
Peak C Weeks JR Does community context influence reproductive outcomes of Mexican origin women in San Diego, California? J Immigrant Health 2002 4 125 36 10.1023/A:1015646800549
Suarez L Hendricks KA Cooper SP Sweeney AM Hardy RJ Larsen RD Neural tube defects among Mexican Americans living on the US-Mexico border: Effects of folic acid and dietary folate Am J Epidemiol 2000 152 1017 23 11117610 10.1093/aje/152.11.1017
Hendricks KA Simpson JS Larsen RD Neural tube defects along the Texas-Mexico border, 1993–1995 Am J Epidemiology 1999 149 1119 1127
Medical Assistance Eligibility Health Care Policy and Financing Staff Manual Volume 8, Section 8.100.53 accessed Feb. 8, 2005
Emergency Medicaid for Non-citizens Department of Health Care Policy and Financing Accessed Feb. 8, 2005
Westfall JM McGloin J Impact of double counting and transfer bias on estimated rates and outcomes of acute myocardial infarction Medical Care 2001 9 459 68 11317094 10.1097/00005650-200105000-00006
Committee to Study the Prevention of Low Birthweight Preventing Low Birth Weight 1985 Washington, DC: National Academy Press
U.S. Congress Office of Technology Assessment Healthy Children: Investing in the Future 1988 Washington, D: US Government Printing Office OTA H-345
Enkin M Keirse MJ Neilson J Crowther C Duley L Hodnett E Hofmeyr J A Guide to Eeffective Care in Pregnancy and Childbirth 2000 New York, NY:Oxford University Press
|
16202159
|
PMC1262728
|
CC BY
|
2021-01-04 16:28:57
|
no
|
BMC Public Health. 2005 Oct 4; 5:100
|
utf-8
|
BMC Public Health
| 2,005 |
10.1186/1471-2458-5-100
|
oa_comm
|
==== Front
BMC Public HealthBMC Public Health1471-2458BioMed Central London 1471-2458-5-1011620737510.1186/1471-2458-5-101Study ProtocolCase management for the treatment of patients with major depression in general practices – rationale, design and conduct of a cluster randomized controlled trial – PRoMPT (Primary care Monitoring for depressive Patient's Trial) [ISRCTN66386086] – Study protocol Gensichen Jochen [email protected] Marion [email protected] Monika [email protected] Heike [email protected] Martin [email protected] Thomas [email protected] Christian [email protected] Heiner [email protected] Josef B [email protected] Ferdinand M [email protected] Institute for General Practice, Chronic Care and Health Services Research University of Frankfurt, Theodor-Stern-Kai 7, 60590 Frankfurt a. M., Germany2 Department of General Practice and Health Services Research, University of Heidelberg, Voβbstr. 2, 69115 Heidelberg, Germany3 Department of Epidemiology, Social Medicine and Health System Research, Hannover Medical School, Carl-Neuberg-Str. 1, 30625 Hannover, Germany4 Institute for Social Medicine, University of Lübeck, Beckergrube 43-47, 23552 Lübeck, Germany5 Center for Integrative Psychiatry – University Hospital Schleswig-Holstein – Campus Kiel, Niemannsweg 147, 24105 Kiel, Germany2005 5 10 2005 5 101 101 5 8 2005 5 10 2005 Copyright © 2005 Gensichen et al; licensee BioMed Central Ltd.2005Gensichen et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Depression is a disorder with high prevalence in primary health care and a significant burden of illness. The delivery of health care for depression, as well as other chronic illnesses, has been criticized for several reasons and new strategies to address the needs of these illnesses have been advocated. Case management is a patient-centered approach which has shown efficacy in the treatment of depression in highly organized Health Maintenance Organization (HMO) settings and which might also be effective in other, less structured settings.
Methods/Design
PRoMPT (PRimary care Monitoring for depressive Patients Trial) is a cluster randomised controlled trial with General Practice (GP) as the unit of randomisation. The aim of the study is to evaluate a GP applied case-management for patients with major depressive disorder. 70 GPs were randomised either to intervention group or to control group with the control group delivering usual care. Each GP will include 10 patients suffering from major depressive disorder according to the DSM-IV criteria. The intervention group will receive treatment based on standardized guidelines and monthly telephone monitoring from a trained practice nurse. The nurse investigates the patient's status concerning the MDD criteria, his adherence to GPs prescriptions, possible side effects of medication, and treatment goal attainment. The control group receives usual care – including recommended guidelines. Main outcome measure is the cumulative score of the section depressive disorders (PHQ-9) from the German version of the Prime MD Patient Health Questionnaire (PHQ-D). Secondary outcome measures are the Beck-Depression-Inventory, self-reported adherence (adapted from Moriskey) and the SF-36. In addition, data are collected about patients' satisfaction (EUROPEP-tool), medication, health care utilization, comorbidity, suicide attempts and days out of work.
The study comprises three assessment times: baseline (T0) , follow-up after 6 months (T1) and follow-up after 12 months (T2).
Discussion
Depression is now recognized as a disorder with a high prevalence in primary care but with insufficient treatment response. Case management seems to be a promising intervention which has the potential to bridge the gap of the usually time-limited and fragmented provision of care. Case management has been proven to be effective in several studies but its application in the private general medical practice setting remains unclear.
==== Body
Background
Depression is a significant burden of illness [1,2]. Most depressed patients are diagnosed and treated by general practitioners [3,4]. Depression is the third most common reason for a primary care consultation [5]. Patients with depression account for 50% higher health care costs patients than patients who are not depressed [6]. When improving primary health care for chronic conditions, a number of problems have to be resolved: discontinuity and fragmentation of the process of care, lack of co-ordination between different providers, and "the tyranny of urgency" [7]. In the management of depression, these deficits lead to frequent interruption or premature termination of drug therapy, which increases the risk of unfavourable depression outcomes [8,9]. Patient-centred approaches with a focus on empowerment and self management have been recommended [10,7,11]. Case Management may be one approach to improve depression care. Case Management has been defined as "taking responsibility for following-up patients; determining whether patients were continuing the prescribed treatment as intended; assessing whether depressive symptoms were improving; and taking action when patients were not adhering to guideline based treatment or when they were not showing expected improvement" [12]. Case Management consists of five essential components: (1) identification of patients in need of services, (2) assessment of individual patient needs, (3) developing a treatment plan, (4) coordination of care, and (5) monitoring outcomes and altering care when favourable outcomes are not achieved [13].
Reviews of controlled trials including our meta-analysis concluded that Case Management improves patient outcomes – with a moderate effect [12,14,15]. However, most of the studies available were conducted in highly organized Health Maintenance Organizations (HMO) setting using central based stuff to run the intervention. However, it still remains unclear what the effects in a peripheral setting as a private general practice based case management for patients with depression will be like.
Methods/Design
Aim of the study
The study examines the efficacy of a GP based case management intervention for patients with depression.
Scientific hypothesis
Case management (intervention) leads to greater reductions in depressive symptoms than usual care (control group). We hypothesize further that case management leads to a greater increase in adherence (medication) and quality of life.
Study design
The study is a (prospective) cluster-randomized two-armed intervention study with the practices being clusters. The design of a cluster randomized study was chosen because with this type of study internal validity (absence of confounders) can be optimized, contamination of interventions associated with patient randomization is not possible.
Sample size
Sample size calculations for cluster randomized trials differ completely from sample size calculations for common RCTs [16,17]. Based on the main outcome parameter depression symptom and the main outcome-assessment instrument (GERMAN-PHQ) [18] we performed a power calculation with the method of Hayes and Bennett [19]. It is provided to record a clinical relevant intervention effect of the Case Management from yet 10% up to the primary size, thus an improvement of the depression score (PHQ-D). Assuming an effect of 35 % (PHQ – Score difference from 18,6 to 12.9) in intervention group and a effect of 25 % in control [20] (PHQ Score – 18.6 to 12.9) we have to detect a minimal difference of 1,86. With an alpha of 5%, a beta of 20%, a SD of 6.1 and a ICC 0.1 we are calculating N = 680 (drop-out 30% included). Every arm of the study (34 practices) has to include 10 Patients.
Recruitment of GPs and randomization
As described above, the GPs form the unit of randomization (cluster). It was decided to address and to include only those practices that have a contract with all German insurances, because 90 % of care provision is covered by this type of practices. About 1600 GPs in the city of Frankfurt and the region near by (federal state of Hessen) were informed by mail about the study and invited to a meeting. After this meeting, 72 GPs gave their written consent to participate in the study. Based on detailed information about the practice and the GP, the inclusion criteria were checked (no exclusive clinical specialization) and two practices had to be excluded due to the inclusion criteria. The 70 GPs were stratified according to the size of the city where the practice is located and then randomized to the intervention group or the control group. The procedure was done by an independent assistant who is not familiar to one of the participating doctors and is not a person of the project team. A randomization protocol was written. (Figure 1)
Figure 1 Case manager inclusion criteria
For the practice nurses who wanted to engage in the study, the following criteria are recommended: formal training as practice nurse and at least one year work experience after completion of formal training, and participation in the study case management work shops.
Patient inclusion criteria
Adult Patients, diagnosed with a Major Depression Episode, aged from 18 to 80 years, capability to give informed consent, sufficient knowledge of the German language and indication for antidepressive treatment (medication and/or psychotherapy) can be included. Diagnostic procedure consists in self-report of depressive symptoms via PHQ-9 [18] and subsequent validation in a clinical interview. Participating practices keep an alphabetic record of their patients who are already on treatment. Five Patients from this list are contacted in consecutive order of appearance in the practice and informed about the option to participate in the study. After checking the inclusion criteria and receiving the informed consent, patients receive the questionnaires. Afterwards, five new patients, fulfilling the same criteria, are to be recruited on a screening-day.
Data collection
After giving their informed and written consent to participate in the study patients will receive the questionnaires. The completed questionnaires will be turned back to the practice where there a collected from a member of the project team when the practice has recruited 10 patients. The envelopes are opened at the university and scanned ("eyes and hands ™ FORMS "-Software, Version 5-2 of Read Soft). A TIF-file is generated out of each questionnaire to avoid any data-manipulation and to have a medium for data storage. The scanned data are transferred into the SPSS files. For documentation and data reporting in publications CONSORT recommendations for cluster randomized trials are considered [21].
Outcome-parameter
Table 1 displays the outcome parameters and additionally used instruments. The primary outcome is the depression score assessed by the PHQ-9 questionnaire. The PHQ is an internationally applied instrument for the screening in primary care of the major mental disorders. The German version is also validated [18]. Secondary outcomes include:
Table 1 Outcomes and instruments
Outcome-Parameter (Patient) Instrument
Primary Outcome
Depression PHQ-9
Secondary outcome
Depression BDI (Beck-Depression-Inventar)
Quality of life SF-36
Adherence to medication Adapted version of Moriskey adherence questionnaire
Health care utilization questionnaire, retrospective chart review
Patient satisfaction Modified EUROPEP
Suicide attempts questionnaire
Days out of work questionnaire
• the Beck-Depression-Inventory (BDI) [22]
• self-reported adherence (adapted from Moriskey) [23]
• Quality of life (SF-36) [24]
• Health Care utilization (referrals to specialists, days in clinic); data retrieved form patients chart
• Patient satisfaction (modified EUROPEP-questionnaire) [25]
• Suicide attempts
• Days out of work
These data will be compiled from patient questionnaires and patients chart review. All instruments are well validated and frequently used in international studies. Assessments are done three times: baseline (T0), follow-up after 6 months (T1) and follow up after 12 months (T2).
Intervention
The practice nurses of the intervention group will be trained to case manager in two work-shops with the following contents: main features of the disorder, communication skills, telephone monitoring and documentation. A follow-up work-shop is scheduled to supervise the telephone-monitoring, and each practice is contacted bimonthly from the project team to advice concerning assessment, documentation and intervention.
The intervention comprises the following aspects:
1. The case manager begins with an introduction session which aims at establishing contact, explaining the patient his/her function, delivering information about the disorder and the self-management tools (education). GPs will also receive a written patient leaflet which provides information about the cause and the treatment possibilities as well as coping strategies and a list of self help books.
2. The case manager contacts the patient for telephone monitoring (10–15 minutes). Based on a structured interview, the case manager will ask the patient first two-weekly (the first two months) and than in intervals of 4 weeks about the status of depression, adherence, side effects of medication and goal-attainment (Depression Monitoring List – DeMol, Gensichen unpublished)
3. After each patient contact, a short report will be given to the family doctor.
In both the intervention and the control group, GPs will receive a summary of evidence based treatments of depression in a primary care setting and these information contain amongst others, the NHG guidelines. There is no implementation strategy in the control group.
Timeframe of the study
The study team has already randomized the 70 GPs who have declared their willingness to participate in the study and who accepted random assignment to the different groups. The study protocol was approved by the ethics commission of the University of Frankfurt (Approval-Nr. E 26/05).
The patient enrollment has already started and up to now, about 220 patient are included who completed baseline assessment.
Description of risks
To our knowledge, serious risks or undesired effects of completing questionnaires are not reported in the literature. There are no specific risks related to the study.
Ethical principles
The study is planned and conducted in accordance with medical professional codex and the Helsinki Declaration of 1996 as well as the German Federal Data Security Law (BDSG).
Patients participate in the study voluntarily. They are informed that they can cancel at any time their participation without disclosing reasons for their cancellation and without negative consequences to their future medical care.
Patient informed consent
Patients receive written and spoken information about the main features of the study; i.e. about potential benefits for their health and potential risks prior to their consent and participation in the study. In case of acceptance, they sign the informed consent sheet.
In case of study discontinuation data will be extinguished, except the patient affirms explicitly the further analysis of his data.
Vote of the ethics committee
The study protocol was approved by the ethics committee of the University of Frankfurt previous to the start of the study in April 25, 2005. Inclusion of patients/participants did not start unless there was a written and unrestricted positive vote of the ethics committee. This vote was received in March 2005.
Data security/disclosure of original documents
The patient names and other confidential information are secured by the medical confidentiality rules and are treated according to German Federal Data Security Law (BDSG). The results of the patient questionnaires are not accessible to the GPs.
All study related data and documents are stored on a protected central server of the Frankfurt University Clinic. Only members of the study team have access to the respective files.
Intermediate and final reports are stored in the office of the Department of General Practice and Health Services Research at the Frankfurt University Clinic.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
JG, MT, MP and MB conceived and performed the study and draft manuscript. HWH and HR developed the data-management. CK performed health economical aspects TR, JA and FMG participated in the study design. All authors read and approved the final manuscript.
Pre-publication history
The pre-publication history for this paper can be accessed here:
Acknowledgements
This study is financed by the German Ministry of Education and Research (BMBF), grant-number DLR 01 GK 0302.
==== Refs
Ormel J von Korff M Ustun T Pini S Korten A Oldehinkel T Common mental disorders and disabilities across cultures. Results from the WHO Collaborative Primary Care Study JAMA 1994 272 1741 1748 7966922 10.1001/jama.272.22.1741
Ormel J Oldehinkel AJ Nolen WA Vollebergh W Psychosocial disability before, during and after a major depressive episode: A 3-wave population-based study of state, scar, and trait effects Archives of General Psychiatry 2004 61 387 392 15066897 10.1001/archpsyc.61.4.387
Kamerow D Research on mental health in primary care settings. Rationale, topics and support Family Practice Research Journal 1986 6 5 11 3455110
Ustun T Sartorius N Mental illness in general health care – an international Study 1995 John Wiley & Sons: Chichester
Shah A The burden of psychiatric disorder in primary care International Review of Psychiatry 1992 4 243 250
Simon G Ormel J Health care costs associated with depression and anxiety disorders in primary care Archives General Psychiatry 1995 52 850 856
Bodenheimer T Wagner E Grumbach K Improving primary care for patients with chronic illness – The chronic care model. Part 2 JAMA 2002 288 1909 1914 12377092 10.1001/jama.288.15.1909
Ahrens B Linden M [Factors of chronification for depression] Faktoren der Chronifizierung von Depressionen Münchner Medizinische Wochenschrift 1991 133 49 50
Fritze J [How long to prescribe antidepressants?] Wie lange sollten Antidepressiva verordnet werden? Psychopharmakotherapie 1997 5 7 25
Coulter A Elwyn G What do patients want from high quality general practice and how do we involve them in improvement? Br J Gen Pract 2002 52 S22 26 12389766
Os van T Ormel J Brink van den R Jenner J Meer van der K Tiemens B Doom W van der Smit A Brink van den W Training primary care physicians improve the management of depression General Hospital Psychiatry 1999 21 168 176 10378110 10.1016/S0163-8343(98)00080-2
Von Korff M Goldberg D Improving outcomes in depression – The whole process of care needs to be enhanced British Medical Journal 2001 323 948 949 11679372
Norris S Nicols P Caspersen C Glasgow R Engelgau M Jack L Isham G Snyder S Carande-Kulis V Isham G Garfield S Briss P McCulloch D The effectiveness of disease and case management for people with diabetes – A systematic review Am J Prev Med 2002 22 15 38 11985933 10.1016/S0749-3797(02)00423-3
Gilbody S Whitty P Grimshaw J Thomas R Educational and organizational interventions to improve the management of depression in primary care: a systematic review JAMA 2003 289 3145 51 12813120 10.1001/jama.289.23.3145
Gensichen J Beyer M Muth C Gerlach FM v Korff M Ormel H Case Management to improve major depression in primary health care – a systematic review Psychological Medicine 2005 in press
Campbell MK Grimshaw J Steen N Sample size calculations for cluster randomised trials. Changing Professional Practice in Europe Group (EU BIOMED II Concerted Action) J Health Serv Res Policy 2000 5 12 16 10787581
Campbell MK Mollison J Grimshaw JM Cluster trials in implementation research: estimation of intracluster correlation coefficients and sample size Stat Med 2001 20 391 399 11180309 10.1002/1097-0258(20010215)20:3<391::AID-SIM800>3.0.CO;2-Z
Löwe B Spitzer RL Zipfel L Herzog W Gesundheitsfragebogen für Patienten (PHQ-D) Komplettversion mit vorläufigem Manual, Fragebogen, Schablonen Pfizer GmbH, Karlsruhe 2001
Hayes RJ Bennett S Sample size calculation for cluster-randomized trials Int J Epidemiol 1999 28 319 26 10342698 10.1093/ije/28.2.319
Banerjee S Shamash K Macdonald A Mann A Randomised controlled trial of effect of intervention by psycho-geriatric team on depression in frail elderly people at home British Medical Journal 1996 313 1058 1061 8898601
Campbell MK Elbourne DR Altman DG CONSORT statement: extension to cluster randomised trials BMJ 2004 328 702 708 15031246 10.1136/bmj.328.7441.702
Beck A Steer R Beck Depressions Inventar (BDI) Bearbeitung der deutschen Ausgabe Hautzinger M, Bailer M, Worall H et Testhandbuch 2 Aufl, Huber Bern 1995
Moriskey D Green L Levine D Current and predictive validity of a Self-reported measure of medication adherence Medical Care 1986 24 67 74 3945130
Bullinger M Kirchberger I Ware J Der deutsche SF-36 Health Survey Übersetzung und psychometrische Testung eines krankheitsübergreifendes Instrumentes zur Erfassung der gesundheitsbezogenen Lebensqualität Zeitschrift für Gesundheitswissenschaften 1995 1 21 36
EuroQOL-Group EuroQOL- a new facility for the measurement of health-related quality of life Health Policy 1990 16 199 208 10109801 10.1016/0168-8510(90)90421-9
|
16207375
|
PMC1262729
|
CC BY
|
2021-01-04 16:28:57
|
no
|
BMC Public Health. 2005 Oct 5; 5:101
|
utf-8
|
BMC Public Health
| 2,005 |
10.1186/1471-2458-5-101
|
oa_comm
|
==== Front
BMC Plant BiolBMC Plant Biology1471-2229BioMed Central London 1471-2229-5-191618149310.1186/1471-2229-5-19Research ArticlePurification and kinetic studies of recombinant gibberellin dioxygenases Lester Diane R [email protected] Andy [email protected] Peter [email protected] Inger [email protected] Institute for Cell and Molecular Biology, Uppsala University, Box 596, 751 24 Uppsala, Sweden2 Rothamsted Research, Harpenden, Herts, AL5 2JQ UK3 Department of Molecular Biology, Swedish University of Agricultural Sciences, Box 590, 751 24 Uppsala, Sweden2005 25 9 2005 5 19 19 27 5 2005 25 9 2005 Copyright © 2005 Lester et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The 2-oxoglutarate-dependent dioxygenases (2ODDs) of gibberellin (GA) biosynthesis have a key role in the metabolism of a major plant hormone. The activity of recombinant GA 2ODDs from many species has been characterised in detail, however little information relates to enzyme purification. Native GA 2ODDs displayed lability during purification.
Results
Two GA 2ODDs were expressed in Escherichia coli and purified to homogeneity. The GA 2-oxidase from Pisum sativum L., PsGA2OX1, was expressed as a glutathione s-transferase (GST) fusion. It was purified in the three steps of affinity chromatography, GST removal and gel filtration. Highly pure PsGA2OX1 was obtained at a yield of 0.3 mg/g of cells. It displayed a Km of 0.024 μM and a Vmax of 4.4 pkat/mg toward [1β,2β,3β-3H3]GA20. The GA 3-oxidase from Arabidopsis thaliana, AtGA3OX4, was expressed as a poly(His)-tagged thioredoxin fusion. It was purified by Immobilised Metal Affinity Chromatography followed by gel filtration. Cleavage of the fusion took place between the two purification steps. Highly pure AtGA3OX4 was obtained at a yield of 0.01 mg/g of cells. It displayed a Km of 0.82 μM and Vmax of 52,500 pkat/mg toward [1β,2β,3β-3H3]GA20.
Conclusion
Fusion tags were required to stabilise and solubilise PsGA2OX1 and AtGA3OX4 during E. coli expression. The successful purification of milligram quantities of PsGA2OX1 enables mechanistic and structural studies not previously possible on GA 2ODDs. A moderate yield of pure AtGA3OX4 requires the further optimisation of the latter stages of the enzyme purification schedule. PsGA2OX1's action in planta as deduced from the effect of the null mutation sln on GA levels in seeds is in agreement with the kinetic parameters of the recombinant enzyme.
==== Body
Background
The 2-oxoglutarate-dependent dioxygenases (2ODDs) of gibberellin (GA) biosynthesis have a key role in plant hormone metabolism. Three principal classes of 2ODDs control the later stages of the pathway where bioactive GA is produced and inactivated (Figure 1). Activation and deactivation of the GA molecule are performed by GA 3-oxidase and GA 2-oxidase enzymes, respectively.
The genes encoding GA 2ODDs have been described from many species. They belong to multigene families which show spatial and temporal patterns of expression [1]. Factors including hormone levels and light have been shown to influence GA 2ODD gene expression [2], suggesting that the enzymes are important in regulating the overall GA pathway. The genetic manipulation of GA levels for agricultural purposes has often involved 2ODD-catalysed steps.
The GA 2ODDs are well suited to in vitro studies because they retain good activity when expressed in Escherichia coli and their natural substrates are mostly known and available in synthetic form. Thus, a reasonable amount of data relates to the activity of recombinant GA 2ODDs, e.g. [3-7]. Their catalysis generally appears to reflect that of the native enzymes.
Unfortunately, the excellent literature on the activity of recombinant GA 2ODDs is accompanied by a lack of progress on their purification and structural characterisation. In native form the enzymes displayed a lability that hindered their complete purification [8-11]. The problems associated with this lability have ostensibly not been overcome through using recombinant expression methods. The sole report on a pure GA 2ODD relates to a native GA 20-oxidase from pumpkin endosperm [12].
Herein, we report the successful purification of a recombinant GA 2-oxidase from Pisum sativum L. (pea) and a GA 3-oxidase from Arabidopsis thaliana (Arabidopsis), findings that represent a significant step for an important group of enzymes.
Results
Expression and purification of the GA 2-oxidase, PsGA2OX1
The pea GA 2-oxidase, PsGA2OX1, accumulated intact and in soluble form in E. coli when expressed as a GST fusion in pGEX. The expression conditions employed were determined to be optimal with respect to incubation temperature and induction conditions. Trials with a GST fusion expression system driven by the stronger T7 promoter did not improve expression levels (results not shown).
Affinity chromatography of GST-PsGA2OX1 from the soluble cell fraction gave a high degree of purification (Figure 2A). Subsequent removal of the GST tag yielded PsGA2OX1 with only two major contaminants (Figure 2A). Gel filtration removed these contaminants without significant loss of PsGA2OX1 and increased its purity to > 95% as estimated from SDS-PAGE analysis (Figure 2B). Pure PsGA2OX1 was obtained in milligram quantities and was concentrated to 10 mg/mL (Figure 2B, Table 1).
PsGA2OX1 was active at all stages of purification (Table 1). EDTA had an inhibitory effect on activity that was removed most effectively by dialysis rather than dilution. However, when crude samples were dialysed, activity diminished, presumably because of enzyme inactivation or degradation. In addition, endogenous components of E. coli lysates are suspected to inhibit GA 2ODDs [7]. Therefore, the activity measurements after the first two stages of purification (Table 1) gave apparent values.
Concentrated highly pure PsGA2OX1 retained activity after storage at -70°C.
Kinetic analysis of PsGA2OX1
The activity of this pure PSGA2OX1 displayed a saturation curve of the Michaelis Menten type in response to increasing concentration of substrate [3H3]GA20 (Figure 3). The KM was calculated to be 0.024 μM and the Vmax 4.4 pkat/mg.
Expression and purification of the GA 3-oxidase, AtGA3OX4
The Arabidopsis GA 3-oxidase, AtGA3OX4, accumulated intact and in soluble form when expressed as a poly(His)-tagged TRX fusion in pET32 to a level that was discernible by SDS-PAGE analysis of the soluble cell fraction (Figure 4A).
During purification AtGA3OX4 tended to precipitate, although precipitation could be minimised during the first step by dilution of the soluble cell lysate and the use of a buffer of pH 6.3 with 10 mM β-ME.
TRX-AtGA3OX4 was purified to a high degree by Immobilised Metal Affinity Chromatography (Figure 4A, Table 2). Cleavage of the TRX tag from AtGA3OX4 was enabled by a thrombin site engineered into the expression construct – the Factor Xa site encoded by pET32 failed to cleave.
AtGA3OX4 was difficult to separate from TRX, remaining strongly associated with it subsequent to thrombin cleavage. The single factor identified that most obviously mitigated this association was β-ME, with low pH further aiding separation. Resolution of the two proteins was achieved by gel filtration in a buffer of pH of 5.8 with 10 mM β-ME. Highly pure AtGA3OX4 was thus obtained in microgram quantities (Figure 4B, Table 2).
Once pure, AtGA3OX4 resisted concentration. The problem was partially overcome by the addition of Triton X-100. However, the maximum concentration of AtGA3OX4 achieved was 1.5 mg/mL and its activity did not tolerate freezing.
Otherwise, AtGA3OX4 was active at all stages of purification, permitting a kinetic study on highly pure AtGA3OX4.
Removal of the TRX tag had no obvious affect on AtGA3OX4 activity levels. Triton X-100 had an inhibitory effect but this seemed to disappear with dilution. The use of low pH buffers during purification did not obviously dampen AtGA3OX4's activity. Indeed, we found AtGA3OX4 displayed 80 % activity when assayed in the phosphate buffer of pH 6.3 used for the first purification step, providing the reducing agent DTT was present (results not shown). However, the buffer of pH 5.8 in the second step was not compatible with activity assays.
In light of the difficulties associated with obtaining AtGA3OX4 in high yield, further purification of the fusion, TRX-AtGA3OX4, was conducted. Milligrams of highly pure TRX-AtGA3OX4 were obtained. The protein was amenable to concentration (Figure 4B) and could be frozen at -70°C without activity loss.
Kinetic analysis of AtGA3OX4
The activity of highly pure AtGA3OX4 displayed a saturation curve of the Michaelis Menten type in response to increasing concentration of [3H3]GA20 (Figure 5). The KM was calculated to be 0.82 μM and the Vmax 52, 500 pkat/mg.
Discussion
High levels of activity are frequently obtained for GA 2ODDs in E. coli expression systems in the absence of accumulated soluble enzyme [13]. The instability generally reported in native 2ODDs appears to be a feature of the recombinant enzymes. In early trials of GA 2ODDs, the pea GA 2-oxidase, PsGA2OX1, and the Arabidopsis GA 3-oxidase, AtGA3OX4, showed promise as candidates for purification (D.R. Lester, unpublished). In both cases, a N-terminal fusion tag was needed to obtain soluble intact protein.
The purification of the pea GA 2-oxidase, PsGA2OX1, was straightforward. No problems were encountered during chromatography, tag removal or enzyme concentration. All steps of purification were performed in a buffer that preserved enzyme activity. The pH was within physiological range when the chelator EDTA and the reducing agent DTT were present. (The latter were considered desirable because PsGA2OX1 is Fe++-dependent and its active site is likely to be susceptible to metal-catalysed oxidation). The milligram amounts of PsGA2OX1 that we obtained are unprecedented for a highly pure GA 2ODD.
The gene encoding PsGA2OX1 is affected by the null mutation, sln, which causes the accumulation of GA20 within maturing pea seed [14,15]. The mutation has a lesser effect on GA20 levels in the stem, presumably because other GA 2-oxidases compensate for its loss in this tissue. Mature seeds contain at least two GA 2-oxidases [11,14]. The sln phenotype indicates that PsGA2OX1 is the main catalyst for the conversion GA20 to GA29 at latter stages of seed maturation, when 2-oxidation of GA20 takes place at high levels. Although the sln mutation is not deleterious to developing seeds, the mature seeds contain abnormally high amounts of GA20, which, after germination, is converted to the bioactive GA1 resulting in the slender (overgrowth) phenotype of the seedling [16]. The levels of GA20present in mature SLN and sln seeds are 11 ng/g and 4,118 ng/g, respectively [16]. The difference can be solely attributed to the action of PsGA2OX1 in planta.
Multiple GA 2-oxidases have been previously partially purified in low yield from pea seeds of unspecified maturity [11]. The incomplete purification was attributed to a low abundance of the enzymes in the tissue, and instability during purification. It can be assumed that native PsGA2OX1 was among these GA 2-oxidases because its gene is expressed throughout seed maturation [15]. Recombinant expression methods have now enabled us to complete the successful purification of PsGA2OX1.
At 0.024 μM, the KM of recombinant PsGA2OX1 toward GA20 was much lower than that obtained for the native pea seed GA 2-oxidases (1.55 μM) [11]. The comparison of KM results is complicated, however, because the native preparation contained multiple activities. Certainly, a high substrate affinity of PsGA2OX1 is consistent with its ability to clear the maturing seed of GA20.
The gene encoding PsGA2OX1 is expressed relatively strongly throughout the latter weeks of pea seed maturation [15]. That the enzyme is abundant in mature seeds was supported by a study of GA 2-oxidases from this tissue in the related species Phaseolus vulgaris [8]. If present in abundance, it is plausible that PsGA2OX1 (Vmax of 4.4 pkat/mg) could account for the difference in GA20 levels between SLN and sln seeds.
The Vmax of the partially pure native GA 2-oxidases toward GA20 was 24.8 pmol/h mg (0.0068 pkat/mg) [11]. The higher figure of the recombinant enzyme suggests that its purity and/or retention of activity was enhanced compared with the native preparation. Interestingly, the Vmax of PsGA2OX1 was several orders of magnitude lower than the Vmax of AtGA3OX4 toward GA20 (Figure 5). However, this difference is difficult to interpret considering the two enzymes are from different species and are likely to be expressed differentially within the plant alongside isoenzymes.
With their role in GA activation, the GA 3-oxidases are particularly interesting to study, but in native form the enzymes are difficult to purify. It required 6,661 immature seeds for Kwak et al. [9] to obtain 0.086 mg of enzyme described as partially pure. Recombinant methods enabled us to partially overcome the problems associated with purification of a recombinant GA 3-oxidase. Our experience provides valuable insight for future work.
The first step we used in the purification of the Arabidopsis GA 3-oxidase, AtGA3OX4, was IMAC chromatography. This method does not permit the inclusion of DDT or EDTA in buffers and therefore has been considered unsuitable for purifying 2ODDs [17]. Nonetheless, AtGA3OX4 with high specific activity was obtained using this method. Evidently, 10 mM β-ME was able to protect the enzyme active site from metal-catalysed oxidation during chromatography. The result suggests that this widely used and highly efficient purification strategy does not necessarily compromise the activity of a 2ODD.
AtGA3OX4 showed a tendency to precipitate, one that became more pronounced when it was cleaved, but not physically separated, from TRX. This diminished the yield of the protein, particularly in the final step. A beneficial effect from low pH and β-ME implied that an exposed cysteine was causing the precipitation. However, a combination of factors were probably to blame because the aforementioned conditions did not fully overcome the problem. Furthermore, replacement by site-directed mutagenesis of the cysteine predicted as most likely to be exposed (residue 9) brought no improvement. Certainly, AtGA3OX4 displayed hydrophobic behaviour when pure, sticking to tubes unless Triton X-100 was present. Hydrophobicity was also a feature of the GA 3-oxidase of bean [9].
Good scope exists for further optimisation of the AtGA3OX4 purification method. Efficient separation of AtGA3OX4 and TRX will lead to a greatly increased yield. Hydrophobic interaction chromatography is a possible alternative to gel filtration in separating the two proteins.
AtGA3OX4 was amenable to concentration and freezing in the presence of TRX, whether or not it was chemically attached to it. The identification of buffer ingredients that replace TRX's stabilising role, such as alternative detergents, salts or organic solvents may improve the behaviour of purified AtGA3OX4.
AtGA3OX4 is encoded by a member of a family of four GA 3-oxidase genes in Arabidopsis. Generally the non-13-hydroxylated GA pathway predominates in Arabidopsis, however recent studies found comparable levels of non-13-hydroxylated and 13-hydroxylated GAs in developing seeds [18]. This tissue is the main site of expression of the gene encoding AtGA3OX4 (AtGenExpress microarray data), therefore GA20 is likely to be a natural substrate of AtGA3OX4. The high Vmax gained here shows that AtGA3OX4 can efficiently metabolise GA20. The value of 52,500 pkat/mg is higher than the Vmax of 62 nmol/min mg (1,033 pkat/mg) reported for a purified GST fusion of another Arabidopsis GA 3-oxidase, AtGA3OX1 [7]. AtGA3OX4's KM of 0.82 μM toward GA20 was lower than that of AtGA3OX1 (10 μM) [6,7], yet higher than the value for the bean seed GA 3-oxidase (0.29 μM) [9]. Accumulated kinetic information with respect to both 13-hydroxylated and non-13-hydroxylated GAs on each of the GA 3-oxidases of Arabidopsis will give valuable insight into the control of active hormone levels in this model species.
Conclusion
The previous lack of reports on the complete purification of GA 2-oxidases and GA 3-oxidases is probably due to a combination of their lability and their low abundance in most plant tissues. One might speculate that a crucial role in the control of active hormone levels requires a short enzyme half-life. This study on PsGA2OX1 and AtGA3OX4 shows that GA 2-oxidases and GA 3-oxidases are not necessarily inherently unstable and can be purified active from E. coli expression systems under standard conditions. Indeed, the GA 2-oxidase, PsGA2OX1, was exceptionally stable throughout purification. Problems were encountered with the Arabidopsis GA 3-oxidase, AtGA3OX4, due to its hydrophobicity and the presence of hydrostatic charges, rather than metal-catalysed autocleavage or proteolytic breakdown.
Indeed, plant 2ODDs from secondary metabolic pathways are also problematic during purification [17,19]. Yet in some cases, they have been purified in high yield permitting detailed structural and mechanistic studies [20,21]. We note the recent elucidation of the structure of the labile 2ODD-related enzyme, ACC-oxidase, of ethylene biosynthesis [22]. Our work opens the door to similar studies on the GA 2ODDs.
Methods
Expression of PsGA2OX1
The cDNA PsGA2ox1 (GenBank AF100954) was ligated into the EcoRI site of the tac promoter-based E. coli expression vector pGEX4T2 (Amersham Biosciences).
A culture of 100 mL LB broth with 100 μg/mL ampicillin was inoculated with E. coli BL21 (DE3) pGEX4/PsGA2ox1 and grown overnight at 37°C with shaking. Thirty mL of starter culture was used to inoculate each of 3 flasks containing 1 L of LB broth with 100 μg/mL ampicillin. Flasks were shaken at 220 rev/min on a circular orbit of 2.54 cm at 37°C. Protein expression was induced when cells reached the OD600 of 1.0. After 2 h cells were harvested, washed with 1X PBS and frozen at -70°C.
Purification of PsGA2OX1
The GST-PsGA2OX1 purification was performed with glutathione agarose beads (Amersham Biosciences) at 4°C according to the manufacturer's instructions with the following additional details. The buffer used for the lysis of cells and washing of beads was 50 mM Tris-HCl pH 8, 300 mM NaCl, 5 mM EDTA, 5 mM DTT and 0.4 mM PMSF. Cells (approximately 6 g) were resuspended in 120 mL of buffer and lysozyme was added to 0.1 μg/mL. After incubation at room temperature for 15 min, cells were lysed with a Vibra Cell sonicator at 65% amplitude for a total of 60 sec. The cell lysate was spun at 20,000 g and then GST-PsGA2OX1 was extracted from the supernatant using a bed volume of 2 mL of beads in batch mode. After a final wash, the beads were resuspended in 2 mL of buffer containing 10 U of bovine thrombin (Sigma) and incubated at 4°C overnight. Thrombin was extracted from the supernatant containing cleaved PsGA2OX1 using benzamidine agarose beads (Amersham Biosciences).
The protein solution was concentrated to a volume of 300 μL using a Vivascience spin concentrator and applied to a Superdex 75 column (Amersham Biosciences). Gel filtration was performed at 6°C in the same buffer as was used for affinity chromatography. After separation, fractions containing PsGA2OX1 were pooled and dialysed versus 10 mM Tris-HCl pH 8.0, 1 mM DTT.
Pure PsGA2OX1 protein was concentrated to 10 mg/mL using a Vivascience spin concentrator and frozen.
Expression of AtGA3OX4
The cDNA AtGA3ox4 (GenBank AAG52440) from Arabidopsis thaliana was amplified by PCR using Advantage cDNA polymerase mix (Clontech) with the forward primer 5'- GGTATTGAGGGTCGCCTGGTGCCACGCGGTTCTATGCCTTCACTAGCAGAAGAG-3' and reverse primer 5'-AGAGGAGAGTTAGAGCCTTAATTGGTGGGATTAACGAC-3'.
The cDNA was inserted into the T7 promoter-based expression plasmid pET32 Xa/LIC (Novagen) using ligation independent cloning according to the manufacturer's instructions. The resulting construct expressed AtGA3OX4 with an N-terminal tag of both poly(His) and thioredoxin (TRX). A thrombin site was adjacent to the N-terminal of AtGA3OX4, encoded by the forward PCR primer.
A culture of 100 mL LB broth with 100 μg/mL ampicillin was inoculated with E. coli BL21 (DE3) pET32/AtGA3ox4 and grown overnight at 37°C with shaking. Thirty mL of starter culture was used to inoculate each of 3 flasks containing 1 L of LB broth with 100 μg/mL ampicillin. The flasks were incubated at 37°C with shaking at 220 rev/min on a circular orbit of 2.54 cm until the OD600 reached 1.0. IPTG was added to 1 mM and the cultures were grown for a further 3 h. Cells were harvested by centrifugation, washed with 1X PBS and frozen at -20°C.
Purification of AtGA3OX4
Cells (approximately 7 g) were thawed and resuspended at room temperature in 120 mL of 5 mM imidazole, 10 mM β-ME, 1X PBS pH 6.3 with 100 μg/mL lysozyme. Cell lysis was performed in a tube on ice using a Vibra Cell sonicator at 65% amplitude for 6 bursts of 10 s. The lysate was centrifuged at 20,000 g at 4°C and the soluble fraction was retained and kept on ice. Two more 10 s bursts of sonication were performed to remove high molecular weight DNA. The soluble fraction was diluted to a total volume of 300 mL with 1X PBS pH 6.3, 5 mM imidazole, 10 mM β-ME.
All column chromatography steps in the purification of AtGA3OX4 were performed at 6°C.
The crude protein solution was passed over a 5 mL HiTrap Chelating column (Amersham Biosciences) loaded with Ni++ using an Äkta Purifier chromatography system. The column was washed with 50 mL of 5 mM imidazole, 10 mM β-ME, 1X PBS pH 6.3 and the poly(His) TRX-AtGA3OX4 was eluted with a linear gradient of 0 -200 mM imidazole, 10 mM β-ME, 1X PBS pH 6.3. Fractions containing TRX-AtGA3OX4 were pooled and bovine thrombin (Sigma, 20 U) was added.
After incubation at 6°C overnight, cleaved protein was concentrated and loaded onto a Superdex 75 column equilibrated with 1X PBS pH 5.8, 10 mM β-ME. Fractions containing AtGA3OX4 were pooled and concentrated to 1.5 mg/mL with 1% Triton X-100.
In parallel experiments, the intact fusion TRX-AtGA3OX4 was further purified by gel filtration on Superdex 200 with 1X PBS pH 5.8, 10 mM β-ME. The purest fractions were pooled, protein was concentrated to 10 mg/mL and dialysed against 10 mM Tris-HCl pH 8.0, 2 mM β-ME.
Protein concentrations were measured with the Bradford method.
Activity assays on the substrate GA20
The substrate [1β,2β,3β-3H3]GA20 (approx. 0.4 TBq/mmole [23]), enabled the measurement of activity from both PsGA2OX1 and AtGA3OX4 via the formation of tritiated water. Quantitative estimates of product formation were possible because the 3H was incorporated at a ratio of 1:4:4 at the 1β, 2β and 3β positions [23]. The conversion GA29 to GA29-catabolite was assumed to make a relatively insignificant contribution to 3H release by PsGA2OX1 under the conditions of assay. This assumption was based on previous findings [4,14,15]. Similarly, it was assumed that AtGA3OX4 produced insignificant amounts of GA5 [4]. It should be noted that the kinetic parameters may be influenced by the presence of the label in the substrate due to a tritium isotope effect.
The standard assays were conducted using cofactors concentrations of established methods e.g. [14]. Reactions were in a volume of 200 μL with 100 mM Tris-HCl pH 7.9, 4 mM ascorbate, 4 mM DTT, 4 mM 2-oxoglutarate, 0.5 mM FeSO4, 2 mg/mL BSA and 1 mg/mL catalase. Substrate (5 μL in methanol) was added to a concentration of 0.02 μM for PsGA2OX1 assays and 0.1 μM for AtGA3OX4 assays. The enzyme solution was added in volumes ranging from 1 to 50 μL. Highly active solutions were serially diluted using 1 mg/mL BSA before addition. The reactions took place at 30°C and were stopped with 25 μL acetic acid after 1 h. Labelled GA20 was precipitated by adding 780 μL of 25 mM EDTA containing 50 mg/mL activated charcoal and centrifuging at 5,000 rev/min for 5 min. The release of tritiated water was measured by radiocounting a 300 μL reaction aliquot in scintillation fluid with tritium standards alongside. The specific activity figures for each purification stage were derived from enzyme samples (or dilutions of samples) showing similar rates of product conversion.
The substrate, [3H3]GA20, was varied in the assays for kinetic studies as shown in Figures 3 and 5. At higher substrate concentrations the maximum volume of methanol added was 10 μL (the extra methanol was shown not to be inhibitory). The amount of enzyme added was 1 μg for PsGA2OX1 and 0.1 μg for AtGA3OX4. Results were analysed using Origin by MicroCal.
List of abbreviations
GA – gibberellin
GST – glutathione-s-transferase
IMAC – immobilised metal affinity chromatography
kat – katal
β-ME – β-mercaptoethanol
2ODD – 2-oxoglutarate-dependent dioxygenase
PBS – phosphate-buffered saline
PCR – polymerase chain reaction
PMSF – phenylmethylsulfonyl fluoride
TRX – thioredoxin
Authors' contributions
DRL designed and performed most of the experimental work. She also drafted the manuscript. AP and PH developed the vector that enabled expression and purification of AtGA3OX4. IA conceived of the project, participated in experimental design and helped draft the manuscript. All authors read and contributed to the final draft of the manuscript.
Acknowledgements
This work was supported by the Swedish Research Council for Environment, Agricultural Sciences and Spatial Planning and the Swedish Research Council. Rothamsted Research is supported by the Biotechnology and Biological Sciences Research Council of the U.K.
Thanks to John Ross, School of Plant Science, University of Tasmania, for assistance with activity assays.
Figures and Tables
Figure 1 The latter steps of biosynthesis of non-13-hydroxylated/13-hydroxylated GAs. The steps surrounding production of bioactive GAs in higher plants are catalysed by the 2-oxoglutarate-dependent dioxygenases (2ODDs); 1 GA 20-oxidase, 2 GA 3-oxidase and 3 GA 2-oxidase. In pea the 13-hydroxylated pathway predominates; in Arabidopsis the non-13-hydroxylated pathway generally predominates, however developing seeds contain comparable levels of non-13-hydroxylated and 13-hydroxylated GAs [18]. The sln phenotype of pea arises from a null mutation affecting the predominant GA 2-oxidase (PsGA2OX1) of maturing seeds (steps 3). In this study the activities of a pea GA 2-oxidase (PsGA2OX1) and arabidopsis GA 3-oxidase (AtGA3OX4) are both assayed using the substrate [1β,2β,3β-3H3]GA20. It should be noted that in addition both enzymes are likely to use GA9 as a natural substrate.
Figure 2 Purification of recombinant PsGA2OX1 shown by SDS-PAGE on a 12% gel with Coomassie Blue staining. (A) Lanes: 1, soluble cell fraction; 2, GST-PsGA2OX1 after affinity chromatography; 3, PSGA2OX1 after thrombin digest and GST removal; 4, thrombin digest of GST-PsGA2OX1 with GST present; 5, Bio Rad molecular weight standards, (from top) 93, 49.8, 35.8, 29.2 and 21.3 kDa. (B) Lanes: 1, concentrated GST-PsGA2OX1 (uncleaved, 10 mg/mL); 2, concentrated PsGA2OX1 (10 mg/mL) after gel filtration on Superdex 75. The molecular weights expected for GST-PsGA2OX1, PsGA2OX1 and GST are 62.8, 36.8 and 26 kDa, respectively. The arrows indicate the position of PsGA2OX1.
Figure 3 Michaelis-Menten plot for highly pure PsGA2OX1 with the substrate [3H3]GA20. Origin (Microcal) was used to draw the plot and calculate the kinetic parameters. The KM was calculated as 0.024 μM and the Vmax as 4.4 pkat/mg.
Figure 4 Purification of recombinant AtGA3OX4 shown by SDS-PAGE on a 8–15% gradient gel with Coomassie Blue staining.(A) Lanes: 1, soluble cell fraction; 2, TRX-AtGA3OX4 after IMAC chromatography; 3, TRX-AtGA3OX4 after thrombin digest; 4, concentrated TRX-AtGA3OX4 (10 mg/ mL); 5, Bio Rad molecular weight standards, (from top) 115, 93, 49.8, 35.8, 29.2, 21.3 and 6.4 kDa.(B) Lanes: 1, TRX-AtGA3OX4 after thrombin digest; 2, AtGA3OX4 (1.5 mg/mL, maximal concentration achieved) after gel filtration on Superdex 75; 3, TRX-AtGA3OX4 (10 mg/mL) after gel filtration on Superdex 200. The molecular weights expected for TRX-AtGA3OX4, AtGA3OX4 and TRX are 54.1, 39.1 and 15 kDa, respectively. The arrows indicate the position of AtGA3OX4.
Figure 5 Michaelis-Menten plot for highly pure AtGA3OX4 with the substrate [3H3]GA20. Origin (Microcal) was used to draw the plot and calculate the kinetic parameters. The KM was calculated as 0.85 μM and the Vmax as 52,500 pkat/mg.
Table 1 Purification of PsGA2OX1 from E. coli cells (approx. 6 g)
Purification step Protein (mg) Specific activity* (pkat/mg) Total activity (pkat)* Apparent purification factor* Yield*
Soluble cell fraction 670 0.013 8.71 - -
Affinity chromatography (after GST removal) 3.8 0.320 1.2 24X 14%
Superdex 75 1.8 0.89 1.6 2.8X 18%
*Figures influenced by EDTA of buffers in the first two steps
Table 2 Purification of AtGA3OX4 from E. coli cells (approx. 7 g)
Purification step Protein (mg) Specific activity (pkat/mg) Total activity (pkat) Apparent purification Factor Yield (%)
Soluble cell fraction 790 360 284,400 - -
IMAC chromatography 4.0 4,325 17,300 12X 6%
Superdex 75 0.075 5,466 410 1.26X 0.14%
==== Refs
Sakamoto T Miura K Itoh H Tatsumi T Ueguchi-Tanaka M Ishiyama K Kobayashi M Agrawal GK Takeda S Abe K Miyao A Hirochika H Kitano H Ashikari M Matsuoka M An overview of gibberellin metabolism enzyme genes and their related mutants in rice Plant Physiology 2004 134 1642 1653 15075394 10.1104/pp.103.033696
Hedden P Phillips AL Gibberellin metabolism: new insights revealed by the genes Trends in Plant Science 2000 5 523 530 11120474 10.1016/S1360-1385(00)01790-8
Frisse A Pimenta MJ Lange T Expression studies of gibberellin oxidases in developing pumpkin seeds Plant Physiology 2003 131 1220–1227 12644672 10.1104/pp.015206
King RW Junttila O Mander LN Beck EJ Gibberellin structure and function: biological activity and competitive inhibition of gibberellin 2- and 3-oxidases Physiologia Plantarum 2004 120 287 297 15032864 10.1111/j.0031-9317.2004.0227.x
Schomburg FM Bizzell CM Lee DJ Zeevaart JAD Amasino RM Overexpression of a novel class of gibberellin 2-oxidases decreases gibberellin levels and creates dwarf plants Plant Cell 2003 15 151 163 12509528 10.1105/tpc.005975
Williams J Phillips AL Gaskin P Hedden P Function and substrate specificity of the gibberellin 3β-hydroxylase encoded by the Arabidopsis GA4 gene Plant Physiology 1998 117 559 563 9625708 10.1104/pp.117.2.559
Zhou R Yu M Pharis RP 16,17-Dihydro gibberellin competitively inhibits a recombinant Arabidopsis GA 3β-hydroxylase encoded by the GA4 gene Plant Physiology 2004 135 1000 1007 15181202 10.1104/pp.104.041509
Griggs DL Hedden P Lazarus CM Partial purification of two gibberellin 2β-hydroxylases from cotyledons of Phaseolus vulgaris Phytochemistry 1991 30 2507 2512 10.1016/0031-9422(91)85090-M
Kwak SS Kamiya Y Sakurai A Takahashi N Graebe JE Partial purification and characterization of gibberellin 3β-hydroxylase from immature seeds of Phaseolus vulgaris L. Plant Cell Physiology 1988 29 935 943
Lange T Schweimer A Ward DA Hedden P Graebe JE Separation and characterization of three 2-oxoglutarate-dependent dioxygenases from Cucurbita maxima L. endosperm involved in gibberellin biosynthesis Planta 1994 195 98 107
Smith VA MacMillan J The partial purification and characterisation of gibberellin 2β-hydroxylases from seeds of Pisum sativum Planta 1986 167 9 18 10.1007/BF00446362
Lange T Purification and partial amino-acid sequence of gibberellin 20-oxidase from Cucurbita maxima endosperm Planta 1994 195 108 115 7765793
Lester DR Ross JJ Davies PJ Reid JB Mendel's stem length gene (Le) encodes a gibberellin 3β-hydroxylase Plant Cell 1997 9 1435 1443 9286112 10.1105/tpc.9.8.1435
Lester DR Ross JJ Smith JJ Elliott RC Reid JB Gibberellin 2-oxidation and the SLN gene of Pisum sativum Plant Journal 1999 19 65 73 10417727 10.1046/j.1365-313X.1999.00501.x
Martin DN Proebsting WM Hedden P The SLENDER gene of pea encodes a gibberellin 2-oxidase Plant Physiology 1999 121 775 781 10557225 10.1104/pp.121.3.775
Ross JJ Reid JB Swain SM Control of stem elongation by gibberellin A1 : Evidence from genetic studies including the slender mutant sln Australian Journal of Plant Physiology 1993 20 585 599
Lukacin R Gröning I Schilz E Britsch L Matern U Purification of recombinant flavanone 3β-hydroxylase from Petunia hybrida and assignment of the primary site of proteolytic degradation Archives of Biochemistry and Biophysics 2000 375 364 370 10700394 10.1006/abbi.1999.1676
Curaba J Moritz T Blervaque R Parcy F Raz V Herzog M Vachon G AtGA3ox2, a Key Gene Responsible for Bioactive Gibberellin Biosynthesis, Is Regulated during Embryogenesis by LEAFY COTYLEDON2 and FUSCA3 in Arabidopsis Plant Physiology 2004 136 3660 3669 15516508 10.1104/pp.104.047266
Anzellotti D Ibrahim RK Molecular characterization and functional expression of flavonol 6-hydroxylase BMC Plant Biology 2004 4 20 15596008 10.1186/1471-2229-4-20
Turnbull JJ Prescott AG Schofield CJ Wilmouth RC Purification, crystallization and preliminary X-ray diffraction of anthocyanidin synthase from Arabidopsis thaliana Acta Crystallographica D 2001 57 425 427 10.1107/S0907444900019818
Wilmouth RC Turnbull JJ Welford RWD Clifton IJ Prescott AG Schofield CJ Structure and mechanism of anthocyanidin synthase from Arabidopsis thaliana. Structure 2002 10 93 103 11796114 10.1016/S0969-2126(01)00695-5
Zhang Z Ren JS Clifton IJ Schofield CJ Crystal structure and mechanistic implications of 1-aminocyclopropane-1-carboxylic acid oxidase – the ethylene-forming enzyme Chemistry and Biology 2004 11 1383 1394 15489165 10.1016/j.chembiol.2004.08.012
Willis CL Gaskin P MacMillan J [1β,2β,3β-3H3]gibberellin GA20: Confirmation of structure by 3H NMR and by mass spectrometry Phytochemistry 1988 27 3970 3972 10.1016/0031-9422(88)83059-0
|
16181493
|
PMC1262730
|
CC BY
|
2021-01-04 16:03:52
|
no
|
BMC Plant Biol. 2005 Sep 25; 5:19
|
utf-8
|
BMC Plant Biol
| 2,005 |
10.1186/1471-2229-5-19
|
oa_comm
|
==== Front
BMC Struct BiolBMC Structural Biology1472-6807BioMed Central London 1472-6807-5-181619754810.1186/1472-6807-5-18Research ArticlePrediction of "hot spots" of aggregation in disease-linked polypeptides de Groot Natalia Sánchez [email protected]és Irantzu [email protected]és Francesc X [email protected] Josep [email protected] Salvador [email protected] Departament de Bioquímica i Biologia Molecular, Facultat de Ciències, Universitat Autònoma de Barcelona, E-08193 Bellaterra, Spain2 Institut de Biotecnologia i de Biomedicina, Universitat Autònoma de Barcelona, E-08193 Bellaterra, Spain2005 30 9 2005 5 18 18 28 9 2005 30 9 2005 Copyright © 2005 de Groot et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The polypeptides involved in amyloidogenesis may be globular proteins with a defined 3D-structure or natively unfolded proteins. The first class includes polypeptides such as β2-microglobulin, lysozyme, transthyretin or the prion protein, whereas β-amyloid peptide, amylin or α-synuclein all belong to the second class. Recent studies suggest that specific regions in the proteins act as "hot spots" driving aggregation. This should be especially relevant for natively unfolded proteins or unfolded states of globular proteins as they lack significant secondary and tertiary structure and specific intra-chain interactions that can mask these aggregation-prone regions. Prediction of such sequence stretches is important since they are potential therapeutic targets.
Results
In this study we exploited the experimental data obtained in an in vivo system using β-amyloid peptide as a model to derive the individual aggregation propensities of natural amino acids. These data are used to generate aggregation profiles for different disease-related polypeptides. The approach detects the presence of "hot spots" which have been already validated experimentally in the literature and provides insights into the effect of disease-linked mutations in these polypeptides.
Conclusion
The proposed method might become a useful tool for the future development of sequence-targeted anti-aggregation pharmaceuticals.
==== Body
Background
In the last decade, protein aggregation has moved beyond being a mostly ignored area of protein chemistry to become a key topic in medical sciences [1], mainly because the presence of insoluble deposits in human tissues correlates with the development of many debilitating human disorders including the amyloidoses and several neurodegenerative diseases [2]. The proteins involved in these diseases are not related in terms of sequence or secondary structure content. From the conformational point of view, two major classes can be distinguished: globular proteins with a stable unique conformation in the native state and intrinsically unstructured proteins [3]. Globular proteins rarely aggregate from their native states and destabilization, resulting in an increased population of unfolded molecules, is well established as a trigging factor in disorders associated with the deposition of proteins that are globular in their normal functional states [4], as in the cases of β2-microglobulin, lysozyme, transthyretin and the prion protein. Interestingly enough, many proteins involved in depositional disorders are mostly unstructured within the cell [3]. These include amylin, amyloid-β-protein, and α-synuclein, among others. In these cases, protein deposition does not require unfolding and can occur by direct self-assembly of the unstructured polypeptide chains.
One of the major unanswered questions of protein aggregation is the specificity with which the primary sequence determines the aggregation propensity from totally or partially unfolded states. Deciphering the answer to this question will give us a chance to control the unwanted protein deposition events through specific sequence-targeted therapeutics. A first advance in this direction is the recent discovery that not all regions of a polypeptide are equally important for determining its aggregation tendency, both in natively unfolded and globular proteins. In this way, some authors, including ourselves, have proved recently that very short specific amino acid stretches can act as facilitators or inhibitors of amyloid fibril formation [5,6]. These relevant regions are usually known as aggregation "hot spots". Aggregation-prone regions are likely to be blocked in the native state of globular proteins because their side chains are usually hidden in the inner hydrophobic core or already involved in the network of contacts that stabilizes a protein. This accounts for the protective role of the native structure against aggregation [7]. In contrast, aggregation-prone regions are already exposed to solvent in natively unfolded proteins, available for the establishment of inter-molecular contacts that may finally lead to the formation of aggregates. Accordingly, the presence of putative "hot spots" of aggregation is much more frequent in the sequences of globular proteins than in those coding for natively unfolded proteins [8]. The presence of aggregation-prone regions has been described in most of the peptides and proteins underlying neurodegenerative and systemic amyloidogenic disorders [9].
We have used a simple in vivo system to study the aggregation effects of a complete set of mutations in one of the best characterized "hot spots" in a disease-linked protein: the central hydrophobic cluster (CHC) of the Amyloid-β-protein (Aβ) [10,11]. The results in this and other studies on protein models not related to disease [12], suggested that common and simple principles underlie protein aggregation, at least from totally or partially unfolded states, and that the propensities of proteins backbones to aggregate are sharply modulated by the sequences that dress them. Based on these assumptions, we have developed a simple approach that identifies the presence of "hot-spots" of aggregation in globular and unstructured disease-linked polypeptides and predicts the aggregation effects of mutations in their sequences.
Results and discussion
Aggregation propensities of natural amino acids
The rationale behind our study is based on two recent observations in the field. First, not all the polypeptide sequence is relevant for the aggregation of a given protein, but rather there exist specific regions that drive the process [5,6] and second, similar simple rules appear to underlie the aggregation propensities of unrelated proteins from unfolded states [12]. According to these two assumptions one may expect that the conclusions obtained from the study of a relevant "hot spot" of aggregation in a specific protein could apply to other unrelated proteins involved in disease. As commented upon previously, we have exploited an in vivo reporter method to calculate the relative aggregation propensities of each individual natural amino acid when placed in the central position of the CHC of Aβ (see Material and Methods). The highest aggregation propensities correspond to isoleucine, phenylalanine, valine, and leucine, whereas aspartic, glutamic, asparagine, and arginine exhibit the lowest (Table 1). In general, hydrophobic residues tend to induce aggregation whereas polar ones promote solubility, matching the general assumption that hydrophobic interactions are supposed to play an important role in protein aggregation [13].
Generation of protein aggregation profiles and prediction of the effects of protein mutation on the aggregation propensity
Provided that a given polypeptide aggregates from an at least partially unstructured state, the experimental intrinsic aggregation propensities shown in Table 1 should apply independently of the protein context. Thus, a profile can be theoretically generated for any protein sequence to detect those regions with aggregation propensities above the average value of the whole sequence. This leads directly to the definition of "hot spot" of aggregation as a certain region that displays higher aggregation propensity than the rest of the sequence. Interestingly, a related approach has been reported very recently for the analysis of unstructured proteins associated with neurodegenerative diseases[14].
A good number of natural occurring mutations have been reported in proteins associated to depositional diseases. In many cases they result in changes in the global protein aggregation propensity and sometimes in the appearance of premature or acute pathological symptoms. The change in average aggregation propensity (ΔAP) between the wild type and the different mutants should predict the effect of sequence variations on the aggregation propensities, provided that they rely on changes in the intrinsic polypeptide properties.
Analysis of disease-related polypeptide sequences
In this section the above described analysis is applied to a set of proteins linked to depositional diseases and the obtained results are compared with the available experimental data.
Intrinsically unstructured proteins
Amyloid-β-protein
As a proof of principle our approach was first tested in the molecule from which the experimental amino acid aggregation propensities were derived. Alzheimer's disease (AD) is a progressive neurodegenerative disorder characterized by the patient's memory loss and impairment of cognitive abilities. The extracellular amyloid is found in the brain and is widely believed to be involved in the progression of the disease [15]. The principal component of the lesions is the hydrophobic polypeptide Aβ. The most abundant forms found in amyloid plaques are a 40-mer (Aβ40) and a 42-mer (Aβ42). Although less abundant, Aβ42 is more amyloidogenic than Aβ40 and is the major component of neuritic plaques [16]. Two main regions with high aggregation propensity can be distinguished in the aggregation profile for this polypeptide (Fig. 1). The second region arises from the contribution of two sequence stretches comprising residues 30–36 and 38–42, respectively. The predicted aggregation-prone regions are in excellent agreement with the experimental data in the literature. Residues 16–21 overlap with the CHC sequence comprising residues 17–21, a particular region recognized to play a key role in Aβ aggregation and that is defined as specially relevant for the amyloidogenesis of the Aβ40 and Aβ42 peptides by two recent proline-scanning-mutagenesis studies [17,18]. In addition, structural studies using solid state-NMR [19] and site-directed spin labeling [20] have revealed that residues 16–21 are located in the core of Aβ fibrils. Accordingly, a short 7 residues fragment comprising residues 16–22 is able to form ordered amyloid fibrils [21] and, more interestingly, 16-LVAFF-20 and derived peptides have been shown to bind to Aβ42 and act as potent inhibitors of amyloid formation [22]. The region 30–42, including both 30–36 and 38–42 stretches, has been also implicated in Aβ aggregation. Proline-scanning-mutagenesis revealed that the region 31–36 is sensitive to proline replacement and likely to include a β-sheet portion of the Aβ fibrils [17,18]. The contribution of the C-terminal region 38–42 to Aβ amyloidogenesis becomes clear from the observation that, although Aβ40 is produced in greater abundance in vivo, the prevalence of the full-length 42-mer in plaques is much higher [16]. Experiments with truncated synthetic Aβ peptides have confirmed that Aβ39 and Aβ40 are kinetically soluble for several days, whereas Aβ42 immediately aggregates into amyloid fibrils [23]. The relevance of the predicted 30–42 region is confirmed by structural studies that demonstrate that residues 30–40 are located in the core of the Aβ fibrils [20].
A set of mutations in the CHC and adjacent positions of Aβ42 is intimately associated to early-onset familial Alzheimer diseases (FAD). The substitutions include A21G (Flemish), E22Q (Dutch) and E22G (Arctic) [24]. Aβ42 congeners bearing these mutations display distinct aggregation kinetics. The rate of fibril formation by the Flemish mutant is decreased relative to WT Aβ42, whereas the Dutch mutant peptide aggregates substantially faster. The Arctic peptide does not shows an overall change in the rate of fibrillogenesis relative to WT Aβ42, but rather accelerated protofibril formation. To assess whether the effect of such mutations could be predicted by the present approach we calculated ΔAP for the different sequences. The results obtained describe accurately the effects documented in the literature (Table 2).
Adding to the mutations present in the population, a large set of mutations has been artificially introduced on Aβ that result in changes in its aggregation propensity. ΔAP values were also calculated for several of them and the results compared with the experimental data (Table 2). The calculated changes in aggregation propensity are in excellent agreement with the trends reported in the literature. Briefly, we predict the changes in aggregation of F19 mutants, those of I31 and I32 in the 30–36 region and those of I41 and A42 in the C-terminal region, as well as the effects of deletions both in the N and the C ends. Finally, we also predict the high solubility of Aβ versions generated by random mutagenesis [25].
Islet amyloid polypeptide
Type II diabetes is associated with progressive beta-cell failure manifested as a decline in insulin secretion and increasing hyperglycemia. A growing body of evidence suggests that beta-cell failure in type II diabetes correlates with the formation of pancreatic islet amyloid. Islet amyloid polypeptide (IAPP, amylin), the major component of islet amyloid, is co-secreted with insulin from beta-cells. In type II diabetes, this peptide aggregates to form amyloid fibrils that are toxic to beta-cells [26]. IAPP is an unstructured peptide hormone of 37 amino acid residues. Two "hot spots" of aggregation comprising residues 12–18 and 22–28 are detected for this peptide (Fig. 1). Interestingly enough, a 8–37 IAPP-fragment including both "hot spots", has been shown to form amyloid fibrils under physiological conditions [27]. The two aggregation prone regions sharply coincide with those protected in the core of the fibrils in a recently described structural model of IAPP aggregates [28]. In this study, residues 12–17 and 22–27 are proposed to form the inner β-sheets in the fibril protofilament structure. According to this hypothesis, peptides corresponding to residues 8–20, 10–19, 20–29 of human IAPP, which include one of the "hot spots" described here, all form amyloid [29-31]. Smaller peptides derived from these regions have also been shown to form amyloid, and a recent investigation suggests that the minimal amyloid forming fragment of IAPP consists of residues 22–27. This hexapeptide fragment, NFGAIL, forms β-sheet-containing fibrils that coil around each other in typical amyloid fibril morphology [32].
The analysis also explains the available mutational data on IAPP. Diabetes-associated IAPP amyloid occurs in primates and cats but not in rodents [33]. Consistently, the sequences of peptides 20–29 of rodents display reduced average aggregation propensity relative to that of cat and human (Table 3). We also predict the slightly increased aggregation propensities of single or multiple mutations of rat IAPP to the corresponding residues of human IAPP [33]: R18H, L23F or V26I, as well as the results from alanine-scanning-mutagenesis in a peptide encompassing residues 22–27 [32] (Table 3). It has been found that a substitution at position 20 (S20G) in the IAPP molecule in a reduced subpopulation of Japanese people with type II diabetes is associated with an earlier onset and more severe form of disease [34]. In this case, our approach does no predict an increased but a slightly reduced aggregation propensity in the mutant, suggesting that the pathological symptoms in this variant may arise from non-intrinsic factors. In fact, it has been suggested that the accelerated aggregation of the S20G variant could be related to structural reasons, resulting from a better packing of the turns connecting the β-sheets in the final protofilament structure [28] that cannot be predicted by the present approach.
Several mechanisms have been proposed for IAPP fibril formation in type II diabetes. One widely accepted mechanism is that in type II diabetes, increased production and secretion of IAPP associated with increased demand for insulin might result in accumulation and aggregation of IAPP [35]. A second view considers that impaired processing of the IAPP precursor molecule, proIAPP, by islet beta-cells may lead to hypersecretion of unprocessed or partially processed forms of proIAPP that may have a higher tendency for aggregation compared to mature IAPP [35]. Our calculated average aggregation propensities for proIAPP and processed IAPP support this view (Table 3).
α-Synuclein
Parkinson disease is the most common neurodegenerative movement disorder and is pathologically characterized by the presence of neuronal intracytoplasmatic deposits of aggregated protein called Lewy bodies [36]. Lewy bodies also occur in other cognitive disorders, globally known as α-synucleinopathies. α-Synuclein is the major component of the fibrils that form the Lewy bodies [36]. It is a small (137 residues), natively unfolded, soluble, presynaptic and highly conserved protein without a well-defined function. The aggregation profile for this polypeptide is shown in Fig. 1. Several large aggregation-prone stretches were predicted for the α-Synuclein sequence: region 1–18, region 27–56 and specially region 61–94. Again, our predictions are in complete agreement with the experimental data in the literature, as many studies suggest that the central region of the protein, known as the non-Aβ component of amyloid plaques (NAC, amino acids 61–95), is the responsible for its aggregation process [37]. A peptide comprising residues 68–78 of α-synuclein has been shown to be the minimum fragment that, like α-synuclein itself, forms amyloid fibrils and exhibits toxicity towards cells in culture [38]. This fragment is included in the region 62–80 which we predict as the sequence stretch with the highest aggregation propensity. All the α-synucleinopathies are characterized by the accumulation of the 35 residues NAC fragment in the insoluble deposits [37]. Accordingly, this central region is predicted to have a much higher average aggregation propensity than its soluble precursor (ΔAP = +38.47). The importance of this hydrophobic stretch is further supported by its absence in β-synuclein, a homologue of α-synuclein, with strongly reduced propensity for fibril formation. It has been shown that the deletion of amino acids 71–82 within the hydrophobic region abrogated the ability of human α-synuclein to polymerize into fibrils [39]. Protease digestion studies suggest that the core region of α-synuclein in the fibrils could be longer, since a 7-kDa fragment (comprising residues 31–109) was shown to be protected from proteinase K digestion [40]. This region contains the putative 12-residue core domain, as well as the NAC region and includes the second and third "hot spots" in our profile. A structural study on the organization of α-synuclein in the fibrilar state using site-directed spin labelling confirms that the 34–101 residues region constitutes the core of the fibrils forming a parallel in-register β-sheet structure whereas the N terminus is structurally more heterogeneous and the C terminus (40 amino acids) is completely unfolded [41].
Several α-Synuclein mutations appear associated with familial early-onset Parkinson Disease: A30P, A53T and E46K. All they map into our predicted second "hot spot". The rates of fibril assembly of the E46K and A53T mutants have been shown to be greater than those of the wild type and A30P proteins [42]. We predict a similar average aggregation propensity for the wild-type and the A30P mutant and an slightly increased aggregation propensity for the E46K mutant, but fail to foresee the effect of the A53T mutation in promoting the formation of protofibrils. Obviously, other functional factors apart from the intrinsic aggregation propensities can strongly influence the aggregation tendency of unfolded polypeptide chains within the cell. In fact the effects of α-synuclein mutations have been associated either to an impaired degradation inside lysosomes or to a reduced axonal transport of the variants [43,44]. Both situations may result in increased concentrations of the protein in certain regions of the neuron that may favor the nucleation step of amyloid formation. According to this, α-synuclein gene triplication identified in two independent families [45] has been shown to accelerate the development of Parkinson disease. Thus, an increase in the amount of cellular α-synuclein appears to be important for the pathogenesis of Parkinson disease, suggesting that the effects of the different α-synuclein mutations on protein aggregation could be quantitative, in terms of local concentration, rather that qualitative. Thus, experimental deviations from the theoretical predictions in natively unfolded proteins, in addition to reflect limitations of the approach, might also contain relevant information, prompting to find alternative structural, as in the case of amylin, or functional, as in the case of α-synuclein, explanations for the observed behavior.
Globular proteins
β2-Microglobulin
β2-Microglobulin-related amyloidosis is a common and serious complication in patients on longterm hemodialysis [46]. Intact β2-microglobulin is a major structural component of the amyloid fibrils. β2-Microglobulin (β2-m) is a small (99 residues) non-glycosilated protein with an immunoglobulin-like fold consisting in two antiparallel pleated β-sheets linked by a disulfide bond (Fig. 2). β2-m has been shown to form amyloid fibrils in vitro under different conditions, but in all cases β2-m populates unfolded non-native states as precursors to fibril assembly [47]. Under these conditions aggregation-prone regions, if present, may promote and drive the aggregation event. According to the analysis of the aggregation profile, shown in Fig. 3, this protein displays four "hot spots" encompassing residues 21–31, 56–69 and 79–85, and 87–91. These regions sharply coincide with four different secondary structure elements in β2-m: β-strand 2, formed by residues 21–31; β-strand 6, formed by residues 61–71; β-strand 7; formed by residues 77–85 and β-strand 8, formed by residues 86–95 (Fig. 2). In agreement with our prediction a peptide comprising residues 21–41 has been shown to form fibrils in isolation [48]. In addition, a N-terminal fragment of this short peptide corresponding exactly to our first "hot spot" [21-31] is also able to self-assemble into fibrillar structures [49]. Interestingly enough, the peptides 23–31 and 21–29 exhibited reduced amyloidogenesis [49]. Thus, in this particular "hot spot" the prediction delimits not only the overall region important for aggregation but also its precise size. The amino acid stretches 59–79 and its shorter version 59–71 which overlap with the predicted second aggregation-prone region of β2-m have been also shown to form fibrils [50]. The C-terminal fragment 72–99 of β2-m has been also reported to form amyloid [51]. This 29 residues sequence includes our third and fourth "hot spots" of aggregation. The peptide 91–99 does not aggregate, indicating that the last 9 residues of β2-m are not relevant for amyloidogenesis as predicted here [49]. The N-terminal region, for which no aggregation propensity is predicted, is probably not involved inthe aggregation process as evidenced by the fact that the fragment 6–12 does not form fibrils [49]. This observation could be physiologically relevant since the N-terminus of β2-m is truncated in 30% of the molecules extracted from ex vivo fibrils [52].
In contrast to the human protein, mouse β2-m does not form fibrils even at high concentration [53]. Based on this observation a seven residues region corresponding to residues 83–89 of human β2-m has been suggested to be particularly important for aggregation, since it corresponds to the sequence with the highest divergence between both species. This hypothesis has been tested experimentally, since a heptapeptide bearing the human sequence is able to self-assemble whereas the mouse version is not [53]. The complete mouse sequence is predicted to have a strongly reduced aggregation propensity (ΔAP = -47.86).
Overall, our predictions on the presence and location of "hot spots" in β2-m are extremely accurate and overlap with the experimentally found relevant regions (Fig. 2). The observation that short peptides including the aggregation-prone regions described here form amyloids implies that exposure of previously hidden short segments can nucleate native proteins into the amyloid state and reinforces the hypothesis that fibril formation is sequence specific.
One of the most urgent issues in the study of amyloid fibrils is to reproduce the formation of fibrils under physiological conditions. Recently, it has been found that low concentrations of SDS around the critical micelle concentration induce the extensive growth of β2-m amyloid fibrils at physiological pH, probably through the SDS-induced conformational change of β2-m monomers [54]. Contrarily to what was expected, the presence of low concentration of SDS had little effect on the stability of the protein and did not promote global protein unfolding. Our results strongly suggest that in β2-m the parts of the molecule involved in aggregation are located in pre-formed β-strands. Therefore, it is possible that local unfolding events may allow anomalous intermolecular interaction between this preformed elements leading to the formation of an aggregated β-sheet structure. This would explain the formation of amyloid deposits in hemodyalisis patients in which no major unfolding of the protein is expected to occur, as well as the effect of seeds, which may have exposed aggregation prone β-strands, in strongly accelerating the aggregation process of β2-m under physiological conditions [55].
Lysozyme
Human lysozyme has been shown to form amyloid fibrils in individuals suffering from nonneuropathic systemic amyloidosis. The disease is always associated to point mutations in the lysozyme gene and fibrils are deposited widely in tissues [56]. The properties of two amyloidogenic lysozyme mutants (I56T and D67H) have been studied in detail and, when compared to those of the wild-type protein, the mutants were found to have reduced structural stability allowing unfolding to take place at least partially at physiologically relevant temperatures [57,58]. Thus, the formation of amyloid fibrils by human lysozyme is likely to occur by the exposure of aggregation-prone region previously hidden in the native structure. The aggregation profile of lysozyme identifies three main "hot spots" corresponding to residues 20–34, 50–62 and 73–104 (Fig. 3). The last large aggregation-prone region includes several local maxima. The first "hot spot" maps in helix B, the second in a β-hairpin of the β-domain and the third includes helix C and a large flanking unstructured region at its N-terminus (Fig. 2). Although there is no experimental characterization of amyloidogenic regions in human lysozyme in the literature, this information is available for the homologous hen lysozyme molecule, which displays an almost identical 3D-structure. The aggregation profile for the hen protein is very similar to that of the human one despite the fact that our input consists solely on the sequence and the identity between both molecules is only of 40%. The equivalent "hot spots" in hen lysozyme comprise residues 24–34, 50–62 and 76–98. Experimental data suggests that the sequence of the β-domain could be of particular relevance for lysozyme aggregation since it unfolds prior to the α-domain [58]. Two peptides encompassing the β-domain of native lysozyme displayed very different behavior: peptide 61–82 appeared to be predominantly unstructured whereas peptide 41–60 showed a high tendency to aggregate and form extended β-sheet structures [59]. The first peptide coincides with a region of very low aggregation propensity in the aggregation profiles, whereas the second one covers the region with the highest aggregation propensity in the profile (residues 50–64). Interestingly enough, a peptide spanning residues 49–64 has been shown to form fibrils with the typical structure of amyloid showing that the first residues of the 41–60 peptide are not relevant for aggregation, as predicted by our approach [60]. Another study has reported that the major fragment incorporated in the core of the fibril structure, as monitored using proteolysis, encompasses the chain region 49–101 [61]. These lysozyme fragments contain helix C and two of the three β-strands of the β-domain of the native protein structure and coincide with the limits of the second and third regions in our predictions (Fig. 2 and 3). This observation could be biologically relevant, since the β-domain and C-helix of the human lysozyme have been shown to unfold locally in the amyloidogenic variant D67H, which is associated with the familial cases of systemic amyloidosis linked to lysozyme deposition [58]. The C-helix is the α-helix with the lowest helical propensity of hen lysozyme according to both theoretical and peptide based studies [59]. This low propensity might be related to the ability of this region to be incorporated into the β-sheet rich fibrillar structured as have been reported for other protein systems [62]. Limited proteolysis of hen lysozyme renders fragments 57–107 and 1–38/108–129 [61]. In the 1–38/108–129 fragment the N-terminal and C-terminal ends of the molecule are joined by a disulfide bond. Only fragment 57–107, but not fragment 1–38/108–129, is able to generate well defined amyloid [61]. Whereas the behavior of the 57–107 fragment is expected from the analysis, one should also expect the fragment 1–38 to have a high tendency to aggregate. Two explanations are possible to account for this discordance. First, it could occur that the helical structure of this region prevents its conversion to β-sheet conformation, since the A-helix displays the highest helical propensity out of all lysozyme helices [59]. The second possibility is that, being joined to the 108–129 region, predicted to have lower aggregation propensity, steric hindrances limit self-assembly or alternatively the average aggregation tendency of this peptide becomes reduced. The analysis supports this last hypothesis reporting a decrease in aggregation propensity (ΔAP = -5.34) in the joined peptide respect the 1–38 peptide alone.
Transthyretin
Transthyretin (TTR) is a homotetramer of 127-amino acid subunits. TTR is found in human plasma and cerebral spinal fluid, the plasma form being the amyloidogenetic precursor. TTR constitutes the fibrillar protein found in familial amyloidotic polyneuropathy (FAP) and senile systemic amyloidosis (SSA) [63]. In the case of FAP, the amyloid is associated with a point mutation in the TTR gene. To date, 100 different TTR mutations have been reported, many of which are amyloidogenic [64]. The FAP-associated variants characterized thus far although tetrameric, are destabilized [65]. This destabilization allows tetramer dissociation to the amyloidogenic monomeric intermediate to occur under the influence of mild denaturing denaturation conditions. More than 10 FAP-related variants crystal structures have been solved, revealing that the tertiary and quaternary structures are essentially identical to the wild type form [65]. This observation suggests that the partial denaturation of TTR is a requirement for amyloidogenesis. In this state, the presence of "hot spots" of aggregation could play an especially important role in promoting/driving amyloid formation. According to the analysis of the aggregation profile shown in Fig. 3, the TTR monomer displays three main "hot spots" encompassing residues 10–20, 23–33 and 105–118. Also in this case, aggregation-prone sequences appear to be located in preformed β-sheet structures: A β-strand (11–19), part of the B β-strand (28–36) and G and beginning of H β-strands (104–123) (Fig. 3). Most of these secondary structure elements are involved in the formation of the tetrameric structure: H strands mediate the dimerization whereas A and G provide the contacts for the tetramerization of two preformed dimmers. This explains the protective role played by the TTR quaternary structure against aggregation, since it hides or blocks most of the aggregation prone regions. Dissociation of the tetramer has been reported as a prerequisite for amyloidosis and according to our results might be associated to the exposure of previously hidden amyloidogenic sequences. We detect several short peaks exhibiting high aggregation propensities in the central region (63–94) of TTR. These result from the presence of almost regularly placed residues with low aggregation propensity (Asp, Glu, Arg, Lys, Gly) in this rather hydrophobic sequence, which probably act as disrupters, significantly lowering the aggregation tendency of this particular region, a strategy suggested to be used by nature to avoid edge-to-edge aggregation [66].
To date two different fragments of TTR have been shown to form amyloid fibrils. The peptide 105–115 can be assembled into homogeneous amyloid fibrils with favorable spectroscopic properties [67]. This has allowed to solve its fibrillar structure at high-resolution, showing that it adopts an antiparallel extended beta-strand conformation in the amyloid fibrils [68]. This peptide coincides with the region with the highest aggregation propensity in the profile. Also in excellent agreement with the prediction, the peptide 10–20 is the only other fragment of TTR reported to form amyloid fibrils [69]. No data are available on the region 23–33 but the success of the present method in predicting relevant regions in TTR suggests that it is worth to characterize its in vitro aggregation capabilities.
Prion protein
Misfolded isoforms of the naturally occurring prion protein (PrP) have been shown to be the causative agents in many mammalian neurodegenerative disorders, including Cruetzfeldt-Jakob disease (CJD) in human, scrapie in sheep, and bovine spongiform encephalopathy in cows. Prion infectivity is unique in that the pathogenic prion form (PrPSc) is involved in the conversion of the endogenous conformation (PrPC) into transformed PrPSc. The "protein-only" hypothesis [70] asserts further that no extraneous agents are necessary to explain the unusual behavior of prions. Prion diseases can have infectious, familial, and sporadic origins. The basic infectious mechanism is thought to be a conformational change of the normal prion protein (PrPC) into the pathogenic PrPSc catalyzed by PrPSc itself.
The normal prion protein (PrPC) is a GPI-anchored glycoprotein constitutively expressed on the surface of primarily neuronal cells. It consists of two structurally different parts; a C-terminal, globular part mainly α-helical in nature (Fig. 2) and an unstructured, N-terminal part [71]. Misfolding of PrPC into PrPSc occurs posttranslationally and results in increased β-sheet content and gain of protease-resistance. Fig. 3 shows the predicted "hot spots" in the aggregation profile of the full-length human prion protein. They are located at the N-terminus (1–32), in the central region (105–146) and the C-terminus (208–252), respectively.
The role of the detected aggregation-prone sequence at the N-terminus is uncertain since it is out of the protease resistant core of PrPSc. Little information exits about the role of this region, although it appears to be unnecessary both for prion transmission and aggregation. The predicted C-terminal "hot spot" includes almost all the C-terminal α-helix, named C, from the globular domain (Fig. 2). Interestingly, some of the human mutations linked to Creutzfeldt-Jakob disease occur in this region of the prion protein and it has been related to the conversion of PrPC into the toxic PrPSc. Moreover, some strains of PrP resistant to conversion to PrPSc have been found to bear mutations in helix C, and positions 214 and 218 have been shown to modulate PrPSc formation [72]. It is also important to note that the main structural differences between prion proteins from different species have been found at the end of helix C [71].
The central region of PrPC linking the unstructured N-terminal part with the globular C-terminal domain is believed to play a pivotal role in the PrPC conformational changes. Extensive studies on the secondary structure and fibrillogenic properties of synthetic peptides of PrP have established that the continuous segment of the prion protein spanning residues 106–147, coincident with the second "hot spot" predicted using our approach, is important for the fibrillogenic properties of the protein [73]. One of the synthetic peptides, that named PrP106-126 within the central region of PrP and near the N-terminal of the protease resistant core of PrPSc, shares many properties with the infectious form as it readily forms amyloid fibrils with a high β-sheet content, shows partial proteinase K resistance and is neurotoxic in vivo [74]. The neurotoxicity of PrP106-126 depends on the expression of endogenous PrPC which makes PrP106-126 a relevant model for PrPSc neurotoxicity [74]. Also another prion derived peptide – PrP118-135 – has been found to cause neuronal death via induction of apoptosis [75]. The toxicity of PrP118-135 is, however, independent of endogenous PrPC expression. Both peptides map in our predicted central aggregation-prone region of PrPC.
Conclusion
Overall, the method described here appears as a useful tool for the identification of protein regions that are especially relevant for protein aggregation and amyloidogenesis both in natively unfolded and properly folded globular proteins (Table 4). The results provide support to the hypothesis that short specific amino acid stretches can act as triggers for the incorporation of polypeptides into amyloid structures. It is interesting to note that in those cases in which structural information allows to delimitate the region incorporated in the core of the fibrillar structure, our predicted "hot spots" and those proved experimentally are considerably shorter than the whole region, suggesting that the role of "hot spots" is to act as specific nucleation points from which the ordered fibrillar structure is expanded.
Nature has provided globular proteins with a reasonable conformational stability in the native state in which, as proved here, aggregation-prone sequences are buried or involved in intra-molecular interactions. This appears as a very successful evolutive strategy to avoid aggregation, since few proteins aggregate from their stable native conformation. Accordingly, amyloid-related mutations in globular proteins usually result in destabilization of the folded state allowing the exposure of previously hidden "hot spots", as those reported here. This explains the scarce success in predicting the effect of mutations in the aggregation of globular proteins (data not shown), whereas the prediction of fatal sequence changes in intrinsically unstructured proteins involved in disease is generally accurate. The effects of such mutations can be explained in most cases by intrinsic factors, as they directly result in changes on the average propensity of the full polypeptide to aggregate.
Besides providing important clues about the mechanism of protein aggregation, this study may be relevant for the therapeutics of amyloid disease, since the identified "hot spots" could be regarded as preferential targets to tackle the deleterious disorders linked to protein deposition. According to our results, different specific strategies should be employed when designing methods to avoid aggregation, depending on the disease being caused by natively unfolded or by globular proteins. In Alzheimer, type II diabetes and Parkinson diseases, shielding the already exposed aggregation-prone regions in the polypeptides by using small compounds or antibodies appears as a promising approach, whereas compounds that will stabilize the native conformation and avoid the exposure of the deleterious "hot spots" will be more effective in the case of globular proteins. Additionally, when gene therapy eventually comes to age, mutations that disrupt aggregation-prone regions in unstructured polypeptides or those which over-stabilize the native state of globular aggregation-prone proteins are expected to be useful approaches to avoid protein deposition and meliorate neurodegenerative and systemic amyloidogenic disorders.
Methods
Experimental determination of amino acids aggregation propensities
The CHC of Aβ42 peptide was chosen as a paradigmatic aggregation-prone region for the calculation of the individual effect of each natural amino acid on protein aggregation. The specific effect on Aβ42's deposition promoted by the 20 different natural amino acids when located in the central position of this model "hot spot" were evaluated. Briefly, the wild type Aβ42 gene and its 19 mutants were inserted as a fusion protein upstream of the green fluorescence protein (GFP) and expressed individually in bacteria. In this system, the levels of GFP fluorescence in the cells depend exclusively on the in vivo aggregation propensity of the Aβ42 variant [10,25], in such a way that changes in aggregation propensities promoted by the different mutations can be easily monitored by measuring the fluorescence emission of the cells expressing each particular variant and normalizing it relative to that emitted by the cells bearing the wild type sequence. Three independent clones were analyzed for each mutation and each clone was analyzed at least by triplicate to generate consistent data. To obtain the individual aggregation propensities in Table 1, the change promoted by each amino acid was normalized relative to the average change of the pool of 20 amino acids.
Generation of aggregation profiles and identification of "hot spots"
Different experimental data suggest that the aggregation of Aβ42 occurs from a mostly unfolded conformation in which the CHC is exposed to solvent [76]. Assuming that the individual intrinsic aggregation propensities obtained analyzing this particular protein region will probably apply for any unfolded sequence; an aggregation profile was generated for every protein in this study through a simple assignment of the values in Table 1 to each individual residue in the corresponding sequences. Since "hot spots" are clusters of consecutive residues, the sequence was scanned by using a five residues sliding window. "Hot spots" in the sequence were identified as those protein regions at least five residues in length (the minimal size shown to date to be required for a peptide to form amyloid fibrils similar to those formed by whole polypeptides [77], in which the aggregation propensity is above the average aggregation propensity of the complete sequence. The average propensity of the polypeptide was calculated as the sum of the aggregation propensities of its individual amino acids divided by the number of residues.
Analysis of the effect of changes in the polypeptide sequence on aggregation
The concept of "hot spot" of aggregation implies that the contribution of a particular residue in a protein sequence on protein aggregation is somehow modulated by its immediate neighbors. According to this, the effects of mutation on protein aggregation can not be properly calculated by a simple subtraction of the intrinsic aggregation propensities of the wild type and mutant residues. Instead, to provide a more general description of the effect of the change on the overall aggregation propensity, the individual aggregation profiles for the wild type protein and the different mutants are obtained and the differences between the areas below the corresponding profiles are calculated. The area between each profile was always normalized by the number of residues in the considered species to compare between the aggregation propensities of the complete protein and fragments coming from proteolysis, chemical synthesis or other processes. The difference between normalized areas, multiplied by a 100 factor, was designed as the change in average aggregation propensity (ΔAP). ΔAP will be positive if the mutation is predicted to increase the aggregation propensity of the polypeptide chain and negative if it is predicted to increase solubility.
Abbreviations
AD Alzheimer's disease
Aβ Amyloid-β-protein
CHC Central hydrophobic cluster
FAD Familial Alzheimer diseases
FAP Amyloidotic polyneuropathy
GFP Green fluorescent protein
IAPP Islet amyloid polypeptide
NAC Non-Aβ component of amyloid plaques
PrP Prion protein
PrPSc Pathogenic prion form
SSA Senile systemic amyloidosis
TTR Transthyretin
ΔAP Change in average aggregation propensity
β2-m β2-Microglobulin
Authors' contributions
NSG and IP performed most of the experiments and prepared the final data and figures. FXA and JV contributed to data interpretation and manuscript redaction. SV directed the work and prepared the manuscript.
Acknowledgements
The vector expressing the Aβ42-GFP fusion was a generous gift of Michael Hecht's group. This work has been supported by Grants BIO2001-2046 and BIO2004-05879 (Ministerio de Ciencia y Tecnología, MCYT, Spain), by the Centre de Referència en Biotecnologia (Generalitat de Catalunya, Spain) and by PNL2004-40 (Universitat Autònoma de Barcelona (UAB). S.V. is supported by a "Ramón y Cajal" project awarded by the MCYT and co-financed by the UAB.
Figures and Tables
Figure 1 Aggregation profile of natively unfolded proteins related to disease. The average aggregation propensity of the different polypeptides is shown as a dashed line. Minimal protein regions which have been experimentally proven to be involved in aggregation are shown at the top of the plot as black bars. Regions in the core of the fibrils are shown as grey bars (when information available). The NAC fragment of α-Synuclein is shown as a dashed bar.
Figure 2 Representation of the 3D structure of globular proteins related to disease. The chain segments in which the prediction and the experimental data coincide are colored in green. Those identified experimentally to be relevant for amyloid formation but not predicted by the present approach are colored in blue. The regions predicted to be important for amyloid formation from which experimental data are not available or indicates that they are not involved in aggregation are shown in yellow.
Figure 3 Aggregation profile of globular proteins related to disease. Minimal protein regions which have been proved experimentally to be involved in aggregation are shown at the top of the plot as black bars. Regions in the core of the fibrils are shown as a grey bars (when information available).
Table 1 Relative experimental aggregation propensities of the 20 natural amino acids derived from the analysis of mutants in the central position of the CHC in amyloid-β-protein.
Amino acid
I 1.822
F 1.754
V 1.594
L 1.380
Y 1.159
W 1.037
M 0.910
C 0.604
A -0.036
T -0.159
S -0.294
P -0.334
G -0.535
K -0.931
H -1.033
Q -1.231
R -1.240
N -1.302
E -1.412
D -1.836
Table 2 Comparison of predicted and experimental changes in aggregation for Aβ variants.
Mutation ΔAP* Observed aggregation‡
A21G -1.22 -
E22G +2.14 +
E22Q +0.44 +
F19P -5.09 -
F19T -4.73 -
I31L -1.07 -
I32L -1.07 -
I41G -5.76 -
I41A -4.52 -
I41L -1.075 -
A42G -1.21 -
A42V +3.95 +
Δ1–4 +4.82 +
Δ1–9 +21.86 +
Δ40–42 -8.34 -
Δ41–42 -4.26 -
V12E+V18E+M35T+I41N -18.96 -
F19S+L34P -9.18 -
* Change in average aggregation propensity
‡ Changes in aggregation determined experimentally.
Table 3 Comparison of predicted and experimental changes in aggregation for IAPP variants, relative to the corresponding human IAPP sequence.
Variant ΔAP* Observed aggregation‡
(20–29) Cat -5.12 =
(20–29) Rat -16.46 -
(20–29) Hamster -32.73 -
R18H +0.94 +
L23F +1.70 +
V26I +0.42 +
R18H+L23F+V26I +3.06 +
(22–27) N22A +31.53 +
(22–27) F23A -42.96 -
(22–27) G24A +11.96 +
(22–27) I26A -44.5872 -
(22–27) L27A -33.99 -
S20G -1.09 +
ProIAPP +31.40 +?
* Change in average aggregation propensity
‡ Changes in aggregation determined experimentally.
? Not yet proved experimentally.
Table 4 List of the predicted "hot spots" in the different disease-linked polypeptides in this study and comparison with the available experimental data. Experimental "hot spots" refer to those protein regions shown to be involved in the aggregation process of the corresponding polypeptide. It is also noted if the predicted "hot spot" has been described as a structural element of the amyloid fibrils formed by the different peptides and proteins in the study.
Protein Predicted "Hot Spots" Experimental "Hot Spots" Regions in the fibrils
Amyloid-β-protein 16–21 + +
30–36 + +
38–42 + +
Islet amyloid polypeptide 12–18 + uncertain
22–28 + uncertain
1–18 No experimental data available uncertain
α-Synuclein 27–56 + uncertain
61–94 + +
β2-Microgobulin 21–31 + +
56–69 + +
79–85 + +
87–91 + +
Lysozyme (hen) 24–34 - -
50–62 + +
76–98 + +
Transthyretin 10–20 + +
23–33 No experimetal data available uncertain
105–118 + +
Prion Protein 1–32 No experimetal data available uncertain
105–146 + +
208–252 No experimetal data available uncertain
==== Refs
Smith A protein misfolding Nature 2003 426 883 883 10.1038/426883a
Rochet JC Lansbury PT Jr Amyloid fibrillogenesis: themes and variations Curr Opin Struct Biol 2000 10 60 68 10679462 10.1016/S0959-440X(99)00049-4
Uversky VN Fink AL Conformational constraints for amyloid fibrillation: the importance of being unfolded Biochim Biophys Acta 2004 1698 131 153 15134647
Kelly JW The alternative conformations of amyloidogenic proteins and their multi-step assembly pathways Curr Opin Struct Biol 1998 8 101 106 9519302 10.1016/S0959-440X(98)80016-X
Ventura S Zurdo J Narayanan S Parreno M Mangues R Reif B Chiti F Giannoni E Dobson CM Aviles FX Serrano L Short amino acid stretches can mediate amyloid formation in globular proteins: the Src homology 3 (SH3) case Proc Natl Acad Sci U S A 2004 101 7258 7263 15123800 10.1073/pnas.0308249101
Ivanova MI Sawaya MR Gingery M Attinger A Eisenberg D An amyloid-forming segment of beta2-microglobulin suggests a molecular model for the fibril Proc Natl Acad Sci U S A 2004 101 10584 10589 15249659 10.1073/pnas.0403756101
Dobson CM Protein misfolding, evolution and disease Trends Biochem Sci 1999 24 329 332 10470028 10.1016/S0968-0004(99)01445-0
Linding R Schymkowitz J Rousseau F Diella F Serrano L A comparative study of the relationship between protein structure and beta-aggregation in globular and intrinsically disordered proteins J Mol Biol 2004 342 345 353 15313629 10.1016/j.jmb.2004.06.088
Fernandez-Escamilla AM Rousseau F Schymkowitz J Serrano L Prediction of sequence-dependent and mutational effects on the aggregation of peptides and proteins Nat Biotechnol 2004 22 1302 1306 15361882 10.1038/nbt1012
de Groot N Aviles FX Vendrell J Ventura S 2005 Submitted
Bitan G Vollers SS Teplow DB Elucidation of primary structure elements controlling early amyloid beta-protein oligomerization J Biol Chem 2003 278 34882 34889 12840029 10.1074/jbc.M300825200
Chiti F Stefani M Taddei N Ramponi G Dobson CM Rationalization of the effects of mutations on peptide and protein aggregation rates Nature 2003 424 805 808 12917692 10.1038/nature01891
Fink AL Protein aggregation: folding aggregates, inclusion bodies and amyloid Fold Des 1998 3 R9 23 9502314 10.1016/S1359-0278(98)00002-9
Pawar AP Dubay KF Zurdo J Chiti F Vendruscolo M Dobson CM Prediction of "Aggregation-prone" and "Aggregation-susceptible" Regions in Proteins Associated with Neurodegenerative Diseases J Mol Biol 2005 350 379 392 15925383 10.1016/j.jmb.2005.04.016
Selkoe DJ Alzheimer's disease: genes, proteins, and therapy Physiol Rev 2001 81 741 766 11274343
Nagele RG Wegiel J Venkataraman V Imaki H Wang KC Wegiel J Contribution of glial cells to the development of amyloid plaques in Alzheimer's disease Neurobiol Aging 2004 25 663 674 15172746 10.1016/j.neurobiolaging.2004.01.007
Williams AD Portelius E Kheterpal I Guo JT Cook KD Xu Y Wetzel R Mapping abeta amyloid fibril secondary structure using scanning proline mutagenesis J Mol Biol 2004 335 833 842 14687578 10.1016/j.jmb.2003.11.008
Morimoto A Irie K Murakami K Masuda Y Ohigashi H Nagao M Fukuda H Shimizu T Shirasawa T Analysis of the secondary structure of beta-amyloid (Abeta42) fibrils by systematic proline replacement J Biol Chem 2004 279 52781 52788 15459202 10.1074/jbc.M406262200
Petkova AT Ishii Y Balbach JJ Antzutkin ON Leapman RD Delaglio F Tycko R A structural model for Alzheimer's beta-amyloid fibrils based on experimental constraints from solid state NMR Proc Natl Acad Sci U S A 2002 99 16742 16747 12481027 10.1073/pnas.262663499
Torok M Milton S Kayed R Wu P McIntire T Glabe CG Langen R Structural and dynamic features of Alzheimer's Abeta peptide in amyloid fibrils studied by site-directed spin labeling J Biol Chem 2002 277 40810 40815 12181315 10.1074/jbc.M205659200
Balbach JJ Ishii Y Antzutkin ON Leapman RD Rizzo NW Dyda F Reed J Tycko R Amyloid fibril formation by A beta 16–22, a seven-residue fragment of the Alzheimer's beta-amyloid peptide, and structural characterization by solid state NMR Biochemistry 2000 39 13748 13759 11076514 10.1021/bi0011330
Findeis MA Musso GM Arico-Muendel CC Benjamin HW Hundal AM Lee JJ Chin J Kelley M Wakefield J Hayward NJ Molineaux SM Modified-peptide inhibitors of amyloid beta-peptide polymerization Biochemistry 1999 38 6791 6800 10346900 10.1021/bi982824n
Jarrett JT Berger EP Lansbury PT Jr The carboxy terminus of the beta amyloid protein is critical for the seeding of amyloid formation: implications for the pathogenesis of Alzheimer's disease Biochemistry 1993 32 4693 4697 8490014 10.1021/bi00069a001
Yamamoto N Hasegawa K Matsuzaki K Naiki H Yanagisawa K Environment- and mutation-dependent aggregation behavior of Alzheimer amyloid beta-protein J Neurochem 2004 90 62 9 15198667 10.1111/j.1471-4159.2004.02459.x
Wurth C Guimard NK Hecht MH Mutations that reduce aggregation of the Alzheimer's Abeta42 peptide: an unbiased search for the sequence determinants of Abeta amyloidogenesis J Mol Biol 2002 319 1279 1290 12079364 10.1016/S0022-2836(02)00399-6
Clark A Cooper GJ Lewis CE Morris JF Willis AC Reid KB Turner RC Islet amyloid formed from diabetes-associated peptide may be pathogenic in type-2 diabetes Lancet 1987 2 231 234 2441214 10.1016/S0140-6736(87)90825-7
Goldsbury C Goldie K Pellaud J Seelig J Frey P Muller SA Kistler J Cooper GJ Aebi U Amyloid fibril formation from full-length and fragments of amylin J Struct Biol 2000 130 352 362 10940238 10.1006/jsbi.2000.4268
Kajava AV Aebi U Steven AC The parallel superpleated beta-structure as a model for amyloid fibrils of human amylin J Mol Biol 2005 348 247 252 15811365 10.1016/j.jmb.2005.02.029
Scrocchi LA Ha K Chen Y Wu L Wang F Fraser PE Identification of minimal peptide sequences in the (8–20) domain of human islet amyloid polypeptide involved in fibrillogenesis J Struct Biol 2003 141 218 27 12648568 10.1016/S1047-8477(02)00630-5
Tracz SM Abedini A Driscoll M Raleigh DP Role of aromatic interactions in amyloid formation by peptides derived from human Amylin Biochemistry 2004 43 15901 15908 15595845 10.1021/bi048812l
Moriarty DF Raleigh DP Effects of sequential proline substitutions on amyloid formation by human amylin 20-29 Biochemistry 1999 38 1811 1818 10026261 10.1021/bi981658g
Azriel R Gazit E Analysis of the minimal amyloid-forming fragment of the islet amyloid polypeptide. An experimental support for the key role of the phenylalanine residue in amyloid formation J Biol Chem 2001 276 34156 34161 11445568 10.1074/jbc.M102883200
Green J Goldsbury C Mini T Sunderji S Frey P Kistler J Cooper G Aebi U Full-length rat amylin forms fibrils following substitution of single residues from human amylin J Mol Biol 2003 326 1147 1156 12589759 10.1016/S0022-2836(02)01377-3
Sakagashira S Sanke T Hanabusa T Shimomura H Ohagi S Kumagaye KY Nakajima K Nanjo K Missense mutation of amylin gene (S20G) in Japanese NIDDM patients Diabetes 1996 45 1279 1281 8772735
Porte D JrKahn SE Hyperproinsulinemia and amyloid in NIDDM. Clues to etiology of islet beta-cell dysfunction? Diabetes 1989 38 1333 1336 2695369
Spillantini MG Schmidt ML Lee VM Trojanowski JQ Jakes R Goedert M Alpha-synuclein in Lewy bodies Nature 1997 388 839 840 9278044 10.1038/42166
Goedert M Alpha-synuclein and neurodegenerative diseases Nat Rev Neurosci 2001 2 492 501 11433374 10.1038/35081564
Bodles AM Guthrie DJ Greer B Irvine GB Identification of the region of non-Abeta component (NAC) of Alzheimer's disease amyloid responsible for its aggregation and toxicity J Neurochem 2001 78 384 395 11461974 10.1046/j.1471-4159.2001.00408.x
Giasson BI Murray IV Trojanowski JQ Lee VM A hydrophobic stretch of 12 amino acid residues in the middle of alpha-synuclein is essential for filament assembly J Biol Chem 2001 276 2380 2386 11060312 10.1074/jbc.M008919200
Miake H Mizusawa H Iwatsubo T Hasegawa M Biochemical characterization of the core structure of alpha-synuclein filaments J Biol Chem 2002 277 19213 10219 11893734 10.1074/jbc.M110551200
Der-Sarkissian A Jao CC Chen J Langen R Structural organization of alpha-synuclein fibrils studied by site-directed spin labeling J Biol Chem 2003 278 37530 37535 12815044 10.1074/jbc.M305266200
Choi W Zibaee S Jakes R Serpell LC Davletov B Crowther RA Goedert M Mutation E46K increases phospholipid binding and assembly into filaments of human alpha-synuclein FEBS Lett 2004 576 363 8 15498564 10.1016/j.febslet.2004.09.038
Cuervo AM Stefanis L Fredenburg R Lansbury PT Sulzer D Impaired degradation of mutant alpha-synuclein by chaperone-mediated autophagy Science 2004 305 1292 1295 15333840 10.1126/science.1101738
Saha AR Hill J Utton MA Asuni AA Ackerley S Grierson AJ Miller CC Davies AM Buchman VL Anderton BH Hanger DP Parkinson's disease alpha-synuclein mutations exhibit defective axonal transport in cultured neurons J Cell Sci 2004 117 1017 1024 14996933 10.1242/jcs.00967
Huang Y Cheung L Rowe D Halliday G Genetic contributions to Parkinson's disease Brain Res Brain Res Rev 2004 46 44 70 15297154 10.1016/j.brainresrev.2004.04.007
Koch KM Dialysis-related amyloidosis Kidney Int 1992 41 1416 1429 1614057
McParland VJ Kad NM Kalverda AP Brown A Kirwin-Jones P Hunter MG Sunde M Radford SE Partially unfolded states of beta(2)-microglobulin and amyloid formation in vitro Biochemistry 2000 39 8735 8746 10913285 10.1021/bi000276j
Kozhukh GV Hagihara Y Kawakami T Hasegawa K Naiki H Goto Y Investigation of a peptide responsible for amyloid fibril formation of beta 2-microglobulin by achromobacter protease I J Biol Chem 2002 277 1310 1315 11687582 10.1074/jbc.M108753200
Hasegawa K Ohhashi Y Yamaguchi I Takahashi N Tsutsumi S Goto Y Gejyo F Naiki H Amyloidogenic synthetic peptides of beta2-microglobulin–a role of the disulfide bond Biochem Biophys Res Commun 2003 304 101 106 12705891 10.1016/S0006-291X(03)00543-6
Jones S Manning J Kad NM Radford SE Amyloid-forming peptides from beta2-microglobulin-Insights into the mechanism of fibril formation in vitro J Mol Biol 2003 325 249 57 12488093 10.1016/S0022-2836(02)01227-5
Ivanova MI Gingery M Whitson LJ Eisenberg D Role of the C-terminal 28 residues of beta2-microglobulin in amyloid fibril formation Biochemistry 2003 42 13536 13540 14622000 10.1021/bi0301486
Bellotti V Stoppini M Mangione P Sunde M Robinson C Asti L Brancaccio D Ferri G Beta2-microglobulin can be refolded into a native state from ex vivo amyloid fibrils Eur J Biochem 1998 258 61 67 9851692 10.1046/j.1432-1327.1998.2580061.x
Ivanova MI Sawaya MR Gingery M Attinger A Eisenberg D An amyloid-forming segment of beta2-microglobulin suggests a molecular model for the fibril Proc Natl Acad Sci U S A 2004 101 10584 10589 15249659 10.1073/pnas.0403756101
Yamamoto S Hasegawa K Yamaguchi I Tsutsumi S Kardos J Goto Y Gejyo F Naiki H Low concentrations of sodium dodecyl sulfate induce the extension of beta 2-microglobulin-related amyloid fibrils at a neutral pH Biochemistry 2004 43 11075 11082 15323566 10.1021/bi049262u
Kihara M Chatani E Sakai M Hasegawa K Naiki H Goto Y Seeding-dependent maturation of beta2-microglobulin amyloid fibrils at neutral pH J Biol Chem 2005 280 12012 12018 15659393 10.1074/jbc.M411949200
Pepys MB Hawkins PN Booth DR Vigushin DM Tennent GA Soutar AK Totty N Nguyen O Blake CCF Terry CJ Feest TG Zalin AM Hsuan JJ Human lysozyme gene mutations cause hereditary systemic amyloidosis Nature 1993 362 553 557 8464497 10.1038/362553a0
Booth DR Sunde M Bellotti V Robinson CV Hutchinson WL Fraser PE Hawkins PN Dobson CM Radford SE Blake CC Pepys MB Instability, unfolding and aggregation of human lysozyme variants underlying amyloid fibrillogenesis Nature 1997 385 787 793 9039909 10.1038/385787a0
Canet D Last AM Tito P Sunde M Spencer A Archer DB Redfield C Robinson CV Dobson CM Local cooperativity in the unfolding of an amyloidogenic variant of human lysozyme Nat Struct Biol 2002 9 308 315 11887182 10.1038/nsb768
Yang JJ Buck M Pitkeathly M Kotik M Haynie DT Dobson CM Radford SE Conformational properties of four peptides spanning the sequence of hen lysozyme J Mol Biol 1995 29 483 491 10.1006/jmbi.1995.0513
Krebs MR Wilkins DK Chung EW Pitkeathly MC Chamberlain AK Zurdo J Robinson CV Dobson CM Formation and seeding of amyloid fibrils from wild-type hen lysozyme and a peptide fragment from the beta-domain J Mol Biol 2000 300 541 549 10884350 10.1006/jmbi.2000.3862
Frare E Polverino De Laureto P Zurdo J Dobson CM Fontana A A highly amyloidogenic region of hen lysozyme J Mol Biol 2004 340 1153 1165 15236974 10.1016/j.jmb.2004.05.056
Kallberg Y Gustafsson M Persson B Thyberg J Johansson J Prediction of amyloid fibril-forming proteins J Biol Chem 2001 276 12945 12950 11134035 10.1074/jbc.M010402200
Saraiva MJ Birken S Costa PP Goodman DS Amyloid fibril protein in familial amyloidotic polyneuropathy, Portuguese type. Definition of molecular abnormality in transthyretin (prealbumin) J Clin Invest 1984 74 104 119 6736244
McCutchen SL Lai Z Miroy GJ Kelly JW Colon W Comparison of lethal and nonlethal transthyretin variants and their relationship to amyloid disease Biochemistry 1995 34 13527 13536 7577941 10.1021/bi00041a032
Hornberg A Eneqvist T Olofsson A Lundgren E Sauer-Eriksson AE A comparative analysis of 23 structures of the amyloidogenic protein transthyretin J Mol Biol 2000 22 649 669 10.1006/jmbi.2000.4078
Richardson JS Richardson DC Natural beta-sheet proteins use negative design to avoid edge-to-edge aggregation Proc Natl Acad Sci U S A 2002 99 2754 2759 11880627 10.1073/pnas.052706099
Jaroniec CP MacPhee CE Astrof NS Dobson CM Griffin RG Molecular conformation of a peptide fragment of transthyretin in an amyloid fibril Proc Natl Acad Sci U S A 2002 99 16748 16753 12481032 10.1073/pnas.252625999
Jaroniec CP MacPhee CE Bajaj VS McMahon MT Dobson CM Griffin RG High-resolution molecular structure of a peptide in an amyloid fibril determined by magic angle spinning NMR spectroscopy Proc Natl Acad Sci U S A 2004 101 711 716 14715898 10.1073/pnas.0304849101
Jarvis JA Kirkpatrick A Craik DJ 1H NMR analysis of fibril-forming peptide fragments of transthyretin Int J Pept Protein Res 1994 44 388 39 7875942
Prusiner SB Scott MR DeArmond SJ Cohen FE Prion protein biology Cell 1998 93 337 348 9590169 10.1016/S0092-8674(00)81163-0
Zahn R Liu A Luhrs T Riek R von Schroetter C Lopez Garcia F Billeter M Calzolai L Wider G Wuthrich K NMR solution structure of the human prion protein Proc Natl Acad Sci U S A 2000 97 145 150 10618385 10.1073/pnas.97.1.145
Kaneko K Zulianello L Scott M Cooper CM Wallace AC James TL Cohen FE Prusiner SB Evidence for protein X binding to a discontinuous epitope on the cellular prion protein during scrapie prion propagation Proc Natl Acad Sci U S A 1997 94 10069 10074 9294164 10.1073/pnas.94.19.10069
Tagliavini F Prelli F Verga L Giaccone G Sarma R Gorevic P Ghetti B Passerini F Ghibaudi E Forloni G Salmona M Bugiani O Frangione B Synthetic peptides homologous to prion protein residues 106–147 form amyloid-like fibrils in vitro Proc Natl Acad Sci U S A 1993 90 9678 9682 8105481
Gu Y Fujioka H Mishra RS Li R Singh N Prion peptide 106–126 modulates the aggregation of cellular prion protein and induces the synthesis of potentially neurotoxic transmembrane PrP J Biol Chem 2002 277 2275 86 11682469 10.1074/jbc.M104345200
Chabry J Ratsimanohatra C Sponne I Elena PP Vincent JP Pillot T In vivo and in vitro neurotoxicity of the human prion protein (PrP) fragment P118-135 independently of PrP expression J Neurosci 2003 23 462 469 12533606
Santini S Wei G Mousseau N Derreumaux P Pathway complexity of Alzheimer's beta-amyloid Abeta16-22 peptide assembly Structure 2004 12 1245 1255 15242601 10.1016/j.str.2004.04.018
Haspel N Zanuy D Ma B Wolfson H Nussinov R A comparative study of amyloid fibril formation by residues 15–19 of the human calcitonin hormone: a single beta-sheet model with a small hydrophobic core J Mol Biol 2005 345 1213 1227 15644216 10.1016/j.jmb.2004.11.002
|
16197548
|
PMC1262731
|
CC BY
|
2021-01-04 16:03:53
|
no
|
BMC Struct Biol. 2005 Sep 30; 5:18
|
utf-8
|
BMC Struct Biol
| 2,005 |
10.1186/1472-6807-5-18
|
oa_comm
|
==== Front
BMC Vet ResBMC Veterinary Research1746-6148BioMed Central London 1746-6148-1-31618804410.1186/1746-6148-1-3Research ArticleSelection of ovine housekeeping genes for normalisation by real-time RT-PCR; analysis of PrP gene expression and genetic susceptibility to scrapie Garcia-Crespo David [email protected] Ramón A [email protected] Ana [email protected] Department of Animal Health, Instituto Vasco de Investigación y Desarrollo Agrario (NEIKER); Berreaga, 1. 48160 Derio, Bizkaia, Spain2005 28 9 2005 1 3 3 15 6 2005 28 9 2005 Copyright © 2005 Garcia-Crespo et al; licensee BioMed Central Ltd.2005Garcia-Crespo et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Cellular prion protein expression is essential for the development of transmissible spongiform encephalopathies (TSEs), and in sheep, genetic susceptibility to scrapie has been associated to PrP gene polymorphisms. To test the hypothetical linkage between PrP gene expression and genetic susceptibility, PrP mRNA levels were measured by real-time RT-PCR in six ovine tissues of animals with different genotypes.
Results
Previous to the PrP gene expression analysis the stability of several housekeeping (HK) genes was assessed in order to select the best ones for relative quantification. The normalisation of gene expression was carried out using a minimum of three HK genes in order to detect small expression differences more accurately than using a single control gene. The expression stability analysis of six HK genes showed a large tissue-associated variation reflecting the existence of tissue-specific factors. Thereby, a specific set of HK genes was required for an accurate normalisation of the PrP gene expression within each tissue. Statistical differences in the normalised PrP mRNA levels were found among the tissues, obtaining the highest expression level in obex, followed by ileum, lymph node, spleen, cerebellum and cerebrum. A tendency towards increased PrP mRNA levels and genetic susceptibility was observed in central nervous system. However, the results did not support the hypothesis that PrP mRNA levels vary between genotypes.
Conclusion
The results on PrP gene expression presented here provide valuable baseline data for future studies on scrapie pathogenesis. On the other hand, the results on stability data of several HK genes reported in this study could prove very useful in other gene expression studies carried out in these relevant ovine tissues.
==== Body
Background
Scrapie is a neurodegenerative disease of the group of transmissible spongiform encephalopathies (TSEs) that affects sheep and goats [1,2]. The lesions appear mainly in the nervous system in the form of vacuoles triggered by the conversion of the cellular prion protein (PrPc) into the abnormal isoform (PrPSc) followed by its pathological accumulation [2]. Although the exact origin of the disease remains unknown, the "protein only" hypothesis supports the concept that PrPSc is the transmissible agent causing the disease [2]. Therefore, the presence of PrPc is essential to develop scrapie [3,4]. However, little is known about the physiological role of PrPc and basic PrPc regulating mechanisms.
The development of clinical signs of scrapie has been linked to some PrP gene polymorphisms [5,6]. According to this genetic profile, animals have been classified into risk groups from Type 1 to Type 5 in increasing order of susceptibility to scrapie according to the UK National Scrapie Plan (NSP) [7]. The reasons for the different genetic susceptibility have been assessed in vitro revealing low conversion efficiencies of PrPc into PrPSc in resistant genotypes [8,9]. However, individual factors, and unknown genes or proteins might be involved in this genetic susceptibility.
The oral route is the main pathway of transmission of prions in nature. Once the agent has entered into the host, an early amplification occurs in the lymphoreticular system followed by the subsequent spread to several tissues through lymphatic routes, blood or peripheral nervous system. This precedes replication in the central nervous system (CNS) [10,11]. Taking into account that the presence of PrPc is essential to develop scrapie [3,4], knowing the distribution of PrPc in lymphoid and nervous tissues is relevant to understanding the pathogenesis of the disease. Likewise, PrP transcript levels and subsequent translation product abundance might play an important role in the transmission and development of the disease. In this sense, absolute quantification studies on bovine and golden hamster PrP gene expression revealed high levels of PrP mRNA in CNS and lymphoid tissues [12,13], however, these results must be considered with caution due to possible artefacts of the quantification method used. Several methods and techniques can be used to measure mRNA levels. Northern blot analysis has traditionally been used, but more recently, real-time RT-PCR technology provides higher sensitivity and more accurate expression profiles [14,15]. In contrast to absolute quantification, relative quantification is not influenced by artefacts during sample preparation and it provides the means to detect small expression differences. However, this technique is more demanding than absolute quantification in the sense that it requires the selection of highly stable housekeeping (HK) genes to normalise the expression of the target gene. Although several studies on ovine gene expression have been reported using a single HK gene, the use of at least three stable HK genes is more suitable [16].
The aim of this work was to select and evaluate the stability of several ovine HK genes for relative expression analyses, and use them to test the hypothetical linkage between PrP gene expression and genetic susceptibility to scrapie.
Results
Primer optimisation and amplification specificity
Primer concentrations that generated the lowest Ct value and a sharp peak, but lacked non-specific fragments and primer-dimers were selected (Table 1). The analysis of melting temperatures, amplicon sizes and sequencing data demonstrated the specificity of the PCR reactions. The efficiency values obtained for the real-time PCR amplification of the six HK genes and PrP gene were near to 2. Efficiency values (E), slope values and correlation coefficient (R2) for each primer pair are shown in Table 1.
Table 1 Primers sequences and real-time PCR amplification parameters
Gene Forward & reverse primers 5' → 3' [C] a Amplicon size (bp) Tm b (°C) Slope R2c E d
ACTB ATGCCTCCTGCACCACCA
GCATTTGCGGTGGACGAT 300 125 85 -3.597 0.999 1.897
YWHAZ TGTAGGAGCCCGTAGGTCATCT
TTCTCTCTGTATTCTCGAGCCATCT 100 102 79 -3.335 0.988 1.995
RPL19 CAACTCCCGCCAGCAGAT
CCGGGAATGGACAGTCACA 200 76 83 -3.342 0.992 1.992
GAPDH ATGCCTCCTGCACCACCA
AGTCCCTCCACGATGCCAA 100 76 84 -3.485 0.991 1.936
G6PDH TGACCTATGGCAACCGATACAA
CCGCAAAAGACATCCAGGAT 300 76 81 -3.363 0.965 1.983
SDHA CATCCACTACATGACGGAGCA
ATCTTGCCATCTTCAGTTCTGCTA 200 90 82 -3.643 0.992 1.881
PrP GCCAAAAACCAACATGAAGCAT
TGCTCATGGCACTTCCCAG 300 95 83 -3.338 0.995 1.993
a Primer concentrations in nM
b Theoretical amplicon melting temperature calculated with Primer Express software (Applied Biosystems)
c Correlation coefficient
d PCR efficiency
Selection of the optimal HK genes and normalisation of gene expression
The analysis of the expression of the HK genes in all six tissues in the 22 animals showed a pairwise variation above the cut off value (Vn/n+1> 0.15) established by Vandesompele et al. [16]. This indicated the invalidity of using a common set of HK genes for all the tissues and therefore, the stability of the six HK genes was assessed within each tissue. The initial comparison of the M values for the six genes (Table 2) showed a large tissue-associated variation in the expression stability of some genes, which in some cases showed even opposite values. The stepwise exclusion of the less stable HK gene according to the geNorm application showed six different stability series, one for each tissue (Table 3), confirming the tissue-associated variation. For instance, the G6PDH gene showed high stability in cerebellum and ileum while it was the less stable gene in spleen. In addition, the frequently used HK gene GAPDH showed the smallest variation when the M value from each tissue was compared (Table 2) and it was selected for the normalisation in five of the six tissues analysed (Table 3), confirming its high stability. On the other hand, another traditionally used HK gene, ACTB, showed the second highest standard deviation among tissues in comparison to other HK genes.
Table 2 Expression stability values (M) of the six candidate HK genes
Tissue ACTB YWHAZ RPL19 SDHA GAPDH G6PDH
Cerebrum 0.602 0.757 0.634 0.540 0.510 0.660
Cerebellum 0.477 0.480 0.452 0.518 0.473 0.457
Obex 0.534 0.493 0.593 0.541 0.451 0.479
Spleen 0.436 0.413 0.550 0.427 0.448 0.569
Mesenteric lymph node 0.556 0.691 0.542 0.525 0.502 0.624
Ileum 0.757 0.652 0.696 0.610 0.517 0.511
Mean 0.560 0.581 0.578 0.527 0.484 0.550
SD a 0.113 0.137 0.084 0.059 0.030 0.081
CV (%)b 20.133 23.669 14.543 11.192 6.268 14.771
M values in this table were calculated for the six HK genes previous to the stepwise exclusion of the less stable HK gene.
a Standard deviation
b Coefficient of variation
Table 3 HK genes stability series for each tissue
Cerebrum Cerebellum Obex Spleen Mesenteric lymph node Ileum
GAPDH-SDHA G6PDH-ACTB GAPDH-YWHAZ GAPDH-SDHA SDHA-RPL19 GAPDH-G6PDH
ACTB YWHAZ G6PDH ACTB GAPDH SDHA
G6PDH RPL19 SDHA YWHAZ ACTB YWHAZ
RPL19 GAPDH ACTB RPL19 G6PDH RPL19
YWHAZ SDHA RPL19 G6PDH YWHAZ ACTB
The stability series are shown from the most stable gene at the top to the least stable gene at the bottom ranked to their expression stability estimated using geNorm. The first two genes in each series cannot be ranked because of the required use of gene ratios for gene stability measurements. The HK genes required for a reliable normalisation of the target gene expression in each tissue are shown in bold.
Normalised PrP gene expression analysis
Classification of animals according to their risk group and results from the expression analyses are listed and graphically represented in Table 4 and Figure 1, respectively. A marked association between PrP mRNA level and the type of tissue (p < 0.0001) was found in the overall analysis including all risk levels. In this sense, the obex showed the highest expression level (38.05) followed by ileum (35.73), lymph node (33.51), spleen (29.99), cerebellum (28.89) and cerebrum (21.58). When the risk group effect was analysed, no significant association was found between it and the PrP mRNA levels. The model did not show any interaction between tissue type and risk group.
Figure 1 PrP gene expression levels in the different tissues of 22 sheep grouped in risk groups. The PrP mRNA levels were obtained by relative quantification real-time RT-PCR analysis using the most stable HK genes within each tissue. Error bars represent standard deviation. Risk groups according to the NSP classification [7].
Table 4 Relative mRNA expression levels of PrP gene (arbitrary units)
Risk group Cerebrum Cerebellum Obex Spleen Ileum Mesenteric lymph node Nervous tissues Lymphoid tissues
n a nPrP b (SDc) nPrP (SD) nPrP (SD) nPrP (SD) nPrP (SD) nPrP (SD) nPrP (SD) nPrP (SD)
1 3 15.829 (4.279) 21.202 (3.052) 29.892 (12.223) 38.851 (16.115) 42.048 (8.800) 31.424 (3.496) 22.308 (9.057) 37.442 (10.470)
2 5 27.132 (4.830) 26.632 (3.495) 37.897 (6.930) 24.504 (16.953) 33.659 (13.712) 31.940 (4.088) 30.554 (7.267) 30.035 (12.551)
3 7 21.025 (5.869) 33.088 (6.515) 42.435 (7.598) 33.468 (18.426) 36.318 (13.434) 36.971 (6.320) 32.183 (11.001) 35.586 (13.054)
5 7 20.645 (5.092) 29.584 (4.439) 37.283 (8.811) 26.628 (10.323) 33.897 (9.467) 32.051 (5.895) 29.171 (9.247) 30.859 (8.903)
Total 22 21.584 (5.991) 28.885 (6.055) 38.054 (8.841) 29.989 (15.268) 35.725 (11.328) 33.506 (5.621) 29.508 (9.736) 33.073 (11.517)
Risk groups are listed from the most resistant one on the top to the most susceptible at the bottom according to the NSP classification [7]. Type 1, included three ARR/ARR animals; Type 2, included five ARR/ARQ animals; Type 3, included one ARH/ARH, one ARQ/ARH and five ARQ/ARQ animals; Type 5, included one ARH/VRQ, five ARQ/VRQ and one VRQ/VRQ animal.
a Number of animals
b Normalised PrP mRNA level
c Standard deviation
When the expression values were considered within each tissue, several specific significant (p < 0.05) or marginally significant (p < 0.10) differences were found between pairwise comparisons. In cerebral samples, animals from Type 1 showed the lowest PrP gene expression value, followed by animals from Type 5, Type 3 and Type 2, but only the lower expression level in Type 1 than in Type 2 animals was marginally significant (p = 0.0700). In cerebellum samples, Type 1 showed the lowest expression value followed by Type 2, Type 5 and Type 3. Expression level in Type 1 was marginally lower than in Type 3 (p = 0.0668). In obex samples, Type 1 showed the lowest expression value followed by Type 5, Type 2 and Type 3. Again, expression level in Type 1 was marginally lower than in Type 3 (p = 0.0633). In spleen samples, Type 1 showed the highest expression value followed by Type 3, Type 5, and Type 2. Expression in Type 1 was significantly higher than in Type 2 (p = 0.0281) and in Type 5 (p = 0.0671), and in Type 2 lower than in Type 3 (p = 0.0937). In ileum samples Type 1 showed the highest expression value followed by Type 3, Type 5, and Type 2. In mesenteric lymph node samples Type 1 showed the lowest expression value followed by Type 2, Type 5, and Type 3. When data were grouped in nervous tissues (cerebrum, cerebellum and obex) versus lymphoid tissues (spleen, ileum and lymph node), the models showed a relationship between the PrP mRNA levels and the effect of tissue type (p = 0.0160). A statistical association was found for the interaction between tissue type and risk group (p = 0.0623). PrP mRNA level in lymphoid tissues was significantly higher than in nervous tissues (p = 0.0160). In nervous tissues, Type 3 showed the highest expression value followed by Type 2, Type 5, and Type 1. The expression found in Type 1 was lower than in Type 2 (p = 0.0496), Type 3 (p = 0.0174), and Type 5 (p = 0.0863). Regarding lymphoid tissues samples, Type 1 showed the highest expression value followed by Type 3, Type 5, and Type 2. Expression in Type 1 was higher than in Type 2 (p = 0.0840).
Discussion
Real-time RT-PCR was chosen among several techniques available to measure the mRNA levels of PrP gene in six important tissues for the transmission and the development of scrapie. Real-time RT-PCR technology provides high sensitivity and accurate expression profiles [14,15] and in that approach, two basic protocols can be followed: absolute quantification and relative quantification. For gene expression studies, relative quantification is more suitable because the influence of unavoidable artefacts during sample preparation is taken into account. Several works on ovine gene expression have been carried out using the common practice of normalising with a single control gene like 18S rRNA, GAPDH or ACTB [17-19]. However, inter-individual variation of traditionally considered stable HK genes can be high enough to bias gene expression profiles when calculated using only one HK gene for normalisation. Therefore, the use of more than a single HK gene is recommended particularly to detect small expression differences more accurately. Thereby, the sensitivity of this approach depends on how well the HK genes are selected. Thus, in this study a robust method described by Vandesompele et al. [16] has been followed where the use of a minimum of three stable HK genes is required for an accurate normalisation of the target gene expression after assessing the stability of a given set of HK genes.
The expression stability observed in the six HK genes analysed in the present work varied with the tissue, and therefore, different sets of HK genes were necessary to normalise the PrP gene expression within each tissue. This variability of the HK genes stability among samples from a variety of sources is consistent with the literature [15,20-22] and reflects the existence of a tissue-specific metabolism and/or unknown tissue-specific factors. These findings clearly demonstrate that there is no single universal control gene for all tissues or cell types. Thereby, these inherent variations have to be taken into consideration and the stability of the HK genes needs to be studied in each scenario prior to any relative quantification study in order to obtain results as accurate as possible. GAPDH and ACTB have been traditionally considered invariable (equally expressed) genes and consequently, they have been widely used as single control genes for gene expression in many studies. However, the expression of these genes can vary from 7 to 23-fold depending on the cell type or tissue [21]. In our study, GAPDH gene showed the lowest variation among the panel of six tissues, whereas ACTB showed the penultimate worst score in variability. This is in agreement with the reported invalidity of ACTB gene for gene expression in ovine interstitial cells from heart valves [23]. Consistent with this variation and in order to improve the accuracy of our results, we normalised the PrP gene expression using the most stable sets of HK genes in each of the six tissues analysed.
In order to test the hypothesis of the linkage between PrP gene expression and PrP genotype-associated susceptibility, 22 Latxa sheep with different genotypes were analysed. Special care was taken with the statistical analysis because the number of samples available for our study was not too large. In this sense, a general linear model was used to control all the effects for which information was available in order to reduce risks of errors linked to repeated separate comparisons and to guarantee that no type α errors were committed. The results revealed statistically significant differences in the PrP gene expression among the panel of six tissues. The highest PrP mRNA expression level in CNS samples was found in obex followed by cerebellum and cerebrum. This circumstance might be translated into high levels of PrPc and perhaps into high levels of PrPSc aggregates considering PrPc as the substrate for conversion into the pathologic isoform. The different PrPc content in the three tissues of the CNS would be in accordance with the spatiotemporal appearance of PrPSc aggregates and would also support the idea that the obex is the best source of material for the detection of PrPSc in TSE rapid tests analyses.
Factors inherent in the nature of the different tissues like transcripts stability or postranscriptional regulation of the PrP gene have also to be considered. In this context, some studies focused on PrP mRNA have shown that there are two PrP mRNA transcripts (2.1 and 4.6 kb mRNA) with a tissue-specific distribution, different stability rates and different efficiency of translation [24-26]. Moreover, a previous work revealed that the isoform profile and the abundance of the PrPc in sheep were tissue-specific, showing lower PrPc abundance in lymphoid tissues (three orders of magnitude) than in CNS tissues [27]. However, our results showed higher PrP transcripts in lymphoid tissues suggesting that a postranscriptional regulation of the PrP gene occurs in these tissues. Therefore, the comparison of PrP gene expression among tissues might be a very complex issue. When expression values from nervous tissues were grouped, a tendency towards increased mRNA expression levels of PrP gene and genetic susceptibility to scrapie was observed in CNS tissues. In these tissues, Type 1 animals showed the lowest expression levels and a gradual increase of PrP gene expression was found towards Type 3. Curiously, Type 3 animals showed more PrP mRNA levels than Type 5, however, no statistical differences were found. On the other hand, in spleen and ileum, Type 1 showed the highest expression levels. Inherent artefacts such as PCR inhibitors in spleen samples or the heterogeneous distribution of the immune system in ileum of different animals might have contributed to the high variability found in these samples.
In general, most of the differences found in this study were only marginally significant, and therefore, the results presented here cannot support the existence of a relationship between PrP mRNA levels and risk group. Interestingly, recent studies using a smaller number of samples have revealed PrP genotype-specific differences in PrPc levels in mononuclear cells of peripheral ovine blood [28] and in the amount of PrPSc (but no PrPc) in experimentally infected sheep brain [29]. Therefore, if PrPc synthesis were PrP genotype-dependent, this study would show that this association does not occur at the transcriptional level. However, an association at later stages, i.e. at the postranscriptional regulation level (including mRNA transport out of the nucleus, transcripts stability and regulation at the level of translation) cannot be excluded. Thus, since expression of genes is controlled at several steps, further studies applying different approaches are needed. In addition, the complexity of scrapie pathogenesis might also be influenced by other still unknown genes or strain-specific factors.
Conclusion
The global overview of scrapie pathogenesis is quite complex, but being PrPc the substrate for the conversion of the pathogenic form, PrP mRNA transcripts play an important role, and in this sense, the results on PrP gene expression presented here provide valuable baseline data for future studies. In any case, whatever the mechanism for susceptibility, this study showed that it is not related to the regulation of the PrP gene transcripts. On the other hand, the results on stability data of several HK genes reported in this study could prove very useful in other gene expression studies carried out in these ovine tissues. Future gene expression studies including a larger and more diverse (i.e. different breeds) set of samples would benefit from these data.
Methods
Sample selection
Twenty-two healthy sheep from Latxa breed were selected according to their susceptibility to scrapie [7] and the distribution of genotypes within the Latxa sheep population [30]. Hence, all the highly susceptible genotypes (Type 5) described in Latxa breed were included, along with those other genotypes present in more than 1.5% of the population. In any case, only the most prevalent genotypes were represented by more than one individual. Thus, three animals with ARR/ARR genotype, five ARQ/ARQ, one VRQ/VRQ, one ARH/ARH, five ARR/ARQ, five ARQ/VRQ, one ARQ/ARH and one animal with ARH/VRQ genotype were selected. PrP genotyping was performed by real-time PCR as previously described [30]. The age of the animals ranged from 3 to 12 years old. All the animals were negative to TSE by the PrPSc detection kit Platelia®BSE (Bio-Rad, Hercules, CA, USA) on an obex sample. Animals were sacrificed under controlled conditions and a sample from the same region of cerebrum (neocortex), cerebellum, obex, spleen, terminal ileum and mesenteric lymph node was aseptically taken from each animal. All tissues were frozen immediately at -80°C until RNA extraction was performed.
RNA extraction and cDNA synthesis
Tissue samples were homogenised with a Ribolyzer (Hybaid, Ashford, UK) and total RNA was isolated using the RNeasy Protect Mini kit (Qiagen, Hilden, Germany). Total RNA was treated with DNase I (Ambion, Austin, TX, USA) to avoid genomic DNA amplification and first strand cDNA was synthesised using random hexamers and MultiScribe™ reverse transcriptase (Applied Biosystems, Foster City, CA, USA) according to manufacturer's instructions. In addition, the effectiveness of the DNase treatment was assessed in RT-negative samples. After reverse transcription, the same batch of diluted cDNA was subjected to real-time PCR to amplify six HK genes and the PrP gene.
Primers design and optimisation
Six commonly used HK genes were selected to normalise the expression of the target gene PrP: β-actin (ACTB), tyrosine 3-monooxygenase (YWHAZ), ribosomal protein L19 (RPL19), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), glucose-6-phosphate dehydrogenase (G6PDH) and succinate dehydrogenase (SDHA). Primers were designed to span one intron using Primer Express software (Applied Biosystems, Foster City, CA, USA) (Table 1). Primers for the ACTB, GAPDH, RPL19, G6PDH and PrP genes were designed from ovine sequences obtained from GenBank. For the SDHA and YWHAZ genes, ovine sequences were not available and therefore, multiple sequence alignments of these genes obtained from different animal species (Bos taurus, Mus musculus, Rattus norvegicus and Homo sapiens) were carried out using the program AlingX (Vector NTI 8.0 suite, Informax Inc., North Bethesda, MD, USA) to identify conserved regions for primer design. Four concentrations of primers (50 nM, 100 nM, 200 nM and 300 nM) were evaluated, and formation of primer-dimers was assessed by melting curve analysis. Thus, only those concentrations of primers which showed dimer-free reactions were used for the final analysis.
Real-time RT-PCR
The same number of samples (22) for each tissue were analysed to prevent bias in the results. PCR reactions were set up with the automatic workstation Biomek 2000 (Beckman-Coulter, Fullerton, CA, USA) to minimise pippetting errors. Each sample was analysed in triplicate in a total reaction volume of 10 μl consisting of 10 ng of cDNA, 2xSYBRGreen buffer (Applied Biosystems, Foster City, CA, USA) and the required amount of forward and reverse primers (Table 1). Reactions were run on an ABI PRISM 7000 thermocycler (Applied Biosystems, Foster City, CA, USA) using the following cycling conditions: 95°C for 10 min and 40 cycles at 95°C for 15 s and 60°C for 1 min. For each experiment, a non-template reaction was included as negative control. The specificity of the PCR reactions was confirmed by melting curves analysis of the products as well as by size verification of the amplicons in a conventional agarose gel. In addition, PCR products from the HK genes and PrP gene were cloned into pCR®4-TOPO vector using TOPO TA Cloning® kit (Invitrogen, CA, USA) and submitted to a commercial subcontractor for automatic dye-terminator cycle sequencing. The sequences of SDHA and YWHAZ genes were deposited in GenBank under accession nos. AY970969 and AY970970, respectively.
The threshold cycle values (Ct) were determined at the same fluorescence threshold line for each gene and the Ct value for each sample was obtained by calculating the arithmetic mean of the triplicate values when the standard deviation was lower than 0.16. Ct values were transformed into raw quantity values (Q) according to the following equation, Q = E (Min Ct-Sample Ct) (geNorm user manual, ), where "E" is the efficiency of the real-time PCR for each gene and "Min Ct" is the minimum Ct value for the samples analysed. E values were calculated for each gene from the given slope after running serial dilutions of cDNA and the following formula E= [10(-1/slope)] [31].
Selection of the optimal HK genes and normalisation of PrP gene expression
The method described by Vandesompele et al. [16] was followed to assess the stability of the expression of the HK genes under study using the MS Excel application (geNorm 3.3). Briefly, this application calculates the expression stability measure (M) for the set of HK genes and selects the minimum number of HK genes needed for the normalisation. Thus, genes with the lowest M values have the most stable expression and following the stepwise exclusion of the less stable HK gene M values are re-calculated and the stability series is obtained. Once ranked, the minimum number of HK genes needed was calculated using a cut-off value of 0.15 for Vn/n+1 [16]. The normalisation factor (NF) was then calculated as the geometric mean of their Q values.
Finally, the normalised expression level of the PrP gene (nPrP) was calculated as the ratio between the Q values of PrP gene amplification and the NF calculated for each sample.
Statistical analysis
In order to compare among tissues, analysis of variance of the reference values (NF) was carried out with the GLM procedure of the SAS statistical package version 8 (SAS institute, Cary, NC, USA). Once the results of this model showed no significant differences among tissues, the PrP/NF ratios (nPrP) were transformed according to the formula arc sin √ (nPrPx100-1) as recommended for use of parametric tests on relative data. Then nPrP values along with all available independent variables (sex, age, tissue and genetic susceptibility) and their interactions were submitted to the GLM procedure of SAS statistical package version 8 (SAS institute, Cary, NC, USA). This analysis showed that sex and age had no significant effects and therefore, only tissue and genetic susceptibility and their interactions as independent variables for effects on nPrP were included in the final model. Genetic susceptibility was considered as risk levels [7]. Comparison of means was carried out using a Student's t test with the SAS statistical package version 8 (SAS institute, Cary, NC, USA).
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
DG carried out the design of the study, the experimental work and drafting of the manuscript. RJ performed the statistical analysis and participated in the critical reading of the publication. AH participated in the design and coordination of the study and drafting of the manuscript. All authors read and approved the final manuscript.
Acknowledgements
Financial support was provided by the Basque Government (Projects PI-00-20 and EC2001-3). D. Garcia-Crespo was the recipient of a Predoctoral Fellowship from the Dpto. de Educación, Universidades e Investigación of the Basque Government. We thank A. L. García-Pérez and J. Barandika for their collaboration in the selection of the animals and the TSE laboratory staff for their contribution in rapid test analyses. We thank N. Gomez for helpful comments on the manuscript.
==== Refs
Hunter N Scrapie Mol Biotechnol 1998 9 225 234 9718582
Prusiner SB Prions Proc Natl Acad Sci U S A 1998 95 13363 13383 9811807 10.1073/pnas.95.23.13363
Brandner S Raeber AJ Sailer A Blattler T Fischer M Weissmann C Aguzzi A Normal host prion protein (PrPc) is required for scrapie spread within the central nervous system Proc Natl Acad Sci U S A 1996 93 13148 13151 8917559 10.1073/pnas.93.23.13148
Blattler T Brandner S Raeber AJ Klein MA Voigtlander T Weissmann C Aguzzi A PrP-expressing tissue required for transfer of scrapie infectivity from spleen to brain Nature 1997 389 69 73 9288968 10.1038/37981
Hunter N Foster JD Dickinson AG Hope J Linkage of the gene for the scrapie-associated fibril protein (PrP) to the Sip gene in Cheviot sheep Vet Rec 1989 124 364 366 2566212
Belt PB Muileman IH Schreuder BE Bos-de RJ Gielkens AL Smits MA Identification of five allelic variants of the sheep PrP gene and their association with natural scrapie J Gen Virol 1995 76 509 517 7897344
UKNational Scrapie Plan risk groups 2003
Bossers A de Vries R Smits MA Susceptibility of sheep for scrapie as assessed by in vitro conversion of nine naturally occurring variants of PrP J Virol 2000 74 1407 1414 10627551 10.1128/JVI.74.3.1407-1414.2000
Sabuncu E Petit S Le Dur A Lan LT Vilotte JL Laude H Vilette D PrP polymorphisms tightly control sheep prion replication in cultured cells J Virol 2003 77 2696 2700 12552009 10.1128/JVI.77.4.2696-2700.2003
Maignien T Lasmezas C Beringue V Dormont D Deslys JP Pathogenesis of the oral route of infection of mice with scrapie and bovine spongiform encephalopathy agents J Gen Virol 1999 80 3035 3042 10580067
van Keulen LJ Vromans ME van Zijderveld FG Early and late pathogenesis of natural scrapie infection in sheep APMIS 2002 110 23 32 12064252 10.1034/j.1600-0463.2002.100104.x
Tichopad A Pfaffl MW Didier A Tissue-specific expression pattern of bovine prion gene: quantification using real-time RT-PCR Mol Cell Probes 2003 17 5 10 12628587 10.1016/S0890-8508(02)00114-7
Ning ZY Zhao DM Yang JM Cui YL Meng LP Wu CD Liu HX Quantification of prion gene expression in brain and peripheral organs of golden hamster by real-time RT-PCR Anim Biotechnol 2005 16 55 65 15926263
Heid CA Stevens J Livak KJ Williams PM Real time quantitative PCR Genome Res 1996 6 986 994 8908518
Bustin SA Quantification of mRNA using real-time reverse transcription PCR (RT-PCR): trends and problems J Mol Endocrinol 2002 29 23 39 12200227 10.1677/jme.0.0290023
Vandesompele J De Preter K Pattyn F Poppe B Van Roy N De Paepe A Speleman F Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes Genome Biol 2002 3 research0034.1 0034.12 12184808 10.1186/gb-2002-3-7-research0034
Laud K Hornez L Gourdou I Belair L Arnold A Peyrat JP Djiane J Expression of BRCA1 gene in ewe mammary epithelial cells during pregnancy: regulation by growth hormone and steroid hormones Eur J Endocrinol 2001 145 763 770 11720902 10.1530/eje.0.1450763
Grubor B Gallup JM Ramirez-Romero R Bailey TB Crouch EC Brogden KA Ackermann MR Surfactant protein D expression in normal and pneumonic ovine lung Vet Immunol Immunopathol 2004 101 235 242 15350753 10.1016/j.vetimm.2004.05.004
Hein WR Barber T Cole SA Morrison L Pernthaner A Long-term collection and characterization of afferent lymph from the ovine small intestine J Immunol Methods 2004 293 153 168 15541285 10.1016/j.jim.2004.07.008
Thellin O Zorzi W Lakaye B De Borman B Coumans B Hennen G Grisar T Igout A Heinen E Housekeeping genes as internal standards: use and limits J Biotechnol 1999 75 291 295 10617337 10.1016/S0168-1656(99)00163-7
Warrington JA Nair A Mahadevappa M Tsyganskaya M Comparison of human adult and fetal expression and identification of 535 housekeeping/maintenance genes Physiol Genomics 2000 2 143 147 11015593
Tichopad A Didier A Pfaffl MW Inhibition of real-time RT-PCR quantification due to tissue-specific contaminants Mol Cell Probes 2004 18 45 50 15036369 10.1016/j.mcp.2003.09.001
Yperman J De Visscher G Holvoet P Flameng W Beta-actin cannot be used as a control for gene expression in ovine interstitial cells derived from heart valves J Heart Valve Dis 2004 13 848 853 15473489
Hunter N Manson JC Charleson FC Hope J Comparison of expression patterns of PrP mRNA in the developing sheep and mouse Ann N Y Acad Sci 1994 724 353 354 8030958
Horiuchi M Yamazaki N Ikeda T Ishiguro N Shinagawa M A cellular form of prion protein (PrPC) exists in many non-neuronal tissues of sheep J Gen Virol 1995 76 2583 2587 7595362
Goldmann W O'Neill G Cheung F Charleson F Ford P Hunter N PrP (prion) gene expression in sheep may be modulated by alternative polyadenylation of its messenger RNA J Gen Virol 1999 80 2275 2283 10466827
Moudjou M Frobert Y Grassi J La Bonnardiere C Cellular prion protein status in sheep: tissue-specific biochemical signatures J Gen Virol 2001 82 2017 2024 11458009
Halliday S Houston F Hunter N Expression of PrPC on cellular components of sheep blood J Gen Virol 2005 86 1571 1579 15831971 10.1099/vir.0.80561-0
McCutcheon S Hunter N Houston F Use of a new immunoassay to measure PrP(Sc) levels in scrapie-infected sheep brains reveals PrP genotype-specific differences J Immunol Methods 2005 298 119 128 15847802 10.1016/j.jim.2005.01.012
Garcia-Crespo D Oporto B Gomez N Nagore D Benedicto L Juste RA Hurtado A PrP polymorphisms in Basque sheep breeds determined by PCR-restriction fragment length polymorphism and real-time PCR Vet Rec 2004 154 717 722 15214515
Rasmussen R Meuer S, Wittwer C and Nakagawara K Quantification on the LightCycler instrument Rapid Cycle Real-time PCR, Methods and Applications 2001 Heidelberg, Springer-Verlag Press 21 34
|
16188044
|
PMC1262732
|
CC BY
|
2021-01-04 16:29:46
|
no
|
BMC Vet Res. 2005 Sep 28; 1:3
|
utf-8
|
BMC Vet Res
| 2,005 |
10.1186/1746-6148-1-3
|
oa_comm
|
==== Front
J CarcinogJournal of Carcinogenesis1477-3163BioMed Central London 1477-3163-4-171620216510.1186/1477-3163-4-17ResearchhTERT protein expression is independent of clinicopathological parameters and c-Myc protein expression in human breast cancer Elkak AE [email protected] G [email protected] M [email protected] B [email protected] JRS [email protected] RF [email protected] K [email protected] The Breast Unit, St George's Hospital, Blackshaw Road, London, SW17 0QT, UK2 Department of Histopathology, King's Mill Hospital, Mansfield, UK3 Department of Surgery, King's Mill Hospital, Mansfield, UK4 Institute of Cancer Genetics and Pharmacogenomics, Brunel University, Uxbridge, Middlesex. UB8 3PH, UK2005 4 10 2005 4 17 17 28 7 2005 4 10 2005 Copyright © 2005 Elkak et al; licensee BioMed Central Ltd.2005Elkak et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Telomerase is a ribonucleoprotein enzyme that synthesises telomeres after cell division and maintains chromosomal length and stability thus leading to cellular immortalisation. The hTERT (human telomerase reverse transcriptase) subunit seems to be the rate-limiting determinant of telomerase and knowledge of factors controlling hTERT transcription may be useful in therapeutic strategies. The hTERT promoter contains binding sites for c-Myc and there is some experimental and in vitro evidence that c-Myc may increase hTERT expression. We previously reported no correlation between c-Myc mRNA expression and hTERT mRNA or telomerase activity in human breast cancer. This study aims to examine the correlation between hTERT expression as determined by immunohistochemistry and c-Myc expression, lymph node status, and tumour size and grade in human breast cancer.
Materials and methods
The immunohistochemical expression of hTERT and c-Myc was investigated in 38 malignant breast tumours. The expression of hTERT was then correlated with the lymph node status, c-Myc expression and other clinicopathological parameters of the tumours.
Results
hTERT expression was positive in 27 (71%) of the 38 tumours. 15 (79%) of 19 node positive tumours were hTERT positive compared with 11 (63%) of 19 node negative tumours. The expression was higher in node positive tumours but this failed to reach statistical significance (p = 0.388). There was no significant association with tumour size, tumour grade or c-Myc expression. However, hTERT expression correlated positively with patients' age (correlation coefficient = 0.415, p = 0.0097).
Conclusion
hTERT protein expression is independent of lymph node status, tumour size and grade and c-Myc protein expression in human breast cancer
==== Body
Introduction
Telomeres are highly specialised structures at the end of each chromosome which contain tandem repeat DNA sequences? In humans, this sequence is TTAGGG [1]. According to the telomere/telomerase hypothesis, the telomeric ends of chromosomes of normal somatic cells replicate inefficiently and progressively shorten at each cell division until cumulative loss impairs vital cellular functions and the cells exit the cell cycle and undergo apoptosis [2].
Telomerase is a ribonucleoprotein enzyme that contains an RNA template complementary to the (TTAGGG) n repeats and thus synthesises telomeres after cell division and maintains chromosomal length and stability leading to immortalisation [3-6]. Accordingly, telomerase activity has been investigated and detected in a wide range of human malignancies, germline cells and immortal cells but not in normal somatic cells [3-8]. Human telomerase consists of an RNA subunit – human telomerase RNA (hTR) [9], a protein component (human telomerase associated protein 1 – hTEP1) [10] and the catalytic subunit hTERT (human telomerase reverse transcriptase) [11-13]. Of these subunits telomerase activity requires the presence of hTR, which is the RNA template for the telomeric repeat, and hTERT, which is the reverse transcriptase. The gene coding for hTERT has been cloned and mapped to 5p15.33 [14]. hTERT is a 127-kDa protein whose function closely correlates with telomerase activity [10,15,16].
Investigation of the mechanisms of hTERT control is important in elucidating the pathways that may be amenable to therapeutic manipulation and one such pathway involves the transcription factor Myc.
An increased level of c-Myc occurs frequently in a wide range of tumours [17-23] due to de-regulated expression of myc through gene amplification, retroviral insertion or chromosomal translocation. Sequence analysis of the hTERT gene promoter has shown the presence of at least 2 and perhaps as many as 29 E boxes [24], potential binding sites for the Myc oncoprotein, and the possibility of a regulatory role for Myc has been explored in a number of studies. It has been found that purified Myc interacts with the E box sequences and that cotransfection of Myc induces activity in the isolated hTERT promoter [25]. It has been shown that retroviral expression of c-myc increases the amount of hTERT mRNA in human mammary epithelial cells and fibroblasts and telomerase activity could thereafter be detected [26]. It has also been reported that expression of c-Myc leads to increased hTERT expression and telomerase activity in human B cells [27]. Moreover, since this does not require protein synthesis this appears to be due to a direct effect of Myc on the hTERT promoter and not secondary to an increase in cellular proliferation by Myc [27]. In addition, the introduction of Myc anti-sense RNA has been shown to lead to a reduction in hTERT promoter activity [25]. Latil et al [28] demonstrated a relationship between hTERT and c-Myc expressions in prostate cancer whilst other investigators [29] have found that Mad1, another member of Myc pathway, is a suppressor of hTERT at the level of transcription.
Using RT-PCR, we found that hTERT mRNA expression was significantly higher in breast cancer tissues compared with non-cancerous breast tissues [30]. However, we did not observe any relationship between hTERT expression and tumour stage, patients' age or c-Myc expression. Due to the potential limitations of mRNA studies which we highlighted in our previous report, we planned to determine the protein expression of both hTERT and c-Myc using immunohistochemistry (IHC) and investigate any potential associations.
The aim of this study was to determine the (IHC) expression of hTERT in human breast cancer and to examine the association between hTERT expression and clinicopathological parameters of the tumours (size, grade, and nodal status) and c-Myc expression.
Methods
Institutional guidelines including ethical approval were followed. Patients were treated with wide local excision or mastectomy and axillary node dissection. Patients with estrogen and /or progesterone positive tumours received tamoxifen. Radiotherapy was administered to all patients who had breast conserving surgery and chemotherapy to patients with lymph node positive or high grade tumours. Using IHC, the expressions of hTERT and c-Myc were determined in 38 breast cancer specimens (19 specimens were lymph node positive tumours and 19 were lymph node negative tumours the specimens).
Immunohistochemistry
Staining
Paraffin sections were dewaxed in xylene (two changes each for 5 minutes). Endogenous peroxidase was blocked in 480 ml of Methanol and 6 ml of H2O2 for 15 minutes and rinsed in running tap water for 5 minutes. Heat mediated antigen retrival was performed as follows: 1 litre of pH6 Citrate buffer was placed in the microwave into the pressure cooker, then heated on full power for 10 minutes.
Sections were placed into the boiling buffer. Pressure cooker was heated on full power until the pressure is attained and then cooked for a further 2 minutes. Sections were then rinsed well in running tap water and placed in Tris Buffer. Primary antibody preparations were applied and incubated for 45 minutes then washed in TRIS Buffer (pH 7.6). The antibodies used are: NCLL-hTERT (Clone 44F12, 1:50 dilution, Novocastra Laboratories Ltd, Newcastle upon Tyne, UK) and NCL-cMYC (Clone 9E11, 1:200 dilution, Novocastra Laboratories Ltd, Newcastle-upon-Tyne, UK). 5% goat serum was used for dilution. Super enhancer was then applied for 20 minutes. Sections were washed in TRIS Buffer (pH 7.6). Poly-HRP was then applied for 30 minutes. Sections were washed in TRIS Buffer (pH 7.6). DAB solution was then applied for 10 minutes. Slides were washed in distilled water, bleached to de-activate the DAB and rinsed in running tap water. Sections were then counterstained with Harris Haematoxylin.
Evaluation of immunohistochemical staining
Two observers assessed the sections using the following criteria:
hTERT
Scores were assigned as follows: 2, strong staining throughout nucleus (Fig 1); 1, moderate staining of nucleus or dotted staining of nucleolus (Fig 2); 0, no staining.
Figure 1 Invasive ductal carcinoma – hTERT positive: 2 (strong nuclear staining).
Figure 2 Invasive ductal carcinoma – hTERT positive: 1 (moderate nuclear staining of tumour cells and strong staining of surrounding lymphocytes.
c-Myc
Cytoplasmic staining intensity was graded as: no staining (0), weak (1), moderate (2) (Fig 3), or strong (3) (Fig 4). The percentage of tumour cells with c-Myc staining was scored as follows: 1, <5%; 2, 5–20%; 3, 21–50%; 4, >50%. Then the multiplication values were grouped into four scores as 0, (multiplication values 0, 1); 1, (multiplication values 2, 3); 2, (multiplication values 4, 6); or 3, (multiplication values 8, 9, 12).
Figure 3 Invasive ductal carcinoma – c-Myc positive: 2 (moderate cytoplasmic staining).
Figure 4 Invasive ductal carcinoma – c-Myc positive: 3 (strong cytoplasmic staining).
Statistical analysis
Chi Square test was used to study the relationship between hTERT expression in lymph node positive and lymph node negative tumours. The expression of hTERT in tumours was then correlated with c-Myc expression, clinicopathological parameters of the tumours (size and grade) and patients' age. A p-value of < 0.05 was considered statistically significant.
Results
hTERT expression was positive in 27 (71%) of 38 tumours. 15 (79%) of the 19 node positive tumours were hTERT positive compared with 11 (63%) of the node negative tumours. Although hTERT expression was higher in node positive tumours (median score 1.0 Vs 0), however this failed to reach statistical significance (p = 0.388). There was no significant correlation with tumour size, tumour grade, hormone receptor status or c-Myc expression. Interestingly, hTERT expression correlated positively with patients' age (correlation coefficient = 0.415, p = 0.0097). Table 1 demonstrates hTERT and c-Myc expression and the clinicopathological characteristics of the tumours studied.
Table 1 hTERT and c-Myc expression and the clinicopathological characteristics of the tumours
No hTERT Int % Multip Final Hist Age Size Grade LN ER
1 1 3 4 12 3 D 77 23 2 P P
2 1 1 1 1 0 D 60 13 3 P P
3 2 2 4 8 3 D 75 28 3 P P
4 2 3 4 12 3 D 68 35 3 P P
5 1 3 4 12 3 D 63 12 2 P P
6 1 2 4 8 3 D 74 14 3 P P
7 1 3 4 12 3 D 48 28 2 P P
8 2 2 4 8 3 D 72 50 3 P N
9 2 3 4 12 3 D 58 30 2 P P
10 0 3 4 12 3 L 50 35 2 P P
11 1 3 3 9 3 D 38 30 3 P N
12 1 2 4 8 3 D 56 20 3 P N
13 1 1 4 4 2 D 54 50 2 P P
14 0 3 4 12 3 L 43 20 1 P P
15 0 3 4 12 3 L 43 20 1 P P
16 0 2 4 8 3 DL 48 23 2 P P
17 1 3 4 12 3 DL 82 35 2 P P
18 1 3 4 12 3 PL 73 50 3 P P
19 1 1 4 4 2 D 50 18 2 P P
20 1 3 4 12 3 DL 61 6 2 N P
21 1 3 4 12 3 D 59 15 2 N P
22 1 3 4 12 3 D 63 26 3 N P
23 0 2 4 8 3 P 66 13 2 N P
24 1 3 4 12 3 D 61 15 3 N P
25 1 3 4 12 3 D 65 6 1 N P
26 2 3 4 12 3 D 51 35 3 N P
27 1 3 4 12 3 D 50 12 2 N P
28 2 2 4 8 3 TL 70 12 1 N P
29 1 3 4 12 3 TL 62 19 2 N P
30 0 3 4 12 3 T 43 8 1 N P
31 2 2 4 8 3 M 54 14 1 N P
32 1 3 4 12 3 D 60 13 1 N P
33 1 3 4 12 3 A 56 11 2 N N
34 0 3 4 12 3 D 57 10 2 N P
35 0 3 4 12 3 D 53 18 2 N P
36 0 2 4 8 3 D 55 10 2 N P
37 0 2 4 8 3 D 52 20 3 N P
38 0 1 2 2 1 D 64 12 3 N N
Discussion
Our observation that hTERT protein is expressed in most breast tumours is expected and consistent with our previous reports using mRNA and enzyme measurements [30-32]. The lack of is correlation between hTERT protein expression and tumour size, grade or nodal status is also consistent with our previous study using mRNA and RT-PCR technology [30]. However we previously reported that telomerase activity correlated with these clinicopathological parameters [31,32]. Although hTERT expression is associated with malignancy, it does not seem to correlate with tumour stage. This is probably a true observation as it has been demonstrated using both RT- PCR and immunohistochemistry and could be explained on the basis of post-transcriptional modification.
Bieche et al [33] reported a positive correlation between hTERT and c-Myc gene expression. Furthermore, other investigators [34-36] demonstrated that hTERT gene is a direct target of c-Myc. Although the hTERT promoter contains E-boxes, consistent with the findings of the present study, we previously observed no correlation between c-Myc mRNA levels and telomerase activity [37] and no association between hTERT and c-Myc at the mRNA level [30]. The control of hTERT is undoubtedly a complex one and it is likely that a number of other transcription factors influence its expression than c-Myc. These might act together with c-Myc, as has been shown for Sp1 [38] or independently. In this respect, it has been shown that transfer of a normal chromosome 3 into human breast carcinoma cells results in abolition of hTERT transcripts without any change in c-Myc levels [39]. Furthermore, it is known that another member of the Myc family, Mad1 forms a complex with Max and acts as a transcriptional repressor at the same binding sites as Myc-Max. It has been shown that, the proportion of Mad1 binding to the hTERT promoter rises and that of Myc falls, during the differentiation of HL60 cells [37,40]. This is associated with reduced acetylation of the hTERT promoter and measurement of the Mad/Myc ratio is likely to be important in establishing the overall level of transcriptional activation of hTERT.
Abbreviations
Int: c-Myc intensity
%: c-Myc percentage positive
Multip: c-Myc multiplication
Final: Final c-Myc score
Hist: Histology
D: ductal
L: Lobular
DL: Ductal and Lobular
P: Papillary
M: Mucoid
A: Apocrine
T: Tubular
LN: Lymph node status, P: positive, N: negative
ER: Oestrogen receptor status, P: positive, N: negative
==== Refs
Morin GB The human telomere terminal transferase enzyme is aribonucleoprotein that synthesises TTAGGG repeats Cell 1989 59 521 9 2805070 10.1016/0092-8674(89)90035-4
Harley CB Telomere loss: Mitotic clock or genetic time bomb? Mutation Res 1991 256 271 82 1722017 10.1016/0921-8734(91)90018-7
Kim NW Piatyszek MA Prowse KR Specific association of human telomerase activity with immortal cells and cancer Science 1994 266 2011 15 7605428
Rhyu MS Telomeres, Telomeres and immortality J Na Cancer Inst 1995 87 884 94
Chadeneau C Hay K Hirte HW Telomerase activity associated with acquisition of malignancy in human colorectal cancer Cancer Res 1995 55 2533 6 7780964
Blackburn EH Structure and formation of telomeres Nature 1991 350 569 73 1708110 10.1038/350569a0
Hiyama E Gollahon L Kataoka TE Telomerase activity in human breast tumours J Natl Cancer Inst 1996 88 116 22 8537972
Hiyama E Yokoyama T Tatsumoto N Telomerase activity in gastric cancer Cancer Res 1995 55 3258 62 7614459
Feng J Funk WD Wang SS The RNA component of human telomerase Science 1995 269 1236 41 7544491
Harrington L McPhail T Mar V A mammalian telomerase-associated protein Science 1997 275 973 7 9020079 10.1126/science.275.5302.973
Yang H Kyo S Takatura M Autocrine transformation growth factor β suppresses telomerase activity and transcription of human telomerase reverse transcriptase in human cancer cells Cell Growth Differ 2001 12 119 27 11243466
Nakamura TM Morin GB Chapman KB Telomerase catalytic subunit homologs from fission yeast and human Science 1997 277 955 959 9252327 10.1126/science.277.5328.955
Weinrich SL Pruzan R Ma L Reconstitution of human telomerase with the template RNA component hTR and the catalytic protein subunit hTERT Nat Genet 1997 17 498 502 9398860 10.1038/ng1297-498
Kirkpatrick KL Mokbel K The significance of human telomerase reverse transcriptase (hTERT) in cancer Eur J Surg Oncol 2001 127 754 60 11735173 10.1053/ejso.2001.1151
Nakayama J Tahara H Tahara E Telomerase activation by hTRT in human normal fibroblasts and hepatocellular carcinomas Nat Genet 1998 18 65 8 9425903 10.1038/ng0198-65
Meyerson M Counter CM Eaton EN hEST2, the putative human telomerase catalytic subunit gene, is up-regulated in tumor cells and during immortalization Cell 1997 90 785 95 9288757 10.1016/S0092-8674(00)80538-3
Stephenson BM Hall NR Mueller RF Cuschieri A, Giles GR, Moosa AR DNA technology, genetics, cancer and surgery Essential surgical practice 1995 Oxford: Butterworth Heinemann 138 176
Rothermund K Rogulski K Fernandes E C-Myc-independent restoration of multiple phenotypes by two C-Myc target genes with overlapping functions Cancer Res 2005 15 2097 107 15781619 10.1158/0008-5472.CAN-04-2928
Oster SK Ho CS Soucie EL The myc oncogene: MarvelouslY Complex Adv Cancer Res 2002 84 81 154 11885563
Fernandez PC Frank SR Wang L Genomic targets of the human c-Myc protein Genes Dev 2003 17 1115 29 12695333 10.1101/gad.1067003
O'Connell BC Cheung AF Simkevich CP A large scale genetic analysis of c-Myc-regulated gene expression patterns J Biol Chem 2003 278 12563 73 12529326 10.1074/jbc.M210462200
Levens DL Reconstructing MYC Genes Dev 2003 17 1071 7 12730130 10.1101/gad.1095203
Nesbit CE Tersak JM Prochowink MYC oncogenes and human neoplastic disease Oncogene 1999 18 3004 16 10378696 10.1038/sj.onc.1202746
Cong YS Wen J Bacchetti S The human telomerase catalytic subunit hTERT: organization of the gene and characterization of the promoter Hum Mol Genet 1999 8 137 142 9887342 10.1093/hmg/8.1.137
Cerni C Telomeres, telomerase, and myc. An update. [Review] [139 refs] Mutat Res 2000 462 31 47 10648922 10.1016/S1383-5742(99)00091-5
Wang J Xie LY Allan S Beach D Hannon GJ Myc activates telomerase Genes Dev 1998 12 1769 1774 9637678
Wu KJ Grandori C Amacker M Simon-Vermot N Polack A Lingner J Dalla-Favera R Direct activation of TERT transcription by c-MYC Nat Genet 1999 21 220 224 9988278 10.1038/6010
Latil A Vidaud D Valeri A hTERT expression correlates with MYC overexpression in human prostate cancer Int J Cancer 2000 89 172 6 10754496 10.1002/(SICI)1097-0215(20000320)89:2<172::AID-IJC12>3.3.CO;2-#
Gunes C Lichtsteiner S Vasserot AP Englert C Expression of the hTERT gene is regulated at the level of transcriptional initiation and repressed by Mad1 Cancer Res 2000 60 2116 21 10786671
Kirkpatrick KL Ogunkolade W Elkak AE hTERT expression in human breast cancer and non-cancerous breast tissue: correlation with tumour stage and c-Myc expression Breast Cancer Res Treat 2003 77 277 84 12602927 10.1023/A:1021849217054
Mokbel K Parris CN Ghilchik M The association between telomerase, histopathological parameters, and KI-67 expression in breast cancer Am J Surg 1999 178 69 72 10456708 10.1016/S0002-9610(99)00128-2
Mokbel KM Parris CN Ghilchik M Telomerase and lymphovascular invasion in breast cancer Eur J Surg Oncol 2000 126 30 3 10718176 10.1053/ejso.1999.0736
Bieche I Nogues C Paradis V Quantitation of hTERT gene expression in sporadic breast tumors with a real-time reverse transcription-polymerase chain reaction assay Clin Cancer Res 2000 6 452 9 10690523
Greenberg RA O'Hagan RC Deng H telomerase reverse transcriptase gene is a direct target of c-Myc but is not functionally equivalent in cellular transformation Oncogene 1999 18 1219 26 10022128 10.1038/sj.onc.1202669
Cong YS Wen J Bacchetti S The human telomerase catalytic subunit hTERT: organization of the gene and characterization of the promoter Hum Mol Genet 1999 8 137 42 9887342 10.1093/hmg/8.1.137
Wang J Xie LY Allan S Myc activates telomerase Genes Dev 1998 12 1769 74 9637678
Kirkpatrick KL Newbold RF Mokbel K There is no correlation between c-Myc mRNA expression and telomerase activity in human breast cancer Int Semin Surg Oncol 2004 1 2 15285810 10.1186/1477-7800-1-2
Kyo S Takakura M Taira T Kanaya T Itoh H Yutsudo M Ariga H Inoue M Sp1 cooperates with c-Myc to activate transcription of the human telomerase reverse transcriptase gene (hTERT) Nucleic Acids Res 2000 28 669 677 10637317 10.1093/nar/28.3.669
Ducrest AL Amacker M Mathieu YD Regulation of human telomerase activity: repression by normal chromosome 3 abolishes nuclear telomerase reverse transcriptase transcripts but does not affect c-Myc activity Cancer Res 2001 61 7594 602 11606399
Xu D Popov N Hou M Switch from Myc/Max to Mad1/Max binding and decrease in histone acetylation at telomerase reverse transcriptase promoter during differentiation of HL60 cells Proc Natl Acad Sci USA 2001 98 3826 31 11274400 10.1073/pnas.071043198
|
16202165
|
PMC1262733
|
CC BY
|
2021-01-04 16:39:20
|
no
|
J Carcinog. 2005 Oct 4; 4:17
|
utf-8
|
J Carcinog
| 2,005 |
10.1186/1477-3163-4-17
|
oa_comm
|
==== Front
J CarcinogJournal of Carcinogenesis1477-3163BioMed Central London 1477-3163-4-181620970810.1186/1477-3163-4-18ResearchTelomerase expression is sufficient for chromosomal integrity in cells lacking p53 dependent G1 checkpoint function Simpson Dennis A [email protected] Elizabeth [email protected] Timothy P [email protected] William K [email protected] Department of Pathology and Laboratory Medicine, Lineberger Comprehensive Cancer Center, and Center for Environmental Health and Susceptibility, University of North Carolina at Chapel Hill, CB 7295, Chapel Hill, NC 27599, USA2005 6 10 2005 4 18 18 17 3 2005 6 10 2005 Copyright © 2005 Simpson et al; licensee BioMed Central Ltd.2005Simpson et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Secondary cultures of human fibroblasts display a finite lifespan ending at senescence. Loss of p53 function by mutation or viral oncogene expression bypasses senescence, allowing cell division to continue for an additional 10 – 20 doublings. During this time chromosomal aberrations seen in mitotic cells increase while DNA damage and decatenation checkpoint functions in G2 cells decrease.
Methods
To explore this complex interplay between chromosomal instability and checkpoint dysfunction, human fibroblast lines were derived that expressed HPV16E6 oncoprotein or dominant-negative alleles of p53 (A143V and H179Q) with or without the catalytic subunit of telomerase.
Results
Cells with normal p53 function displayed 86 – 93% G1 arrest after exposure to 1.5 Gy ionizing radiation (IR). Expression of HPV16E6 or p53-H179Q severely attenuated G1 checkpoint function (3 – 20% arrest) while p53-A143V expression induced intermediate attenuation (55 – 57% arrest) irrespective of telomerase expression. All cell lines, regardless of telomerase expression or p53 status, exhibited a normal DNA damage G2 checkpoint response following exposure to 1.5 Gy IR prior to the senescence checkpoint. As telomerase-negative cells bypassed senescence, the frequencies of chromosomal aberrations increased generally congruent with attenuation of G2 checkpoint function. Telomerase expression allowed cells with defective p53 function to grow >175 doublings without chromosomal aberrations or attenuation of G2 checkpoint function.
Conclusion
Thus, chromosomal instability in cells with defective p53 function appears to depend upon telomere erosion not loss of the DNA damage induced G1 checkpoint.
==== Body
Background
Normal diploid fibroblasts proliferate in secondary cultures for a finite number of population doublings until a growth arrest known as replicative senescence, or M1, is reached [1]. This limitation in lifespan is believed to be due to the continuous shortening of the telomeres with each cell division [2]. Recent evidence has suggested that an alteration in the structure of one or more telomeres may, in fact, be what triggers cells to enter replicative senescence, a permanent p53-dependent G1 arrest [3,4]. Regardless of the exact trigger of senescence, inactivation of p53 allows cells to bypass senescence and continue to divide until a second growth restriction termed crisis, or M2, is reached [5]. Cells in crisis contain numerous structural and numerical chromosomal abnormalities which may be due to cycles of chromosome fusion (dicentric chromosomes) and subsequent resolution of the fusion (chromosome break) during mitosis [2]. A previous study has demonstrated that during the phase of extended proliferation after bypass of M1, telomeres in p53-defective, telomerase-negative cells can erode to the point where little or no telomeric repeat DNA can be detected [6]. Chromosomes without telomeres appear to be substrates for DNA repair pathways resulting in telomere associations and formation of dicentric and ring chromosomes. The resolution of these unstable structures is believed to result in the other structural and numerical abnormalities in chromosomes observed in cells in crisis (i.e., breaks, exchanges, aneuploidy, polyploidy).
Prevention of telomere erosion by ectopic expression of the catalytic subunit of human telomerase (hTERT) has been shown to prevent crisis in cells expressing SV40 large T antigen or HPV16E6 oncoprotein [7-10]. Normal diploid human fibroblasts expressing hTERT have been reported to maintain a normal karyotype and preserve cell cycle checkpoint function for at least 200 population doublings [11,12], although others have suggested that otherwise normal telomerase-expressing human fibroblasts do display alterations in expression of tumor suppressor genes, growth characteristics, and transient genetic instability [13-16] These studies have failed to directly address the question as to whether cells can maintain a stable genome in the absence of a functional DNA damage induced G1 checkpoint. Here we report that in the absence of telomere erosion cells defective for p53 signaling can maintain stable genomes for >175 population doublings. This study found that normal diploid cells expressing hTERT maintain a normal karyotype for at least 100 PD's but eventually did become numerically abnormal. We also report that two independent p53-defective lines which emerged from crisis by reactivation of telomerase displayed remarkably stable karyotypes.
Materials and methods
Plasmids and viruses
All cloning steps were carried out according to standard methods [17]. Plasmids were maintained in the DH5α strain of E. coli. Replication-defective retroviruses used in this study and helper plasmids for packaging are shown in Figure 8. The hTERT retroviral expression vector, pDSWK-8, was created by cloning the hTERT cDNA from pBABE/Hyg-hTERT (Dr. Robert A. Weinberg, Whitehead Institute for Biomedical Research) into the EcoRI and HpaI sites of the pHIT-2 retroviral backbone (Dr. John Olsen, University of North Carolina). cDNA's encoding an alanine to valine substitution at amino acid 143 or a histidine to glutamine substitution at amino acid 179 in p53 (p53-A143V and p53-H179Q respectively) were provided by Drs. David Wynford-Thomas (University of Aberdeen) and Dr. Howard Liber (Massachusetts General Hospital), respectively. Retroviral expression vectors containing these dominant-negative forms of p53 were constructed by cloning the cDNA into the EcoRI site of the pLXIN (Clonetech) retroviral expression vector. The pLXSN-E6 retroviral expression vector containing the HPV16 E6 oncoprotein DNA was a gift from Dr. Denise Galloway (Fred Hutchinson Cancer Center). Vesicular stomatitis virus glycoprotein G-pseudotyped, replication-defective retroviruses were produced as previously described following transient transfection of viral vector and helper plasmids into HEK 293T cells [18-20]. Transfections of plasmids for virus production were done using Superfect™ or Polyfect™ (Qiagen) according to the manufacturer's protocol.
Figure 8 Expression constructs used in this study. CMV/ie Pro/Enh, Cytomegalovirus immediate early promoter/enhancer sequence; VSV-G, vesicular stomatitis virus glycoprotein G; SV40 Pro, simian virus 40 promoter/origin sequence; IRES, internal ribosome entry sequence; MuLV LTR, murine leukemia virus long terminal repeat; hTERT, human telomerase catalytic subunit. A) Plasmid pCI VSV-G expressing VSV-G used to pseudotype replication-defective retrovirus particles. B) Helper plasmid pCI GPZ for packaging replication-defective retrovirus particles. C) pDSWK-8 plasmid vector used to package telomerase cDNA. D) pLXSN-E6 plasmid vector for packaging HPV16 E6. E) pLXIN+p53-A143V and pLXIN+p53-H179Q plasmid vectors for packaging dominant-negative p53 alleles.
Cell Culture
A normal human fibroblast strain designated NHF1 was derived from neonatal foreskin as previously described [21]. All cell culture, including retroviral production, was performed in a humidified, water-jacketed incubator at 37°C with a 5 % CO2 atmosphere. NHF1 cells and all cell lines derived from the parental NHF1 secondary culture were maintained in MEM (Gibco Invitrogen Corp.) supplemented with 10 % defined fetal bovine serum (Hyclone), 2 mM L-glutamine (Gibco Invitrogen Corp.), and 100 μM non-essential amino acids (Gibco Invitrogen Corp.). HEK 293T cells were maintained in DMEM-H (Gibco Invitrogen Corp.) supplemented with 10 % defined fetal bovine serum (Hyclone), 2 mM L-glutamine (Gibco Invitrogen Corp.), 100 μM non-essential amino acids (Gibco Invitrogen Corp.), and 20 mM HEPES pH 7.3 (Sigma Chemical Co.). Transductions were carried out according to standard methods as described previously [22]. Cells at passages 5 or 6 were simultaneously transduced with both the hTERT-expressing virus and one of the viruses disrupting p53 function and/or empty vectors to derive the cell lines listed in Table 1. At the time of transduction, the NHF1 cells were estimated to have undergone 15 – 20 population doublings in vitro. Transductants were selected by 2 weeks growth in media containing 300 ng/ml puromycin (Sigma Chemical Co.) plus 200 μg/ml of active G418 (Gibco Invitrogen Corp.) and, following this initial selection, lines were maintained without antibiotics. Cells were seeded each passage at a density of 5300 – 5500 cells per cm2. The population doubling level (PDL) of the culture was defined as the sum of the population doublings (PD) of each passage. The PD of each passage was determined using the following equation:
Table 1 Status of p53 in Cell Lines
Cell Line hTERT p53 Protein
F1-hTERT+LXIN + WT1
F1-hTERT+p53-A143V + WT/DN2
F1-hTERT+p53-H179Q + WT/DN
F1-hTERT+E6 + -
F1-HIT+LXIN - WT
F1-HIT+p53-A143V - WT/DN
F1-HIT+p53-H179Q - WT/DN
F1-HIT+E6 - -
1. WT: Wild Type
2. DN: Dominant Negative
Cell lines were monitored for mycoplasma contamination using the Gen-Probe kit (Gen Probe Inc. San Diego CA) according the manufacturer's instructions. By this method the cell lines remained free of mycoplasma for the duration of the study.
Cell Cycle Checkpoint Analysis
DNA damage checkpoint responses were assessed following exposure to 1.5 Gy of IR from a 137Cs source (GammaCell 40, MDS Nordion, Canada) at a dose-rate of 86 rads per minute. G1 checkpoint function was assessed by measuring 5-bromo-2'-deoxy-uridine (BrdU, Sigma Chemical Co.) incorporation from six to eight hours following exposure to 1.5 Gy as previously described [23-25]. Flow cytometric and microscopic determination of mitotic indices were shown to yield equivalent results [26,27]. DNA damage G2 checkpoint function was assessed by determining the mitotic index of cultures two hours following irradiation. Mitotic index was determined using flow cytometry to measure the number of cells expressing the phospho-histone H3 mitotic epitope or by directly counting Giemsa- or DAPI-stained mitotic figures as previously described [28-30] Spindle damage checkpoint function was assessed by seeding cells into medium containing 100 ng/ml colcemid (Sigma Chemical Co.) for 24 or 48 hours. The cells were labeled with BrdU during the last two hours of this incubation. Cells were then analyzed by flow cytometry as described above to determine the percentage of cells with >4n DNA content.
Chromosomal Analyses
Metaphase spreads were prepared from the 8 cell lines listed in Table 1 at the earliest possible PDL following selection and then every 10 – 20 PD thereafter until telomerase-negative cells reached crisis. All metaphase preparations were done according to previously described methods [6]. Fifty metaphases from each cell line at each PDL were analyzed and scored for number of chromosomes, and the numbers and types of structural abnormalities. G-banding was done according to standard protocols [31] and representative karyotypes were assembled after analysis of 20 to 25 metaphases.
Western Blot Analysis
Logarithmically growing cells were seeded at 5 × 105 per 100-mm dish and incubated for 48 hr. Cultures were irradiated as described above and incubated for 6 hr at 37°C. Cells were harvested by trypsinization, washed once in PBS, and resuspended in lysis buffer (100 mM sodium phosphate buffer, pH 7.2, 10 mM EDTA, 10 mM EGTA, 1.5 M NaCl, 10% NP40, supplemented with 10 mM 4-(2-aminoethyl) benzenesulfonyl fluoride (AEBSF, Sigma Chemical Co.), 10 mM β-glycerophosphate (Sigma Chemical Co.), 10 mM sodium orthovanadate (Sigma Chemical Co.), and 10 ug/ml of leupeptin (Sigma Chemical Co.) and aprotinin Sigma Chemical Co.). Protein concentrations were determined using the Bio-Rad DC Protein Assay (Bio-Rad Laboratories) according to the manufacturer's protocol. Samples containing 100 μg protein were mixed with an equal volume of 2 × Laemmli sample buffer (125 mM Tris-HCl, pH 6.8, 4% SDS, 20% glycerol) containing 5% β-mercaptoethanol (Sigma Chemical Co.), boiled, and separated by SDS-PAGE. Proteins were transferred to nitrocellulose and probed with antibody against p21Waf1(Neomarkers) and detected with goat anti-rabbit HRP using the ECL substrate (both Amersham Pharmacia Biotech).
Results
Validation of Cell Lines
This study utilized eight isogenic cell lines differing only in expression of telomerase and p53 function as listed in Table 1. Following selection the cells transduced with DSWK-8 (F1-hTERT lines) were assayed for telomerase expression by TRAP assay [32-34]. As shown in Figure 1, all cell lines transduced with DSWK-8 were telomerase-positive. Cell lines transduced with the empty telomerase vector (HIT) were telomerase-negative (data not shown). Western immunoblot analysis confirmed there was significant overexpression of p53 in lines expressing the p53-V143A and p53-H179Q alleles, and no detectable p53 in lines expressing HPV16E6 (not shown).
Figure 1 Assessment of telomerase activity in cell lines transduced with DSWK-8 by TRAP assay. Each of these cell lines exhibits an RNase-sensitive PCR product.
The ability of the cells to delay entry into S-phase following exposure to 1.5 Gy of ionizing radiation (IR) was assessed as a quantitative index of p53-dependent G1 checkpoint function (Figure 2). The F1-HIT+LXIN and F1-hTERT+LXIN cell lines that have an intact p53 signaling pathway displayed an effective G1 checkpoint response to DNA damage. In this line the percent of cells in the first half of S phase 6 – 8 h after irradiation was reduced by >75% due to a G1 arrest. Cells transduced with HPV16E6 and p53-H179Q exhibited severely attenuated G1 checkpoint function. Less than 15% of HPV16E6-expressing cells were delayed in G1 while cells expressing p53-H179Q had <25% arrested in G1 post-irradiation. Cell lines expressing the p53-A143V dominant-negative form of p53 retained approximately half of the normal G1 checkpoint response with about 50% of irradiated cells delayed in G1. Expression of telomerase had no effect on the radiation-induced G1 arrest.
Figure 2 G1 checkpoint analysis of cell lines. Cells at population doubling level 25 – 30 were tested for G1 checkpoint function. Incorporation of BrdU was analyzed 6 – 8 hours after exposure to 1.5 Gy. The radiation-induced reduction in the percentage of cells in the first half of S-phase was determined as a quantitative measure of G1 checkpoint function.
Immunoblot analysis of p21Waf1 expression confirmed the biological analysis of the G1 checkpoint (Figure 3). Cell lines transduced with the empty LXIN vector expressed p21Waf1 in sham-treated controls and expression was induced after treatment with IR. Lines expressing HPV16E6 and p53-H179Q, which displayed severe attenuation of G1 checkpoint function, did not express p21Waf1 in sham-treated controls nor after irradiation. Lines expressing p53-A143V did not display full ablation of expression or induction of p21Waf1 as evident by the low level of expression in sham-treated controls and some induction of protein after irradiation. As was the case for radiation-induced G1 arrest, expression of telomerase did not affect the expression or induction of p21Waf1. Expression of HPV16E6 and p53-H179Q ablated expression of p21Waf1 and induced a severe attenuation of G1 checkpoint function, while expression of p53-A143V attenuated expression of p21Waf1 while reducing G1 checkpoint function by about 50%. Thus the p53-A143V lines displayed only a partial loss of p53 function.
Figure 3 Assessment of p53-dependent induction of p21Waf1 following IR. Western immuno blot to ascertain p21Waf1 induction 6 hr following 1.5 Gy IR was done as described in Methods.
In contrast to their differing responses in the DNA damage G1 checkpoint the non-telomerized F1-HIT+p53-A143V line behaved like the non-telomerized F1-HIT+HPV16E6 and F1-HIT+p53-H179Q lines and bypassed the replicative senescence checkpoint during in vitro aging. Cell population expansion was monitored continuously and all of the telomerase-negative lines initially displayed equivalent growth in vitro (Figure 4A). Cells expressing HPV16E6 or the dominant-negative alleles of p53 continued to grow for 15 – 20 population doublings beyond the 60 PDL at which F1-HIT+LXIN senesced and arrested growth (Figure 4A). After PDL 78 cell death exceeded cell birth in the telomerase-negative, E6-expressing culture, and the culture died by what is classically known as telomere crisis [35]. Although population doublings did not increase beyond PDL 80 – 85 in the p53-A143V and p53-H179Q lines for a period of about 18 weeks and the cells appeared to be in crisis, viable cells nevertheless remained on dishes. After 36 weeks in culture, population expansion resumed and two immortal lines were recovered. The behavior of the cell lines was similar to that detailed previously [25] with p53-effective, telomerase-negative fibroblasts undergoing replicative senescence after 60 population doublings, and the p53-defective, telomerase-negative lines bypassing senescence and then undergoing telomere crisis.
Figure 4 Growth curves of transduced fibroblasts. The x-axis is the number of weeks of continuous culture. The y-axis represents the number of population doublings the cultures had accumulated. A. (◆) F1-HIT+LXIN; (■) F1-HIT+p53-A143V; (▲) F1-HIT+p53-H179Q; F1-HIT+E6. The empty vector control (●) F1-HIT+LXIN cell line underwent senescence at PDL 61 while the cells with defective p53 function continued to divide for an additional 10 – 20 population doublings. At this point cell death equaled cell division, resulting in no net gain of cell number over time. During the approximate 16-week duration of this phase the E6-expressing cells died. B. Continuous growth of hTERT-expressing lines. (◆) F1-hTERT+LXIN; (■) F1-hTERT+p53-A143V; (●) F1-hTERT+p53-H179Q; F1-hTERT+E6.
Population expansion in the lines transduced directly with hTERT was continuous and equivalent to that seen in telomerase-negative lines at PDL 40 – 60 and in the spontaneously immortalized lines at PDL 90 – 110 (Figure 4B). The hTERT-expressing lines were carried to PDL >175 without reduction in growth rate. P53-dependent G1 checkpoint function was monitored during the various phases of cellular aging in vitro and found not to vary substantially (Table 2). Additionally the spontaneously immortalized cell line expressing the p53-A143V dominant negative p53 allele was still able to induce a small amount of p21Waf1 following exposure to 1.5 Gy (Figure 5)
Table 2 G1 Checkpoint Function of Aging Cell Lines
% of Cells Exhibiting a G1 Delay
PDL1 F1-hTERT+LXIN F1-hTERT+p53-A143V F1-hTERT+p53-H179Q F1-hTERT+E6 F1-HIT+LXIN F1-HIT+p53-A143V F1-HIT+p53-H179Q F1-HIT+E6
30 – 35 100 51 11 0 77 56 22 12
40 – 50 94 44 20 6 95 55 29 0
60 – 65 92 65 5 7 85 54 10 5
75 – 85 80 57 17 1
>1002 97 66 0 0 373 03
Avg.4 93 ± 8 57 ± 9 11 ± 8 3 ± 3 86 ± 9 55 ± 1 20 ± 10 6 ± 6
1. Population Doubling Level
2. PDL of hTERT expressing lines 190 – 210
3. Spontaneous Immortalized Derivative
4. Average ± standard deviation. Average for HIT lines excludes spontaneous immortals.
Figure 5 p53-A143V-expressing cells are able to partially induce p21Waf1. Western-immuno blot to demonstrating p21Waf1 induction in p53-A143V-expressing cells. Log-phase cultures were irradiated with 1.5 Gy. Six hours post-irradiation, the cells were harvested and lysed in loading buffer.
A previous report indicated that some dominant-negative p53 alleles induced a gain of function [36]. This gain of function was identified using a "spindle damage" assay that measures the ability of cells to become polyploid when incubated with microtubule poisons such as colcemid. This phenomenon was examined in the four-telomerized cell lines derived for this study. Following 24 or 48 hours incubation in 100 ng/ml colcemid, cells were labeled for two hours with BrdU and then analyzed by flow cytometry to determine the rate of DNA synthesis in diploid and tetraploid nuclei. As shown in Figure 6, all cell lines with defective p53 signaling underwent endoreduplication and displayed increased frequencies of tetraploid S-phase cells when incubated in colcemid. The isogenic F1-hTERT+LXIN line with effective p53-dependent G1 checkpoint function did not display this endoreduplication when incubated with colcemid. Thus, inactivation of p53 expression with HPV16E6 oncoprotein induced the same susceptibility to endoreduplication during incubation in colcemid as was seen using dominant-negative mutant p53 alleles to disrupt p53 signaling.
Figure 6 Assessment of spindle damage checkpoint function in p53-defective fibroblast lines. Endoreduplication was assessed by flow cytometric analysis of BrdU-labeled cells 24 and 48 h after addition of colcemid to culture medium.
Chromosomal Instability
An assessment of chromosomal integrity was done on all eight cell lines within five population doublings of gene transduction to assess the background level of structural and numerical abnormalities in the population and then at PDL 60 (normal replicative senescence point), and PDL 75 – 85 (crisis). Table 3 demonstrates that there were no differences among the various cell lines at the first PDL examined. As was previously observed upon transduction of telomerase-negative fibroblasts with HPV16E6 [6], soon after inactivation of p53 with the dominant-negative alleles, chromosomal number and structure appeared normal. However, as the lines aged and approached the normal replicative senescence point of 60 population doublings, the number of metaphases exhibiting structural and numerical abnormalities increased dramatically in the telomerase-negative cell lines with defective p53 function. At PDL 76 – 77, the majority of metaphases derived from the telomerase-negative HPV16E6-, p53-A143V-, and p53-H179Q-expressing cells exhibited hypodiploidy and/or dicentric chromosomes. The p53-defective lines that were transduced with hTERT to express telomerase did not display these aging-related instabilities in chromosome numbers and structure.
Table 3 Quantification of Structural and Numerical Chromosomal Abnormalities in Cell Lines
% of Metaphases Containing:
Cell Line PDL1 # Chromosomes
≤ 44 45 – 47 48 – 85 86 – 99 ≥ 100 Dicentrics + Rings TA2 Breaks Fragments Other3
F1-hTERT+LXIN 32 9 85 0 7 0 0 0 2 0 0
60 2 93 2 3 0 0 2 2 0 0
76 7 93 0 0 0 0 0 0 0 0
F1-hTERT+p53-A143V 31 5 86 3 5 2 0 3 0 0 0
66 0 90 4 4 2 2 0 2 0 0
85 0 96 2 2 0 0 5 4 0 0
F1-hTERT+p53-H179Q 33 10 88 0 2 0 0 0 0 0 0
63 2 87 2 9 0 0 0 2 2 0
84 2 98 0 0 0 0 0 0 0 0
F1-hTERT+E6 30 0 86 4 11 0 0 0 0 2 0
63 2 93 4 2 0 4 0 0 0 0
84 2 94 2 0 0 2 0 2 0 0
F1-HIT+LXIN4 31 10 85 2 2 2 0 0 0 0 0
59 2 97 0 2 0 5 0 0 0 0
F1-HIT+p53-A143V 32 8 89 0 4 0 0 0 4 0 0
65 27 48 10 10 5 53 2 3 5 7
76 44 32 7 11 4 22 2 2 20 0
F1-HIT+p53-H179Q 33 5 91 2 0 0 0 0 0 2 2
63 7 76 2 15 0 37 4 2 7 2
77 23 62 4 9 0 38 2 2 8 2
F1-HIT+E6 32 7 79 9 4 2 5 2 2 2 2
63 33 60 2 5 0 43 0 3 7 0
77 66 24 0 5 0 39 10 0 24 0
1. Population Doubling Level
2. Telomere Associations
3. Exchange Aberrations + Deletions
4. These cells senesced at 60 population doublings
Attenuation of DNA damage G2 checkpoint function
Previous studies from this laboratory have demonstrated that DNA damage G2 checkpoint function becomes attenuated in congruence with chromosomal instability [6,25,29]. Figure 7 depicts DNA damage G2 checkpoint function in the cell lines at various in vitro PDLs. The F1-HIT+LXIN and F1-hTERT+LXIN lines displayed a typically effective G2 checkpoint response with on average >95% of G2 cells being delayed in their entry to mitosis after treatment with 1.5 Gy. Expression of the dominant-negative p53 alleles and HPV16E6 induced a modest attenuation of G2 checkpoint function measured at PDL 30 – 40 with 7 – 24% of p53-defective cells evading radiation-induced G2 delay. The telomerase-negative lines expressing p53-A143V and HPV16E6 displayed further severe attenuation of G2 checkpoint function with aging in vitro. For these cells at PDL 70 – 80, the mitotic index in irradiated cells was about half of that seen in sham-treated controls. The telomerase-negative, p53-H179Q line displayed a more modest decrement of G2 checkpoint function during aging with at most 17% of cells evading G2 delay. This represents the first example in a total of seven independent analyses of p53-defective human fibroblasts [6,25,37] in which G2 checkpoint function did not appear to be severely attenuated in cells at crisis. Interestingly the G2 checkpoint response of the two spontaneously derived immortal lines (F1-HIT+p53-A143V Immortal and F1-HIT+p53-H179Q Immortal) was very similar to the younger (precrisis) parental lines (Figure 7).
Figure 7 DNA damage G2 checkpoint function in aging fibroblasts. Log-phase cultures at low and high PDL were treated with 1.5 Gy of IR then incubated for 2 h before determination of mitotic index. The percentage of mitotic cells in irradiated cultures was divided by the percentage of mitotic cells in sham-treated controls to determine the percent of cells evading G2 delay [25].
In contrast to the aging-associated attenuation of G2 checkpoint function in the telomerase-negative, p53-defective lines, there was no aging-associated attenuation of G2 checkpoint function in the telomerase-expressing, p53-defective lines. Cells at >150 PDL displayed a response to IR that was nearly equivalent to that seen at PDL 30 – 40. Thus, the age-related attenuation of G2 checkpoint function in p53-defective lines was prevented entirely by expression of telomerase.
Cytogenetics of immortal lines
Spontaneously immortalized cells emerged from crisis after three months in culture (F1-HIT+p53-A143V Immortal and F1-HIT+p53-H179Q Immortal cell lines). These two cell lines had DNA damage G1 and G2 checkpoint responses that were similar to those seen in their low-passage parents (Table 2 and Figure 7, respectively). G-band karyotype analysis of these immortal cell lines revealed that both were near diploid (45 – 46 chromosomes, Table 4). The F1-HIT+p53-H179Q immortal line exhibited few abnormalities (12% 18q+, 8% 18q-) with 92% of metaphases being 45 XY or 46 XY. The F1-HIT+p53-A143V immortal line exhibited a somewhat reduced number of diploid metaphases (72%) and an increased number of polyploid metaphases (20%) as well as a small deletion at 4p16 in greater than 50% of metaphases. The lines that were transduced with hTERT when at low PDL were maintained in culture for greater that 175 population doublings. Up to this point, the telomerase positive, p53-defective lines displayed a normal diploid karyotype with no marker chromosomal aberrations (Table 4). However, the F1-hTERT+LXIN control line exhibited a trisomy for chromosome 8 in 92% of metaphases. Interestingly the F1-hTERT+LXIN cell line had a normal karyotype (25 of 25 metaphases) at an intermediate PDL of 96.
Table 4 Karyotypic Analysis of Cell Lines
Cell Line Chromosome Number1 Karyotype2
F1-hTERT+LXIN (3PDL 173.5) 46, XY (16%) +8 [72%]
47, XY (76%) +8, -Y [8%]
+8, -16
+8, -11
+8, 5q-
F1-hTERT+p53-A143V (PDL 217.6) 46, XY (80%) t(1;4)
12q-
-13
-X
-15
+ marker
F1-hTERT+p53-H179Q (PDL 200.8) 46, XY (84%) +20
12p-
15q-
19q-
-7
+ marker (iso 5p?)
-21
F1-hTERT+E6 (PDL 160.2) 46, XY (80%) -Y
-Y, 11q-
-6, 12p-
-15
+4
9q chromatid break
F1-HIT+p53-A143V Immortal 45, XY (16%) 4p- [24%]
46, XY (56%) 4p-, 4q- [16%]
>90 (20%) two 4p-, two 4q-, dic 13q;13q, four of every chromosome
-Y
6q-
-13, 4p+
-15, -18, 3p-, 2 markers
-15, -18, 3p-
-19, 4p-
4p-, dic 2p;2p
-14, -15, t(5q;14q), t(4p;15q)
-22, 4p-
-20, 10p+
seven 22+
F1-HIT+p53-H179Q Immortal 45, XY (20%) -4, 18q-
46, XY (72%) 4p+, -11, -16, 18q-
18q+ [8%]
16q-
-21, -22, 18q+
-18, 4p+
-15
-17
-18
4q chromatid break
1. Twenty five metaphases analyzed. Number in parenthesis is percentage of metaphases with given number of chromosomes.
2. Karyotype of samples. Number in brackets is percentage of metaphases examined with given karyotype.
3. PDL, population doubling level. PDL of the spontaneously immortalized lines cannot be determined due to the extended period of time the cells spent in crisis.
Discussion
Induced expression of telomerase at the time of inactivation or attenuation of p53 function fully blocked the chromosomal destabilization that is commonly associated with defective p53 in human cells. Thus aneuploidization, polyploidization, and formation of chromosomal aberrations including rings, dicentrics, breaks and exchanges, all of which develop in human cells with inherited or transduced mutant p53 alleles [38,39] or after transduction of viral oncoproteins that inactivate p53 [40-42] appear to be secondary consequences of telomere erosion. Telomerase-expressing fibroblasts with little or no p53-dependent G1 checkpoint function exhibited no detectable numerical or structural chromosomal alterations after being carried through >170 population doublings. This finding has implications for the mechanisms of genetic instability in cancer and its precursors.
In this study eight isogenic cell lines were derived that differed from each other by one or two alleles. All cell lines expressing dominant-negative forms of p53 or the HPV16E6 oncoprotein to interfere with p53 signaling bypassed the senescence checkpoint. However, expression of the p53-A143V allele caused only a 50 % attenuation of DNA damage G1 checkpoint function regardless of telomerase expression. The partial expression of this checkpoint response was associated with partial induction of p21Waf1 protein 6 hr after exposure to 1.5 Gy IR (Figure 3, and Figure 5). This observation is consistent with the report that the V143A allele of p53 binds to and transactivates p53-responsive promoters [43]. When transduced by transfection the A143V allele of p53 has been reported to override the senescence checkpoint and induce severe chromosomal destabilization as seen here [44].
The continued erosion of telomeres in cells bypassing the replicative senescence checkpoint results in a cycle of formation of dicentric chromosomes leading to non-disjunction errors and broken chromosomes that, when repaired, may result in translocations, deletions, or more dicentrics. The telomerase-expressing p53 defective cells used in this study accumulated no structural or numerical abnormalities when grown in culture for greater than 170 population doublings. In contrast to other reports, the telomerase-positive control cell line (F1-hTERT+LXIN) demonstrated no evidence of chromosomal destabilization or morphological changes through approximately 100 PD's [14,45]. However during the next 75 population doublings this cell line did acquire a trisomy for chromosome 8. The trisomy for chromosome 8 seen in the F1-hTERT+LXIN cells at PDL 175 has also been observed in this laboratory in another telomerized human fibroblast line around PDL 200 [46]. Trisomy for chromosome 8 is commonly found in many hyperproliferative disorders and is believed to provide a growth advantage to those cells that acquire it [47-50] Such a growth advantage could have selected for the cell or cells that underwent nondisjunction and acquired the extra chromosome, although population expansion by the F1-hTERT+LXIN line did not increase after PDL 100 (Figure 4B). Recent studies demonstrate that the forced expression of telomerase in skin fibroblasts and other cell types enhances cell proliferation [9]. These effects of telomerase did not appear to alter cellular responses to IR-induced DNA damage as shown here or UV-induced DNA damage as previously reported [51].
Chromosomal destabilization associated with telomere crisis has also been correlated with attenuation and inactivation of DNA damage G2 checkpoint function [6,25,29]. Because this checkpoint blocks mitosis by cells with damaged chromatids, it was hypothesized that chromosomally unstable cells with defects in the G2 checkpoint would have a growth advantage and accumulate in cultures because of selection. An alternative explanation is the severe instability of chromosomes at crisis impeded progression through mitosis, so that upon irradiation damaged cells complete mitosis at a slower than normal rate. Yang et al. described a rad9-dependent checkpoint in S. cerevisiae that responds to the presence of a dicentric chromosome to delay progression through mitosis [52]. If human fibroblasts with dicentric chromosomes are also delayed in their progression through mitosis, this phenomenon could explain the attenuation of G2 checkpoint function seen as p53-defective cells enter crisis. As the efficacy of G2 checkpoint function is quantified by the degree of emptying of the mitotic compartment 2 h after irradiation, a reduced rate of progression through mitosis to G1 will have the effect of sustaining a higher mitotic index in control and irradiated cells. This explanation would view the attenuation of G2 checkpoint function during telomere crisis as another manifestation of chromosomal instability. The correction of G2 checkpoint function in the immortal lines that emerged from crisis with stable chromosomes also supports this explanation. There are other ways to attenuate DNA damage G2 checkpoint function such as inactivation of ATM [53], BRCA1 and 14-3-3ε [54] or overexpression of cyclin B1/Cdk1 kinase [55]. The common attenuation of DNA damage G2 checkpoint function in small cell lung cancer but not squamous cell carcinoma [54] implies that this pathway of DNA damage response protects against carcinogenesis in selected targets.
Human fibroblasts with inherited mutations in p53 or ectopic expression of HPV16E6 spontaneously immortalize at a low frequency [6,38]. During crisis, prior to immortalization, these cell lines exhibit both numerical and structural chromosomal abnormalities. The loss of chromosomes (Table 3) seen in these lines may be due to non-disjunction errors or artifacts associated with the processing of the metaphase preparations. It is not possible to distinguish between these possibilities. However, since this was not observed in the telomerized cell lines or in the non-telomerized cell lines prior to crisis, this phenomenon is likely to be a function of the underlying chromosomal instability. The cell lines, which result from spontaneous immortalization, have usually activated telomerase expression although in some cases telomeres appear to be stabilized by an alternative mechanism (ALT) [56]. These immortal lines are often aneuploid [38]. In this study, the F1-HIT+p53-A143V and F1-HIT+p53-H179Q lines yielded immortal derivatives. These immortal lines were both diploid (45 – 46 chromosomes). The p53-A143V line displayed a marker chromosomal aberration (4p-) in over half of the metaphases. Deletions in this region are associated with Wolf-Hirschhorn syndrome [57]. The p53-H179Q immortal line primarily contained aberrations in chromosome 18 (both losses and gains). Chromosome 18 aberrations have been associated with various malignancies in humans [58,59]. These findings demonstrating only subtle changes in the karyotype following immortalization are consistent with the previous observation that aneuploidy is not an inevitable outcome of in vitro immortalization [60]. Taken together, these results may reflect the fact that all that is required for continuous in vitro growth of human fibroblasts is hTERT expression. Genetic changes that are not cytogenetically detectable apparently are able to derepress hTERT expression and induce immortality.
Authors' contributions
DAS constructed cell lines, carried out checkpoint analysis, participated in karyotype studies, study design, and drafted manuscript. EL carried out the G-banding of chromosomes and fine karyotypic analysis. TPH carried out the Western blot analysis of p53 and p21 induction. WKK was instrumental in study design and manuscript preparation.
Acknowledgements
We are grateful to Dr. Marila Corderio-Stone for helpful comments. This work was supported by PHS grant CA81343 and center grants P30-CA16086 and P30-ES10126. DAS was supported by training grants ES12345 and ES45678. TPH was supported by training grant ES07017.
==== Refs
Houck JC Sharma VK Hayflick L Functional failures of cultured human diploid fibroblasts after continued population doublings Proceedings of the Society for Experimental Biology & Medicine 1971 137 331 333 5581670
Counter CM Avilion AA LeFeuvre CE Stewart NG Greider CW Harley CB Bacchetti S Telomere shortening associated with chromosome instability is arrested in immortal cells which express telomerase activity EMBO Journal 1992 11 1921 1929 1582420
Proctor CJ Kirkwood TB Modelling cellular senescence as a result of telomere state Aging Cell 2003 2 151 157 12882407 10.1046/j.1474-9728.2003.00050.x
Masutomi K Yu EY Khurts S Ben-Porath I Currier JL Metz GB Brooks MW Kaneko S Murakami S DeCaprio JA Weinberg RA Stewart SA Hahn WC Telomerase maintains telomere structure in normal human cells Cell 2003 114 241 253 12887925 10.1016/S0092-8674(03)00550-6
Shay JW Wright WE Hayflick, his limit, and cellular ageing Nature Reviews Molecular Cell Biology 2000 1 72 76 11413492 10.1038/35036093
Filatov L Golubovskaya V Hurt JC Byrd LL Phillips JM Kaufmann WK Chromosomal instability is correlated with telomere erosion and inactivation of G2 checkpoint function in human fibroblasts expressing human papillomavirus type 16 E6 oncoprotein Oncogene 1998 16 1825 1838 9583680 10.1038/sj.onc.1201711
Counter CM Hahn WC Wei W Caddle SD Beijersbergen RL Lansdorp PM Sedivy JM Weinberg RA Dissociation among in vitro telomerase activity, telomere maintenance, and cellular immortalization [see comments] Proceedings of the National Academy of Sciences of the United States of America 1998 95 14723 14728 9843956 10.1073/pnas.95.25.14723
Halvorsen TL Leibowitz G Levine F Telomerase activity is sufficient to allow transformed cells to escape from crisis Molecular & Cellular Biology 1999 19 1864 1870 10022873
Zhu J Wang H Bishop JM Blackburn EH Telomerase extends the lifespan of virus-transformed human cells without net telomere lengthening [see comments] Proceedings of the National Academy of Sciences of the United States of America 1999 96 3723 3728 10097104 10.1073/pnas.96.7.3723
Kaufmann WK Behe CI Golubovskaya VM Byrd LL Albright CD Borchet KM Presnell SC Coleman WB Grisham JW Smith GJ Aberrant cell cycle checkpoint function in transformed hepatocytes and WB-F344 hepatic epithelial stem-like cells Carcinogenesis 2001 22 1257 1269 11470757 10.1093/carcin/22.8.1257
Jiang XR Jimenez G Chang E Frolkis M Kusler B Sage Beeche M Bodnar AG Wahl GM Tlsty TD Chiu CP Telomerase expression in human somatic cells does not induce changes associated with a transformed phenotype Nature Genetics 1999 21 111 114 9916802 10.1038/5056
van Steensel B Smogorzewska A de Lange T TRF2 protects human telomeres from end-to-end fusions Cell 1998 92 401 413 9476899 10.1016/S0092-8674(00)80932-0
Franco S MacKenzie KL Dias S Alvarez S Rafii S Moore MA Clonal variation in phenotype and life span of human embryonic fibroblasts (MRC-5) transduced with the catalytic component of telomerase (hTERT) ExpCell Res 2001 268 14 25 10.1006/excr.2001.5264
MacKenzie KL Franco S May C Sadelain M Moore MA Mass cultured human fibroblasts overexpressing hTERT encounter a growth crisis following an extended period of proliferation ExpCell Res 2000 259 336 350 10.1006/excr.2000.4982
Milyavsky M Shats I Erez N Tang X Senderovich S Meerson A Tabach Y Goldfinger N Ginsberg D Harris CC Rotter V Prolonged culture of telomerase-immortalized human fibroblasts leads to a premalignant phenotype Cancer Res 2003 63 7147 7157 14612508
Noble JR Zhong ZH Neumann AA Melki JR Clark SJ Reddel RR Alterations in the p16(INK4a) and p53 tumor suppressor genes of hTERT-immortalized human fibroblasts Oncogene 2004 14743210
Sambrook J Fritsch EF Maniatis T Molecular Cloning, A Laboratory Manual 1989 2 Cold Spring Harbor Laboratory Press
Comstock KE Watson NF Olsen JC Design of retroviral expression vectors Methods in Molecular Biology 1997 62 207 222 9108523
Johnson LG Mewshaw JP Ni H Friedmann T Boucher RC Olsen JC Effect of host modification and age on airway epithelial gene transfer mediated by a murine leukemia virus-derived vector J Virol 1998 72 8861 8872 9765431
Olsen JC Sechelski J Use of sodium butyrate to enhance production of retroviral vectors expressing CFTR cDNA Human Gene Therapy 1995 6 1195 1202 8527478
Boyer JC Kaufmann WK Brylawski BP Cordeiro-Stone M Defective postreplication repair in xeroderma pigmentosum variant fibroblasts Cancer Research 1990 50 2593 2598 2109654
Olsen JC Johnson LG Wong-Sun ML Moore KL Swanstrom Boucher RC Retrovirus-mediated gene transfer to cystic fibrosis airway epithelial cells: effect of selectable marker sequences on long-term expression Nucleic Acids Research 1993 21 663 669 7680124
Kastan MB Onyekwere O Sidransky D Vogelstein B Craig RW Participation of p53 protein in the cellular response to DNA damage Cancer Res 1991 51 6304 6311 1933891
White RA Terry NH Meistrich ML Calkins DP Improved method for computing potential doubling time from flow cytometric data Cytometry 1990 11 314 317 2318085 10.1002/cyto.990110214
Kaufmann WK Schwartz JL Hurt JC Byrd LL Galloway DA Levedakou E Paules RS Inactivation of G2 checkpoint function and chromosomal destabilization are linked in human fibroblasts expressing human papillomavirus type 16 E6 Cell Growth & Differentiation 1997 8 1105 1114 9342189
Doherty SC McKeown SR McKelvey-Martin V Downes CS Atala A Yoo JJ Simpson DA Kaufmann WK Cell cycle checkpoint function in bladder cancer J Natl Cancer Inst 2003 95 1859 1868 14679155
Kaufmann WK Heffernan TP Beaulieu LM Doherty S Frank AR Zhou Y Bryant MF Zhou T Luche DD Nikolaishvili-Feinberg N Simpson DA Cordeiro-Stone M Caffeine and human DNA metabolism: the magic and the mystery Mutat Res 2003 532 85 102 14643431
Juan G Traganos F James WM Ray JM Roberge M Sauve DM Anderson H Darzynkiewicz Z Histone H3 phosphorylation and expression of cyclins A and B1 measured in individual cells during their progression through G2 and mitosis Cytometry 1998 32 71 77 9627219 10.1002/(SICI)1097-0320(19980601)32:2<71::AID-CYTO1>3.0.CO;2-H
Kaufmann WK Levedakou EN Grady HL Paules RS Stein GH Attenuation of G2 checkpoint function precedes human cell immortalization Cancer Research 1995 55 7 11 7805043
Xu B Kim S Kastan MB Involvement of Brca1 in S-Phase and G2-Phase Checkpoints after Ionizing Irradiation Molecular and Cellular Biology 2001 21 3445 3450 11313470 10.1128/MCB.21.10.3445-3450.2001
Wang HC Fedoroff S Banding in human chromosomes treated with trypsin Nat New Biol 1972 235 52 54 4109406
Falchetti ML Levi A Molinari P Verna R D'Ambrosio E Increased sensitivity and reproducibility of TRAP assay by avoiding direct primers interaction Nucleic Acids Research 1998 26 862 863 9443983 10.1093/nar/26.3.862
Kim NW Wu F Advances in quantification and characterization of telomerase activity by the telomeric repeat amplification protocol (TRAP) Nucleic Acids Research 1997 25 2595 2597 9185569 10.1093/nar/25.13.2595
Krupp G Kuhne K Tamm S Klapper W Heidorn K Rott A Parwaresch R Molecular basis of artifacts in the detection of telomerase activity and a modified primer for a more robust 'TRAP' assay Nucleic Acids Research 1997 25 919 921 9016650 10.1093/nar/25.4.919
Ishikawa F Telomere crisis, the driving force in cancer cell evolution Biochem Biophys Res Commun 1997 230 1 6 9020020 10.1006/bbrc.1996.5928
Gualberto A Aldape K Kozakiewicz K Tlsty TD An oncogenic form of p53 confers a dominant, gain-of-function phenotype that disrupts spindle checkpoint control Proceedings of the National Academy of Sciences of the United States of America 1998 95 5166 5171 9560247 10.1073/pnas.95.9.5166
Kaufmann WK Campbell CB Simpson DA Deming PB Filatov L Galloway DA Zhao XJ Creighton AM Downes CS Degradation of ATM-independent decatenation checkpoint function in human cells is secondary to inactivation of p53 and correlated with chromosomal destabilization Cell Cycle 2002 1 210 219 12429935
Bischoff FZ Yim SO Pathak S Grant G Siciliano MJ Giovanella BC Strong LC Tainsky MA Spontaneous abnormalities in normal fibroblasts from patients with Li- Fraumeni cancer syndrome: aneuploidy and immortalization Cancer Res 1990 50 7979 7984 2253239
Liu PK Kraus E Wu TA Strong LC Tainsky MA Analysis of genomic instability in Li-Fraumeni fibroblasts with germline p53 mutations Oncogene 1996 12 2267 2278 8649766
Schramayr S Caporossi D Mak I Jelinek T Bacchetti S Chromosomal damage induced by human adenovirus type 12 requires expression of the E1B 55-kilodalton viral protein JVirol 1990 64 2090 2095 2325204
Vieten L Belair CD Savelieva L Julicher K Brocker F Bardenheuer W Schutte J Opalka B Reznikoff CA Minimal deletion of 3p13-->14.2 associated with immortalization of human uroepithelial cells Genes ChromosomesCancer 1998 21 39 48 10.1002/(SICI)1098-2264(199801)21:1<39::AID-GCC6>3.0.CO;2-9
Xiong Y Kuppuswamy D Li Y Livanos EM Hixon M White A Beach D Tlsty TD Alteration of cell cycle kinase complexes in human papillomavirus E6- and E7-expressing fibroblasts precedes neoplastic transformation JVirol 1996 70 999 1008 8551641
Zhang W Funk WD Wright WE Shay JW Deisseroth AB Novel DNA binding of p53 mutants and their role in transcriptional activation Oncogene 1993 8 2555 2559 8361764
Liu X Nishitani J McQuirter JL Baluda MA Park NH The temperature sensitive mutant p53-143ala extends in vitro life span, promotes errors in DNA replication and impairs DNA repair in normal human oral keratinocytes Cell Mol Biol (Noisy-le-grand) 2001 47 1169 1178 11838964
Vaziri H Squire JA Pandita TK Bradley G Kuba RM Zhang H Gulyas S Hill RP Nolan GP Benchimol S Analysis of genomic integrity and p53-dependent G1 checkpoint in telomerase-induced extended-life-span human fibroblasts Molecular & Cellular Biology 1999 19 2373 2379 10022923
Cordeiro-Stone M Frank A Bryant M Oguejiofor I Hatch SB McDaniel LD Kaufmann WK DNA damage responses protect xeroderma pigmentosum variant from UVC-induced clastogenesis Carcinogenesis 2002 23 959 965 12082017 10.1093/carcin/23.6.959
Bench AJ Nacheva EP Champion KM Green AR Molecular genetics and cytogenetics of myeloproliferative disorders Baillieres Clin Haematol 1998 11 819 848 10640219
Novak R Dasu S Agamanolis D Herold W Malone J Waterson J Trisomy 8 is a characteristic finding in pleuropulmonary blastoma Pediatr Pathol Lab Med 1997 17 99 103 9050063 10.1080/107710497175056
Sankary S Dickman PS Wiener E Robichaux W Swaney WP Malone PS Gollin SM Consistent numerical chromosome aberrations in congenital fibrosarcoma Cancer Genet Cytogenet 1993 65 152 156 8453601 10.1016/0165-4608(93)90225-B
Sotos JF Genetic disorders associated with overgrowth Clin Pediatr (Phila) 1997 36 39 49 9007346
Heffernan TP Simpson DA Frank AR Heinloth AN Paules RS Cordeiro-Stone M Kaufmann WK An ATR- and Chk1-dependent S checkpoint inhibits replicon initiation following UVC-induced DNA damage Mol Cell Biol 2002 22 8552 8561 12446774 10.1128/MCB.22.24.8552-8561.2002
Yang SS Yeh E Salmon ED Bloom K Identification of a mid-anaphase checkpoint in budding yeast JCell Biol 1997 136 345 354 9015305 10.1083/jcb.136.2.345
Paules RS Levedakou EN Wilson SJ Innes CL Rhodes N Tlsty TD Galloway DA Donehower LA Tainsky MA Kaufmann WK Defective G2 checkpoint function in cells from individuals with familial cancer syndromes Cancer Research 1995 55 1763 1773 7712486
Konishi H Nakagawa T Harano T Mizuno K Saito H Masuda A Matsuda H Osada H Takahashi T Identification of frequent G(2) checkpoint impairment and a homozygous deletion of 14-3-3epsilon at 17p13.3 in small cell lung cancers Cancer Res 2002 62 271 276 11782387
Park M Chae HD Yun J Jung M Kim YS Kim SH Han MH Shin DY Constitutive activation of cyclin B1-associated cdc2 kinase overrides p53-mediated G2-M arrest Cancer Research 2000 60 542 545 10676633
Reddel RR Bryan TM Colgin LM Perrem KT Yeager TR Alternative lengthening of telomeres in human cells RadiatRes 2001 155 194 200
Battaglia A Carey JC Wright TJ Wolf-Hirschhorn (4p-) syndrome Adv Pediatr 2001 48 75 113 11480768
Compton CC Colorectal carcinoma: diagnostic, prognostic, and molecular features Mod Pathol 2003 16 376 388 12692203 10.1097/01.MP.0000062859.46942.93
Oberg K Carcinoid tumors: molecular genetics, tumor biology, and update of diagnosis and treatment Curr Opin Oncol 2002 14 38 45 11790979 10.1097/00001622-200201000-00008
Bunz F Fauth C Speicher MR Dutriaux A Sedivy JM Kinzler KW Vogelstein B Lengauer C Targeted inactivation of p53 in human cells does not result in aneuploidy Cancer Res 2002 62 1129 1133 11861393
|
16209708
|
PMC1262734
|
CC BY
|
2021-01-04 16:39:19
|
no
|
J Carcinog. 2005 Oct 6; 4:18
|
utf-8
|
J Carcinog
| 2,005 |
10.1186/1477-3163-4-18
|
oa_comm
|
==== Front
J CarcinogJournal of Carcinogenesis1477-3163BioMed Central London 1477-3163-4-181620970810.1186/1477-3163-4-18ResearchTelomerase expression is sufficient for chromosomal integrity in cells lacking p53 dependent G1 checkpoint function Simpson Dennis A [email protected] Elizabeth [email protected] Timothy P [email protected] William K [email protected] Department of Pathology and Laboratory Medicine, Lineberger Comprehensive Cancer Center, and Center for Environmental Health and Susceptibility, University of North Carolina at Chapel Hill, CB 7295, Chapel Hill, NC 27599, USA2005 6 10 2005 4 18 18 17 3 2005 6 10 2005 Copyright © 2005 Simpson et al; licensee BioMed Central Ltd.2005Simpson et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Secondary cultures of human fibroblasts display a finite lifespan ending at senescence. Loss of p53 function by mutation or viral oncogene expression bypasses senescence, allowing cell division to continue for an additional 10 – 20 doublings. During this time chromosomal aberrations seen in mitotic cells increase while DNA damage and decatenation checkpoint functions in G2 cells decrease.
Methods
To explore this complex interplay between chromosomal instability and checkpoint dysfunction, human fibroblast lines were derived that expressed HPV16E6 oncoprotein or dominant-negative alleles of p53 (A143V and H179Q) with or without the catalytic subunit of telomerase.
Results
Cells with normal p53 function displayed 86 – 93% G1 arrest after exposure to 1.5 Gy ionizing radiation (IR). Expression of HPV16E6 or p53-H179Q severely attenuated G1 checkpoint function (3 – 20% arrest) while p53-A143V expression induced intermediate attenuation (55 – 57% arrest) irrespective of telomerase expression. All cell lines, regardless of telomerase expression or p53 status, exhibited a normal DNA damage G2 checkpoint response following exposure to 1.5 Gy IR prior to the senescence checkpoint. As telomerase-negative cells bypassed senescence, the frequencies of chromosomal aberrations increased generally congruent with attenuation of G2 checkpoint function. Telomerase expression allowed cells with defective p53 function to grow >175 doublings without chromosomal aberrations or attenuation of G2 checkpoint function.
Conclusion
Thus, chromosomal instability in cells with defective p53 function appears to depend upon telomere erosion not loss of the DNA damage induced G1 checkpoint.
==== Body
Background
Normal diploid fibroblasts proliferate in secondary cultures for a finite number of population doublings until a growth arrest known as replicative senescence, or M1, is reached [1]. This limitation in lifespan is believed to be due to the continuous shortening of the telomeres with each cell division [2]. Recent evidence has suggested that an alteration in the structure of one or more telomeres may, in fact, be what triggers cells to enter replicative senescence, a permanent p53-dependent G1 arrest [3,4]. Regardless of the exact trigger of senescence, inactivation of p53 allows cells to bypass senescence and continue to divide until a second growth restriction termed crisis, or M2, is reached [5]. Cells in crisis contain numerous structural and numerical chromosomal abnormalities which may be due to cycles of chromosome fusion (dicentric chromosomes) and subsequent resolution of the fusion (chromosome break) during mitosis [2]. A previous study has demonstrated that during the phase of extended proliferation after bypass of M1, telomeres in p53-defective, telomerase-negative cells can erode to the point where little or no telomeric repeat DNA can be detected [6]. Chromosomes without telomeres appear to be substrates for DNA repair pathways resulting in telomere associations and formation of dicentric and ring chromosomes. The resolution of these unstable structures is believed to result in the other structural and numerical abnormalities in chromosomes observed in cells in crisis (i.e., breaks, exchanges, aneuploidy, polyploidy).
Prevention of telomere erosion by ectopic expression of the catalytic subunit of human telomerase (hTERT) has been shown to prevent crisis in cells expressing SV40 large T antigen or HPV16E6 oncoprotein [7-10]. Normal diploid human fibroblasts expressing hTERT have been reported to maintain a normal karyotype and preserve cell cycle checkpoint function for at least 200 population doublings [11,12], although others have suggested that otherwise normal telomerase-expressing human fibroblasts do display alterations in expression of tumor suppressor genes, growth characteristics, and transient genetic instability [13-16] These studies have failed to directly address the question as to whether cells can maintain a stable genome in the absence of a functional DNA damage induced G1 checkpoint. Here we report that in the absence of telomere erosion cells defective for p53 signaling can maintain stable genomes for >175 population doublings. This study found that normal diploid cells expressing hTERT maintain a normal karyotype for at least 100 PD's but eventually did become numerically abnormal. We also report that two independent p53-defective lines which emerged from crisis by reactivation of telomerase displayed remarkably stable karyotypes.
Materials and methods
Plasmids and viruses
All cloning steps were carried out according to standard methods [17]. Plasmids were maintained in the DH5α strain of E. coli. Replication-defective retroviruses used in this study and helper plasmids for packaging are shown in Figure 8. The hTERT retroviral expression vector, pDSWK-8, was created by cloning the hTERT cDNA from pBABE/Hyg-hTERT (Dr. Robert A. Weinberg, Whitehead Institute for Biomedical Research) into the EcoRI and HpaI sites of the pHIT-2 retroviral backbone (Dr. John Olsen, University of North Carolina). cDNA's encoding an alanine to valine substitution at amino acid 143 or a histidine to glutamine substitution at amino acid 179 in p53 (p53-A143V and p53-H179Q respectively) were provided by Drs. David Wynford-Thomas (University of Aberdeen) and Dr. Howard Liber (Massachusetts General Hospital), respectively. Retroviral expression vectors containing these dominant-negative forms of p53 were constructed by cloning the cDNA into the EcoRI site of the pLXIN (Clonetech) retroviral expression vector. The pLXSN-E6 retroviral expression vector containing the HPV16 E6 oncoprotein DNA was a gift from Dr. Denise Galloway (Fred Hutchinson Cancer Center). Vesicular stomatitis virus glycoprotein G-pseudotyped, replication-defective retroviruses were produced as previously described following transient transfection of viral vector and helper plasmids into HEK 293T cells [18-20]. Transfections of plasmids for virus production were done using Superfect™ or Polyfect™ (Qiagen) according to the manufacturer's protocol.
Figure 8 Expression constructs used in this study. CMV/ie Pro/Enh, Cytomegalovirus immediate early promoter/enhancer sequence; VSV-G, vesicular stomatitis virus glycoprotein G; SV40 Pro, simian virus 40 promoter/origin sequence; IRES, internal ribosome entry sequence; MuLV LTR, murine leukemia virus long terminal repeat; hTERT, human telomerase catalytic subunit. A) Plasmid pCI VSV-G expressing VSV-G used to pseudotype replication-defective retrovirus particles. B) Helper plasmid pCI GPZ for packaging replication-defective retrovirus particles. C) pDSWK-8 plasmid vector used to package telomerase cDNA. D) pLXSN-E6 plasmid vector for packaging HPV16 E6. E) pLXIN+p53-A143V and pLXIN+p53-H179Q plasmid vectors for packaging dominant-negative p53 alleles.
Cell Culture
A normal human fibroblast strain designated NHF1 was derived from neonatal foreskin as previously described [21]. All cell culture, including retroviral production, was performed in a humidified, water-jacketed incubator at 37°C with a 5 % CO2 atmosphere. NHF1 cells and all cell lines derived from the parental NHF1 secondary culture were maintained in MEM (Gibco Invitrogen Corp.) supplemented with 10 % defined fetal bovine serum (Hyclone), 2 mM L-glutamine (Gibco Invitrogen Corp.), and 100 μM non-essential amino acids (Gibco Invitrogen Corp.). HEK 293T cells were maintained in DMEM-H (Gibco Invitrogen Corp.) supplemented with 10 % defined fetal bovine serum (Hyclone), 2 mM L-glutamine (Gibco Invitrogen Corp.), 100 μM non-essential amino acids (Gibco Invitrogen Corp.), and 20 mM HEPES pH 7.3 (Sigma Chemical Co.). Transductions were carried out according to standard methods as described previously [22]. Cells at passages 5 or 6 were simultaneously transduced with both the hTERT-expressing virus and one of the viruses disrupting p53 function and/or empty vectors to derive the cell lines listed in Table 1. At the time of transduction, the NHF1 cells were estimated to have undergone 15 – 20 population doublings in vitro. Transductants were selected by 2 weeks growth in media containing 300 ng/ml puromycin (Sigma Chemical Co.) plus 200 μg/ml of active G418 (Gibco Invitrogen Corp.) and, following this initial selection, lines were maintained without antibiotics. Cells were seeded each passage at a density of 5300 – 5500 cells per cm2. The population doubling level (PDL) of the culture was defined as the sum of the population doublings (PD) of each passage. The PD of each passage was determined using the following equation:
Table 1 Status of p53 in Cell Lines
Cell Line hTERT p53 Protein
F1-hTERT+LXIN + WT1
F1-hTERT+p53-A143V + WT/DN2
F1-hTERT+p53-H179Q + WT/DN
F1-hTERT+E6 + -
F1-HIT+LXIN - WT
F1-HIT+p53-A143V - WT/DN
F1-HIT+p53-H179Q - WT/DN
F1-HIT+E6 - -
1. WT: Wild Type
2. DN: Dominant Negative
Cell lines were monitored for mycoplasma contamination using the Gen-Probe kit (Gen Probe Inc. San Diego CA) according the manufacturer's instructions. By this method the cell lines remained free of mycoplasma for the duration of the study.
Cell Cycle Checkpoint Analysis
DNA damage checkpoint responses were assessed following exposure to 1.5 Gy of IR from a 137Cs source (GammaCell 40, MDS Nordion, Canada) at a dose-rate of 86 rads per minute. G1 checkpoint function was assessed by measuring 5-bromo-2'-deoxy-uridine (BrdU, Sigma Chemical Co.) incorporation from six to eight hours following exposure to 1.5 Gy as previously described [23-25]. Flow cytometric and microscopic determination of mitotic indices were shown to yield equivalent results [26,27]. DNA damage G2 checkpoint function was assessed by determining the mitotic index of cultures two hours following irradiation. Mitotic index was determined using flow cytometry to measure the number of cells expressing the phospho-histone H3 mitotic epitope or by directly counting Giemsa- or DAPI-stained mitotic figures as previously described [28-30] Spindle damage checkpoint function was assessed by seeding cells into medium containing 100 ng/ml colcemid (Sigma Chemical Co.) for 24 or 48 hours. The cells were labeled with BrdU during the last two hours of this incubation. Cells were then analyzed by flow cytometry as described above to determine the percentage of cells with >4n DNA content.
Chromosomal Analyses
Metaphase spreads were prepared from the 8 cell lines listed in Table 1 at the earliest possible PDL following selection and then every 10 – 20 PD thereafter until telomerase-negative cells reached crisis. All metaphase preparations were done according to previously described methods [6]. Fifty metaphases from each cell line at each PDL were analyzed and scored for number of chromosomes, and the numbers and types of structural abnormalities. G-banding was done according to standard protocols [31] and representative karyotypes were assembled after analysis of 20 to 25 metaphases.
Western Blot Analysis
Logarithmically growing cells were seeded at 5 × 105 per 100-mm dish and incubated for 48 hr. Cultures were irradiated as described above and incubated for 6 hr at 37°C. Cells were harvested by trypsinization, washed once in PBS, and resuspended in lysis buffer (100 mM sodium phosphate buffer, pH 7.2, 10 mM EDTA, 10 mM EGTA, 1.5 M NaCl, 10% NP40, supplemented with 10 mM 4-(2-aminoethyl) benzenesulfonyl fluoride (AEBSF, Sigma Chemical Co.), 10 mM β-glycerophosphate (Sigma Chemical Co.), 10 mM sodium orthovanadate (Sigma Chemical Co.), and 10 ug/ml of leupeptin (Sigma Chemical Co.) and aprotinin Sigma Chemical Co.). Protein concentrations were determined using the Bio-Rad DC Protein Assay (Bio-Rad Laboratories) according to the manufacturer's protocol. Samples containing 100 μg protein were mixed with an equal volume of 2 × Laemmli sample buffer (125 mM Tris-HCl, pH 6.8, 4% SDS, 20% glycerol) containing 5% β-mercaptoethanol (Sigma Chemical Co.), boiled, and separated by SDS-PAGE. Proteins were transferred to nitrocellulose and probed with antibody against p21Waf1(Neomarkers) and detected with goat anti-rabbit HRP using the ECL substrate (both Amersham Pharmacia Biotech).
Results
Validation of Cell Lines
This study utilized eight isogenic cell lines differing only in expression of telomerase and p53 function as listed in Table 1. Following selection the cells transduced with DSWK-8 (F1-hTERT lines) were assayed for telomerase expression by TRAP assay [32-34]. As shown in Figure 1, all cell lines transduced with DSWK-8 were telomerase-positive. Cell lines transduced with the empty telomerase vector (HIT) were telomerase-negative (data not shown). Western immunoblot analysis confirmed there was significant overexpression of p53 in lines expressing the p53-V143A and p53-H179Q alleles, and no detectable p53 in lines expressing HPV16E6 (not shown).
Figure 1 Assessment of telomerase activity in cell lines transduced with DSWK-8 by TRAP assay. Each of these cell lines exhibits an RNase-sensitive PCR product.
The ability of the cells to delay entry into S-phase following exposure to 1.5 Gy of ionizing radiation (IR) was assessed as a quantitative index of p53-dependent G1 checkpoint function (Figure 2). The F1-HIT+LXIN and F1-hTERT+LXIN cell lines that have an intact p53 signaling pathway displayed an effective G1 checkpoint response to DNA damage. In this line the percent of cells in the first half of S phase 6 – 8 h after irradiation was reduced by >75% due to a G1 arrest. Cells transduced with HPV16E6 and p53-H179Q exhibited severely attenuated G1 checkpoint function. Less than 15% of HPV16E6-expressing cells were delayed in G1 while cells expressing p53-H179Q had <25% arrested in G1 post-irradiation. Cell lines expressing the p53-A143V dominant-negative form of p53 retained approximately half of the normal G1 checkpoint response with about 50% of irradiated cells delayed in G1. Expression of telomerase had no effect on the radiation-induced G1 arrest.
Figure 2 G1 checkpoint analysis of cell lines. Cells at population doubling level 25 – 30 were tested for G1 checkpoint function. Incorporation of BrdU was analyzed 6 – 8 hours after exposure to 1.5 Gy. The radiation-induced reduction in the percentage of cells in the first half of S-phase was determined as a quantitative measure of G1 checkpoint function.
Immunoblot analysis of p21Waf1 expression confirmed the biological analysis of the G1 checkpoint (Figure 3). Cell lines transduced with the empty LXIN vector expressed p21Waf1 in sham-treated controls and expression was induced after treatment with IR. Lines expressing HPV16E6 and p53-H179Q, which displayed severe attenuation of G1 checkpoint function, did not express p21Waf1 in sham-treated controls nor after irradiation. Lines expressing p53-A143V did not display full ablation of expression or induction of p21Waf1 as evident by the low level of expression in sham-treated controls and some induction of protein after irradiation. As was the case for radiation-induced G1 arrest, expression of telomerase did not affect the expression or induction of p21Waf1. Expression of HPV16E6 and p53-H179Q ablated expression of p21Waf1 and induced a severe attenuation of G1 checkpoint function, while expression of p53-A143V attenuated expression of p21Waf1 while reducing G1 checkpoint function by about 50%. Thus the p53-A143V lines displayed only a partial loss of p53 function.
Figure 3 Assessment of p53-dependent induction of p21Waf1 following IR. Western immuno blot to ascertain p21Waf1 induction 6 hr following 1.5 Gy IR was done as described in Methods.
In contrast to their differing responses in the DNA damage G1 checkpoint the non-telomerized F1-HIT+p53-A143V line behaved like the non-telomerized F1-HIT+HPV16E6 and F1-HIT+p53-H179Q lines and bypassed the replicative senescence checkpoint during in vitro aging. Cell population expansion was monitored continuously and all of the telomerase-negative lines initially displayed equivalent growth in vitro (Figure 4A). Cells expressing HPV16E6 or the dominant-negative alleles of p53 continued to grow for 15 – 20 population doublings beyond the 60 PDL at which F1-HIT+LXIN senesced and arrested growth (Figure 4A). After PDL 78 cell death exceeded cell birth in the telomerase-negative, E6-expressing culture, and the culture died by what is classically known as telomere crisis [35]. Although population doublings did not increase beyond PDL 80 – 85 in the p53-A143V and p53-H179Q lines for a period of about 18 weeks and the cells appeared to be in crisis, viable cells nevertheless remained on dishes. After 36 weeks in culture, population expansion resumed and two immortal lines were recovered. The behavior of the cell lines was similar to that detailed previously [25] with p53-effective, telomerase-negative fibroblasts undergoing replicative senescence after 60 population doublings, and the p53-defective, telomerase-negative lines bypassing senescence and then undergoing telomere crisis.
Figure 4 Growth curves of transduced fibroblasts. The x-axis is the number of weeks of continuous culture. The y-axis represents the number of population doublings the cultures had accumulated. A. (◆) F1-HIT+LXIN; (■) F1-HIT+p53-A143V; (▲) F1-HIT+p53-H179Q; F1-HIT+E6. The empty vector control (●) F1-HIT+LXIN cell line underwent senescence at PDL 61 while the cells with defective p53 function continued to divide for an additional 10 – 20 population doublings. At this point cell death equaled cell division, resulting in no net gain of cell number over time. During the approximate 16-week duration of this phase the E6-expressing cells died. B. Continuous growth of hTERT-expressing lines. (◆) F1-hTERT+LXIN; (■) F1-hTERT+p53-A143V; (●) F1-hTERT+p53-H179Q; F1-hTERT+E6.
Population expansion in the lines transduced directly with hTERT was continuous and equivalent to that seen in telomerase-negative lines at PDL 40 – 60 and in the spontaneously immortalized lines at PDL 90 – 110 (Figure 4B). The hTERT-expressing lines were carried to PDL >175 without reduction in growth rate. P53-dependent G1 checkpoint function was monitored during the various phases of cellular aging in vitro and found not to vary substantially (Table 2). Additionally the spontaneously immortalized cell line expressing the p53-A143V dominant negative p53 allele was still able to induce a small amount of p21Waf1 following exposure to 1.5 Gy (Figure 5)
Table 2 G1 Checkpoint Function of Aging Cell Lines
% of Cells Exhibiting a G1 Delay
PDL1 F1-hTERT+LXIN F1-hTERT+p53-A143V F1-hTERT+p53-H179Q F1-hTERT+E6 F1-HIT+LXIN F1-HIT+p53-A143V F1-HIT+p53-H179Q F1-HIT+E6
30 – 35 100 51 11 0 77 56 22 12
40 – 50 94 44 20 6 95 55 29 0
60 – 65 92 65 5 7 85 54 10 5
75 – 85 80 57 17 1
>1002 97 66 0 0 373 03
Avg.4 93 ± 8 57 ± 9 11 ± 8 3 ± 3 86 ± 9 55 ± 1 20 ± 10 6 ± 6
1. Population Doubling Level
2. PDL of hTERT expressing lines 190 – 210
3. Spontaneous Immortalized Derivative
4. Average ± standard deviation. Average for HIT lines excludes spontaneous immortals.
Figure 5 p53-A143V-expressing cells are able to partially induce p21Waf1. Western-immuno blot to demonstrating p21Waf1 induction in p53-A143V-expressing cells. Log-phase cultures were irradiated with 1.5 Gy. Six hours post-irradiation, the cells were harvested and lysed in loading buffer.
A previous report indicated that some dominant-negative p53 alleles induced a gain of function [36]. This gain of function was identified using a "spindle damage" assay that measures the ability of cells to become polyploid when incubated with microtubule poisons such as colcemid. This phenomenon was examined in the four-telomerized cell lines derived for this study. Following 24 or 48 hours incubation in 100 ng/ml colcemid, cells were labeled for two hours with BrdU and then analyzed by flow cytometry to determine the rate of DNA synthesis in diploid and tetraploid nuclei. As shown in Figure 6, all cell lines with defective p53 signaling underwent endoreduplication and displayed increased frequencies of tetraploid S-phase cells when incubated in colcemid. The isogenic F1-hTERT+LXIN line with effective p53-dependent G1 checkpoint function did not display this endoreduplication when incubated with colcemid. Thus, inactivation of p53 expression with HPV16E6 oncoprotein induced the same susceptibility to endoreduplication during incubation in colcemid as was seen using dominant-negative mutant p53 alleles to disrupt p53 signaling.
Figure 6 Assessment of spindle damage checkpoint function in p53-defective fibroblast lines. Endoreduplication was assessed by flow cytometric analysis of BrdU-labeled cells 24 and 48 h after addition of colcemid to culture medium.
Chromosomal Instability
An assessment of chromosomal integrity was done on all eight cell lines within five population doublings of gene transduction to assess the background level of structural and numerical abnormalities in the population and then at PDL 60 (normal replicative senescence point), and PDL 75 – 85 (crisis). Table 3 demonstrates that there were no differences among the various cell lines at the first PDL examined. As was previously observed upon transduction of telomerase-negative fibroblasts with HPV16E6 [6], soon after inactivation of p53 with the dominant-negative alleles, chromosomal number and structure appeared normal. However, as the lines aged and approached the normal replicative senescence point of 60 population doublings, the number of metaphases exhibiting structural and numerical abnormalities increased dramatically in the telomerase-negative cell lines with defective p53 function. At PDL 76 – 77, the majority of metaphases derived from the telomerase-negative HPV16E6-, p53-A143V-, and p53-H179Q-expressing cells exhibited hypodiploidy and/or dicentric chromosomes. The p53-defective lines that were transduced with hTERT to express telomerase did not display these aging-related instabilities in chromosome numbers and structure.
Table 3 Quantification of Structural and Numerical Chromosomal Abnormalities in Cell Lines
% of Metaphases Containing:
Cell Line PDL1 # Chromosomes
≤ 44 45 – 47 48 – 85 86 – 99 ≥ 100 Dicentrics + Rings TA2 Breaks Fragments Other3
F1-hTERT+LXIN 32 9 85 0 7 0 0 0 2 0 0
60 2 93 2 3 0 0 2 2 0 0
76 7 93 0 0 0 0 0 0 0 0
F1-hTERT+p53-A143V 31 5 86 3 5 2 0 3 0 0 0
66 0 90 4 4 2 2 0 2 0 0
85 0 96 2 2 0 0 5 4 0 0
F1-hTERT+p53-H179Q 33 10 88 0 2 0 0 0 0 0 0
63 2 87 2 9 0 0 0 2 2 0
84 2 98 0 0 0 0 0 0 0 0
F1-hTERT+E6 30 0 86 4 11 0 0 0 0 2 0
63 2 93 4 2 0 4 0 0 0 0
84 2 94 2 0 0 2 0 2 0 0
F1-HIT+LXIN4 31 10 85 2 2 2 0 0 0 0 0
59 2 97 0 2 0 5 0 0 0 0
F1-HIT+p53-A143V 32 8 89 0 4 0 0 0 4 0 0
65 27 48 10 10 5 53 2 3 5 7
76 44 32 7 11 4 22 2 2 20 0
F1-HIT+p53-H179Q 33 5 91 2 0 0 0 0 0 2 2
63 7 76 2 15 0 37 4 2 7 2
77 23 62 4 9 0 38 2 2 8 2
F1-HIT+E6 32 7 79 9 4 2 5 2 2 2 2
63 33 60 2 5 0 43 0 3 7 0
77 66 24 0 5 0 39 10 0 24 0
1. Population Doubling Level
2. Telomere Associations
3. Exchange Aberrations + Deletions
4. These cells senesced at 60 population doublings
Attenuation of DNA damage G2 checkpoint function
Previous studies from this laboratory have demonstrated that DNA damage G2 checkpoint function becomes attenuated in congruence with chromosomal instability [6,25,29]. Figure 7 depicts DNA damage G2 checkpoint function in the cell lines at various in vitro PDLs. The F1-HIT+LXIN and F1-hTERT+LXIN lines displayed a typically effective G2 checkpoint response with on average >95% of G2 cells being delayed in their entry to mitosis after treatment with 1.5 Gy. Expression of the dominant-negative p53 alleles and HPV16E6 induced a modest attenuation of G2 checkpoint function measured at PDL 30 – 40 with 7 – 24% of p53-defective cells evading radiation-induced G2 delay. The telomerase-negative lines expressing p53-A143V and HPV16E6 displayed further severe attenuation of G2 checkpoint function with aging in vitro. For these cells at PDL 70 – 80, the mitotic index in irradiated cells was about half of that seen in sham-treated controls. The telomerase-negative, p53-H179Q line displayed a more modest decrement of G2 checkpoint function during aging with at most 17% of cells evading G2 delay. This represents the first example in a total of seven independent analyses of p53-defective human fibroblasts [6,25,37] in which G2 checkpoint function did not appear to be severely attenuated in cells at crisis. Interestingly the G2 checkpoint response of the two spontaneously derived immortal lines (F1-HIT+p53-A143V Immortal and F1-HIT+p53-H179Q Immortal) was very similar to the younger (precrisis) parental lines (Figure 7).
Figure 7 DNA damage G2 checkpoint function in aging fibroblasts. Log-phase cultures at low and high PDL were treated with 1.5 Gy of IR then incubated for 2 h before determination of mitotic index. The percentage of mitotic cells in irradiated cultures was divided by the percentage of mitotic cells in sham-treated controls to determine the percent of cells evading G2 delay [25].
In contrast to the aging-associated attenuation of G2 checkpoint function in the telomerase-negative, p53-defective lines, there was no aging-associated attenuation of G2 checkpoint function in the telomerase-expressing, p53-defective lines. Cells at >150 PDL displayed a response to IR that was nearly equivalent to that seen at PDL 30 – 40. Thus, the age-related attenuation of G2 checkpoint function in p53-defective lines was prevented entirely by expression of telomerase.
Cytogenetics of immortal lines
Spontaneously immortalized cells emerged from crisis after three months in culture (F1-HIT+p53-A143V Immortal and F1-HIT+p53-H179Q Immortal cell lines). These two cell lines had DNA damage G1 and G2 checkpoint responses that were similar to those seen in their low-passage parents (Table 2 and Figure 7, respectively). G-band karyotype analysis of these immortal cell lines revealed that both were near diploid (45 – 46 chromosomes, Table 4). The F1-HIT+p53-H179Q immortal line exhibited few abnormalities (12% 18q+, 8% 18q-) with 92% of metaphases being 45 XY or 46 XY. The F1-HIT+p53-A143V immortal line exhibited a somewhat reduced number of diploid metaphases (72%) and an increased number of polyploid metaphases (20%) as well as a small deletion at 4p16 in greater than 50% of metaphases. The lines that were transduced with hTERT when at low PDL were maintained in culture for greater that 175 population doublings. Up to this point, the telomerase positive, p53-defective lines displayed a normal diploid karyotype with no marker chromosomal aberrations (Table 4). However, the F1-hTERT+LXIN control line exhibited a trisomy for chromosome 8 in 92% of metaphases. Interestingly the F1-hTERT+LXIN cell line had a normal karyotype (25 of 25 metaphases) at an intermediate PDL of 96.
Table 4 Karyotypic Analysis of Cell Lines
Cell Line Chromosome Number1 Karyotype2
F1-hTERT+LXIN (3PDL 173.5) 46, XY (16%) +8 [72%]
47, XY (76%) +8, -Y [8%]
+8, -16
+8, -11
+8, 5q-
F1-hTERT+p53-A143V (PDL 217.6) 46, XY (80%) t(1;4)
12q-
-13
-X
-15
+ marker
F1-hTERT+p53-H179Q (PDL 200.8) 46, XY (84%) +20
12p-
15q-
19q-
-7
+ marker (iso 5p?)
-21
F1-hTERT+E6 (PDL 160.2) 46, XY (80%) -Y
-Y, 11q-
-6, 12p-
-15
+4
9q chromatid break
F1-HIT+p53-A143V Immortal 45, XY (16%) 4p- [24%]
46, XY (56%) 4p-, 4q- [16%]
>90 (20%) two 4p-, two 4q-, dic 13q;13q, four of every chromosome
-Y
6q-
-13, 4p+
-15, -18, 3p-, 2 markers
-15, -18, 3p-
-19, 4p-
4p-, dic 2p;2p
-14, -15, t(5q;14q), t(4p;15q)
-22, 4p-
-20, 10p+
seven 22+
F1-HIT+p53-H179Q Immortal 45, XY (20%) -4, 18q-
46, XY (72%) 4p+, -11, -16, 18q-
18q+ [8%]
16q-
-21, -22, 18q+
-18, 4p+
-15
-17
-18
4q chromatid break
1. Twenty five metaphases analyzed. Number in parenthesis is percentage of metaphases with given number of chromosomes.
2. Karyotype of samples. Number in brackets is percentage of metaphases examined with given karyotype.
3. PDL, population doubling level. PDL of the spontaneously immortalized lines cannot be determined due to the extended period of time the cells spent in crisis.
Discussion
Induced expression of telomerase at the time of inactivation or attenuation of p53 function fully blocked the chromosomal destabilization that is commonly associated with defective p53 in human cells. Thus aneuploidization, polyploidization, and formation of chromosomal aberrations including rings, dicentrics, breaks and exchanges, all of which develop in human cells with inherited or transduced mutant p53 alleles [38,39] or after transduction of viral oncoproteins that inactivate p53 [40-42] appear to be secondary consequences of telomere erosion. Telomerase-expressing fibroblasts with little or no p53-dependent G1 checkpoint function exhibited no detectable numerical or structural chromosomal alterations after being carried through >170 population doublings. This finding has implications for the mechanisms of genetic instability in cancer and its precursors.
In this study eight isogenic cell lines were derived that differed from each other by one or two alleles. All cell lines expressing dominant-negative forms of p53 or the HPV16E6 oncoprotein to interfere with p53 signaling bypassed the senescence checkpoint. However, expression of the p53-A143V allele caused only a 50 % attenuation of DNA damage G1 checkpoint function regardless of telomerase expression. The partial expression of this checkpoint response was associated with partial induction of p21Waf1 protein 6 hr after exposure to 1.5 Gy IR (Figure 3, and Figure 5). This observation is consistent with the report that the V143A allele of p53 binds to and transactivates p53-responsive promoters [43]. When transduced by transfection the A143V allele of p53 has been reported to override the senescence checkpoint and induce severe chromosomal destabilization as seen here [44].
The continued erosion of telomeres in cells bypassing the replicative senescence checkpoint results in a cycle of formation of dicentric chromosomes leading to non-disjunction errors and broken chromosomes that, when repaired, may result in translocations, deletions, or more dicentrics. The telomerase-expressing p53 defective cells used in this study accumulated no structural or numerical abnormalities when grown in culture for greater than 170 population doublings. In contrast to other reports, the telomerase-positive control cell line (F1-hTERT+LXIN) demonstrated no evidence of chromosomal destabilization or morphological changes through approximately 100 PD's [14,45]. However during the next 75 population doublings this cell line did acquire a trisomy for chromosome 8. The trisomy for chromosome 8 seen in the F1-hTERT+LXIN cells at PDL 175 has also been observed in this laboratory in another telomerized human fibroblast line around PDL 200 [46]. Trisomy for chromosome 8 is commonly found in many hyperproliferative disorders and is believed to provide a growth advantage to those cells that acquire it [47-50] Such a growth advantage could have selected for the cell or cells that underwent nondisjunction and acquired the extra chromosome, although population expansion by the F1-hTERT+LXIN line did not increase after PDL 100 (Figure 4B). Recent studies demonstrate that the forced expression of telomerase in skin fibroblasts and other cell types enhances cell proliferation [9]. These effects of telomerase did not appear to alter cellular responses to IR-induced DNA damage as shown here or UV-induced DNA damage as previously reported [51].
Chromosomal destabilization associated with telomere crisis has also been correlated with attenuation and inactivation of DNA damage G2 checkpoint function [6,25,29]. Because this checkpoint blocks mitosis by cells with damaged chromatids, it was hypothesized that chromosomally unstable cells with defects in the G2 checkpoint would have a growth advantage and accumulate in cultures because of selection. An alternative explanation is the severe instability of chromosomes at crisis impeded progression through mitosis, so that upon irradiation damaged cells complete mitosis at a slower than normal rate. Yang et al. described a rad9-dependent checkpoint in S. cerevisiae that responds to the presence of a dicentric chromosome to delay progression through mitosis [52]. If human fibroblasts with dicentric chromosomes are also delayed in their progression through mitosis, this phenomenon could explain the attenuation of G2 checkpoint function seen as p53-defective cells enter crisis. As the efficacy of G2 checkpoint function is quantified by the degree of emptying of the mitotic compartment 2 h after irradiation, a reduced rate of progression through mitosis to G1 will have the effect of sustaining a higher mitotic index in control and irradiated cells. This explanation would view the attenuation of G2 checkpoint function during telomere crisis as another manifestation of chromosomal instability. The correction of G2 checkpoint function in the immortal lines that emerged from crisis with stable chromosomes also supports this explanation. There are other ways to attenuate DNA damage G2 checkpoint function such as inactivation of ATM [53], BRCA1 and 14-3-3ε [54] or overexpression of cyclin B1/Cdk1 kinase [55]. The common attenuation of DNA damage G2 checkpoint function in small cell lung cancer but not squamous cell carcinoma [54] implies that this pathway of DNA damage response protects against carcinogenesis in selected targets.
Human fibroblasts with inherited mutations in p53 or ectopic expression of HPV16E6 spontaneously immortalize at a low frequency [6,38]. During crisis, prior to immortalization, these cell lines exhibit both numerical and structural chromosomal abnormalities. The loss of chromosomes (Table 3) seen in these lines may be due to non-disjunction errors or artifacts associated with the processing of the metaphase preparations. It is not possible to distinguish between these possibilities. However, since this was not observed in the telomerized cell lines or in the non-telomerized cell lines prior to crisis, this phenomenon is likely to be a function of the underlying chromosomal instability. The cell lines, which result from spontaneous immortalization, have usually activated telomerase expression although in some cases telomeres appear to be stabilized by an alternative mechanism (ALT) [56]. These immortal lines are often aneuploid [38]. In this study, the F1-HIT+p53-A143V and F1-HIT+p53-H179Q lines yielded immortal derivatives. These immortal lines were both diploid (45 – 46 chromosomes). The p53-A143V line displayed a marker chromosomal aberration (4p-) in over half of the metaphases. Deletions in this region are associated with Wolf-Hirschhorn syndrome [57]. The p53-H179Q immortal line primarily contained aberrations in chromosome 18 (both losses and gains). Chromosome 18 aberrations have been associated with various malignancies in humans [58,59]. These findings demonstrating only subtle changes in the karyotype following immortalization are consistent with the previous observation that aneuploidy is not an inevitable outcome of in vitro immortalization [60]. Taken together, these results may reflect the fact that all that is required for continuous in vitro growth of human fibroblasts is hTERT expression. Genetic changes that are not cytogenetically detectable apparently are able to derepress hTERT expression and induce immortality.
Authors' contributions
DAS constructed cell lines, carried out checkpoint analysis, participated in karyotype studies, study design, and drafted manuscript. EL carried out the G-banding of chromosomes and fine karyotypic analysis. TPH carried out the Western blot analysis of p53 and p21 induction. WKK was instrumental in study design and manuscript preparation.
Acknowledgements
We are grateful to Dr. Marila Corderio-Stone for helpful comments. This work was supported by PHS grant CA81343 and center grants P30-CA16086 and P30-ES10126. DAS was supported by training grants ES12345 and ES45678. TPH was supported by training grant ES07017.
==== Refs
Houck JC Sharma VK Hayflick L Functional failures of cultured human diploid fibroblasts after continued population doublings Proceedings of the Society for Experimental Biology & Medicine 1971 137 331 333 5581670
Counter CM Avilion AA LeFeuvre CE Stewart NG Greider CW Harley CB Bacchetti S Telomere shortening associated with chromosome instability is arrested in immortal cells which express telomerase activity EMBO Journal 1992 11 1921 1929 1582420
Proctor CJ Kirkwood TB Modelling cellular senescence as a result of telomere state Aging Cell 2003 2 151 157 12882407 10.1046/j.1474-9728.2003.00050.x
Masutomi K Yu EY Khurts S Ben-Porath I Currier JL Metz GB Brooks MW Kaneko S Murakami S DeCaprio JA Weinberg RA Stewart SA Hahn WC Telomerase maintains telomere structure in normal human cells Cell 2003 114 241 253 12887925 10.1016/S0092-8674(03)00550-6
Shay JW Wright WE Hayflick, his limit, and cellular ageing Nature Reviews Molecular Cell Biology 2000 1 72 76 11413492 10.1038/35036093
Filatov L Golubovskaya V Hurt JC Byrd LL Phillips JM Kaufmann WK Chromosomal instability is correlated with telomere erosion and inactivation of G2 checkpoint function in human fibroblasts expressing human papillomavirus type 16 E6 oncoprotein Oncogene 1998 16 1825 1838 9583680 10.1038/sj.onc.1201711
Counter CM Hahn WC Wei W Caddle SD Beijersbergen RL Lansdorp PM Sedivy JM Weinberg RA Dissociation among in vitro telomerase activity, telomere maintenance, and cellular immortalization [see comments] Proceedings of the National Academy of Sciences of the United States of America 1998 95 14723 14728 9843956 10.1073/pnas.95.25.14723
Halvorsen TL Leibowitz G Levine F Telomerase activity is sufficient to allow transformed cells to escape from crisis Molecular & Cellular Biology 1999 19 1864 1870 10022873
Zhu J Wang H Bishop JM Blackburn EH Telomerase extends the lifespan of virus-transformed human cells without net telomere lengthening [see comments] Proceedings of the National Academy of Sciences of the United States of America 1999 96 3723 3728 10097104 10.1073/pnas.96.7.3723
Kaufmann WK Behe CI Golubovskaya VM Byrd LL Albright CD Borchet KM Presnell SC Coleman WB Grisham JW Smith GJ Aberrant cell cycle checkpoint function in transformed hepatocytes and WB-F344 hepatic epithelial stem-like cells Carcinogenesis 2001 22 1257 1269 11470757 10.1093/carcin/22.8.1257
Jiang XR Jimenez G Chang E Frolkis M Kusler B Sage Beeche M Bodnar AG Wahl GM Tlsty TD Chiu CP Telomerase expression in human somatic cells does not induce changes associated with a transformed phenotype Nature Genetics 1999 21 111 114 9916802 10.1038/5056
van Steensel B Smogorzewska A de Lange T TRF2 protects human telomeres from end-to-end fusions Cell 1998 92 401 413 9476899 10.1016/S0092-8674(00)80932-0
Franco S MacKenzie KL Dias S Alvarez S Rafii S Moore MA Clonal variation in phenotype and life span of human embryonic fibroblasts (MRC-5) transduced with the catalytic component of telomerase (hTERT) ExpCell Res 2001 268 14 25 10.1006/excr.2001.5264
MacKenzie KL Franco S May C Sadelain M Moore MA Mass cultured human fibroblasts overexpressing hTERT encounter a growth crisis following an extended period of proliferation ExpCell Res 2000 259 336 350 10.1006/excr.2000.4982
Milyavsky M Shats I Erez N Tang X Senderovich S Meerson A Tabach Y Goldfinger N Ginsberg D Harris CC Rotter V Prolonged culture of telomerase-immortalized human fibroblasts leads to a premalignant phenotype Cancer Res 2003 63 7147 7157 14612508
Noble JR Zhong ZH Neumann AA Melki JR Clark SJ Reddel RR Alterations in the p16(INK4a) and p53 tumor suppressor genes of hTERT-immortalized human fibroblasts Oncogene 2004 14743210
Sambrook J Fritsch EF Maniatis T Molecular Cloning, A Laboratory Manual 1989 2 Cold Spring Harbor Laboratory Press
Comstock KE Watson NF Olsen JC Design of retroviral expression vectors Methods in Molecular Biology 1997 62 207 222 9108523
Johnson LG Mewshaw JP Ni H Friedmann T Boucher RC Olsen JC Effect of host modification and age on airway epithelial gene transfer mediated by a murine leukemia virus-derived vector J Virol 1998 72 8861 8872 9765431
Olsen JC Sechelski J Use of sodium butyrate to enhance production of retroviral vectors expressing CFTR cDNA Human Gene Therapy 1995 6 1195 1202 8527478
Boyer JC Kaufmann WK Brylawski BP Cordeiro-Stone M Defective postreplication repair in xeroderma pigmentosum variant fibroblasts Cancer Research 1990 50 2593 2598 2109654
Olsen JC Johnson LG Wong-Sun ML Moore KL Swanstrom Boucher RC Retrovirus-mediated gene transfer to cystic fibrosis airway epithelial cells: effect of selectable marker sequences on long-term expression Nucleic Acids Research 1993 21 663 669 7680124
Kastan MB Onyekwere O Sidransky D Vogelstein B Craig RW Participation of p53 protein in the cellular response to DNA damage Cancer Res 1991 51 6304 6311 1933891
White RA Terry NH Meistrich ML Calkins DP Improved method for computing potential doubling time from flow cytometric data Cytometry 1990 11 314 317 2318085 10.1002/cyto.990110214
Kaufmann WK Schwartz JL Hurt JC Byrd LL Galloway DA Levedakou E Paules RS Inactivation of G2 checkpoint function and chromosomal destabilization are linked in human fibroblasts expressing human papillomavirus type 16 E6 Cell Growth & Differentiation 1997 8 1105 1114 9342189
Doherty SC McKeown SR McKelvey-Martin V Downes CS Atala A Yoo JJ Simpson DA Kaufmann WK Cell cycle checkpoint function in bladder cancer J Natl Cancer Inst 2003 95 1859 1868 14679155
Kaufmann WK Heffernan TP Beaulieu LM Doherty S Frank AR Zhou Y Bryant MF Zhou T Luche DD Nikolaishvili-Feinberg N Simpson DA Cordeiro-Stone M Caffeine and human DNA metabolism: the magic and the mystery Mutat Res 2003 532 85 102 14643431
Juan G Traganos F James WM Ray JM Roberge M Sauve DM Anderson H Darzynkiewicz Z Histone H3 phosphorylation and expression of cyclins A and B1 measured in individual cells during their progression through G2 and mitosis Cytometry 1998 32 71 77 9627219 10.1002/(SICI)1097-0320(19980601)32:2<71::AID-CYTO1>3.0.CO;2-H
Kaufmann WK Levedakou EN Grady HL Paules RS Stein GH Attenuation of G2 checkpoint function precedes human cell immortalization Cancer Research 1995 55 7 11 7805043
Xu B Kim S Kastan MB Involvement of Brca1 in S-Phase and G2-Phase Checkpoints after Ionizing Irradiation Molecular and Cellular Biology 2001 21 3445 3450 11313470 10.1128/MCB.21.10.3445-3450.2001
Wang HC Fedoroff S Banding in human chromosomes treated with trypsin Nat New Biol 1972 235 52 54 4109406
Falchetti ML Levi A Molinari P Verna R D'Ambrosio E Increased sensitivity and reproducibility of TRAP assay by avoiding direct primers interaction Nucleic Acids Research 1998 26 862 863 9443983 10.1093/nar/26.3.862
Kim NW Wu F Advances in quantification and characterization of telomerase activity by the telomeric repeat amplification protocol (TRAP) Nucleic Acids Research 1997 25 2595 2597 9185569 10.1093/nar/25.13.2595
Krupp G Kuhne K Tamm S Klapper W Heidorn K Rott A Parwaresch R Molecular basis of artifacts in the detection of telomerase activity and a modified primer for a more robust 'TRAP' assay Nucleic Acids Research 1997 25 919 921 9016650 10.1093/nar/25.4.919
Ishikawa F Telomere crisis, the driving force in cancer cell evolution Biochem Biophys Res Commun 1997 230 1 6 9020020 10.1006/bbrc.1996.5928
Gualberto A Aldape K Kozakiewicz K Tlsty TD An oncogenic form of p53 confers a dominant, gain-of-function phenotype that disrupts spindle checkpoint control Proceedings of the National Academy of Sciences of the United States of America 1998 95 5166 5171 9560247 10.1073/pnas.95.9.5166
Kaufmann WK Campbell CB Simpson DA Deming PB Filatov L Galloway DA Zhao XJ Creighton AM Downes CS Degradation of ATM-independent decatenation checkpoint function in human cells is secondary to inactivation of p53 and correlated with chromosomal destabilization Cell Cycle 2002 1 210 219 12429935
Bischoff FZ Yim SO Pathak S Grant G Siciliano MJ Giovanella BC Strong LC Tainsky MA Spontaneous abnormalities in normal fibroblasts from patients with Li- Fraumeni cancer syndrome: aneuploidy and immortalization Cancer Res 1990 50 7979 7984 2253239
Liu PK Kraus E Wu TA Strong LC Tainsky MA Analysis of genomic instability in Li-Fraumeni fibroblasts with germline p53 mutations Oncogene 1996 12 2267 2278 8649766
Schramayr S Caporossi D Mak I Jelinek T Bacchetti S Chromosomal damage induced by human adenovirus type 12 requires expression of the E1B 55-kilodalton viral protein JVirol 1990 64 2090 2095 2325204
Vieten L Belair CD Savelieva L Julicher K Brocker F Bardenheuer W Schutte J Opalka B Reznikoff CA Minimal deletion of 3p13-->14.2 associated with immortalization of human uroepithelial cells Genes ChromosomesCancer 1998 21 39 48 10.1002/(SICI)1098-2264(199801)21:1<39::AID-GCC6>3.0.CO;2-9
Xiong Y Kuppuswamy D Li Y Livanos EM Hixon M White A Beach D Tlsty TD Alteration of cell cycle kinase complexes in human papillomavirus E6- and E7-expressing fibroblasts precedes neoplastic transformation JVirol 1996 70 999 1008 8551641
Zhang W Funk WD Wright WE Shay JW Deisseroth AB Novel DNA binding of p53 mutants and their role in transcriptional activation Oncogene 1993 8 2555 2559 8361764
Liu X Nishitani J McQuirter JL Baluda MA Park NH The temperature sensitive mutant p53-143ala extends in vitro life span, promotes errors in DNA replication and impairs DNA repair in normal human oral keratinocytes Cell Mol Biol (Noisy-le-grand) 2001 47 1169 1178 11838964
Vaziri H Squire JA Pandita TK Bradley G Kuba RM Zhang H Gulyas S Hill RP Nolan GP Benchimol S Analysis of genomic integrity and p53-dependent G1 checkpoint in telomerase-induced extended-life-span human fibroblasts Molecular & Cellular Biology 1999 19 2373 2379 10022923
Cordeiro-Stone M Frank A Bryant M Oguejiofor I Hatch SB McDaniel LD Kaufmann WK DNA damage responses protect xeroderma pigmentosum variant from UVC-induced clastogenesis Carcinogenesis 2002 23 959 965 12082017 10.1093/carcin/23.6.959
Bench AJ Nacheva EP Champion KM Green AR Molecular genetics and cytogenetics of myeloproliferative disorders Baillieres Clin Haematol 1998 11 819 848 10640219
Novak R Dasu S Agamanolis D Herold W Malone J Waterson J Trisomy 8 is a characteristic finding in pleuropulmonary blastoma Pediatr Pathol Lab Med 1997 17 99 103 9050063 10.1080/107710497175056
Sankary S Dickman PS Wiener E Robichaux W Swaney WP Malone PS Gollin SM Consistent numerical chromosome aberrations in congenital fibrosarcoma Cancer Genet Cytogenet 1993 65 152 156 8453601 10.1016/0165-4608(93)90225-B
Sotos JF Genetic disorders associated with overgrowth Clin Pediatr (Phila) 1997 36 39 49 9007346
Heffernan TP Simpson DA Frank AR Heinloth AN Paules RS Cordeiro-Stone M Kaufmann WK An ATR- and Chk1-dependent S checkpoint inhibits replicon initiation following UVC-induced DNA damage Mol Cell Biol 2002 22 8552 8561 12446774 10.1128/MCB.22.24.8552-8561.2002
Yang SS Yeh E Salmon ED Bloom K Identification of a mid-anaphase checkpoint in budding yeast JCell Biol 1997 136 345 354 9015305 10.1083/jcb.136.2.345
Paules RS Levedakou EN Wilson SJ Innes CL Rhodes N Tlsty TD Galloway DA Donehower LA Tainsky MA Kaufmann WK Defective G2 checkpoint function in cells from individuals with familial cancer syndromes Cancer Research 1995 55 1763 1773 7712486
Konishi H Nakagawa T Harano T Mizuno K Saito H Masuda A Matsuda H Osada H Takahashi T Identification of frequent G(2) checkpoint impairment and a homozygous deletion of 14-3-3epsilon at 17p13.3 in small cell lung cancers Cancer Res 2002 62 271 276 11782387
Park M Chae HD Yun J Jung M Kim YS Kim SH Han MH Shin DY Constitutive activation of cyclin B1-associated cdc2 kinase overrides p53-mediated G2-M arrest Cancer Research 2000 60 542 545 10676633
Reddel RR Bryan TM Colgin LM Perrem KT Yeager TR Alternative lengthening of telomeres in human cells RadiatRes 2001 155 194 200
Battaglia A Carey JC Wright TJ Wolf-Hirschhorn (4p-) syndrome Adv Pediatr 2001 48 75 113 11480768
Compton CC Colorectal carcinoma: diagnostic, prognostic, and molecular features Mod Pathol 2003 16 376 388 12692203 10.1097/01.MP.0000062859.46942.93
Oberg K Carcinoid tumors: molecular genetics, tumor biology, and update of diagnosis and treatment Curr Opin Oncol 2002 14 38 45 11790979 10.1097/00001622-200201000-00008
Bunz F Fauth C Speicher MR Dutriaux A Sedivy JM Kinzler KW Vogelstein B Lengauer C Targeted inactivation of p53 in human cells does not result in aneuploidy Cancer Res 2002 62 1129 1133 11861393
|
16202133
|
PMC1262735
|
CC BY
|
2021-01-04 16:38:32
|
no
|
Cardiovasc Ultrasound. 2005 Oct 3; 3:31
|
latin-1
|
Cardiovasc Ultrasound
| 2,005 |
10.1186/1476-7120-3-31
|
oa_comm
|
==== Front
Cardiovasc UltrasoundCardiovascular Ultrasound1476-7120BioMed Central London 1476-7120-3-321620971510.1186/1476-7120-3-32ReviewUltrasound imaging versus morphopathology in cardiovascular diseases. Myocardial cell damage Baroldi Giorgio [email protected] Riccardo [email protected] Lauro [email protected] Institute of Clinical Physiology, National Research Council, Milan and Pisa, Italy2 Cardiology, University School of Medicine and "A. De Gasperis" Foundation, Niguarda Hospital, Milan, Italy3 Cardiovascular Unit, "Campo di Marte" Hospital, Lucca, Italy2005 6 10 2005 3 32 32 14 9 2005 6 10 2005 Copyright © 2005 Baroldi et al; licensee BioMed Central Ltd.2005Baroldi et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
This review article summarizes the results of histopathological and clinical imaging studies to assess myocardial necrosis in humans. Different histopathological features of myocardial cell necrosis are reviewed. In addition, the present role of echocardiographic techniques in assessing irreversible myocardial damage is briefly summarized.
==== Body
By myocardial cell damage we mean a primary damage of the myocardial cell. In fact, the myocardium includes several other structures as vessels (arteries, veins, lymphatics), nerves, collagen matrix, interstitium, which can be primarily altered with subsequent secondary damage of myocardial cells. In general, both clinicians and pathologists believe in a unique pattern of myocardial necrosis due to ischemia; less frequently to inflammatory processes or rarely to storage diseases. In reality, three types of myocardial cell necrosis can be recognized [1-4] in relation to contraction cycle.
The myocardial cell may irreversibly arrest in: 1. relaxation 2. contraction 3. after progressive failure.
1. In the first condition the early histologic pattern is characterized by mild eosinophilia, increased length of sarcomeres and elongation of nuclei. This myocellular stretching is due to the action of intraventricular pressure on these elements in flaccid paralysis and visible within 30 minutes. The lesion is pathognomonic for myocardial infarct with its sequelae, namely a polymorphonuclear leukocytic infiltration which starts after 6–8 hours and disappears within 5 days, centripetal removal of necrotic tissue by macrophages and substitution by collagen ending in acellular and avascular, dense, scar (Fig. 1). No repair by granulation tissue is observed. In humans the infarct is monofocal and its size ranges from less than 10% to more than 50% of the total left ventricular mass. Erroneously named "coagulation necrosis" (coagulation never occurs), is better defined as infarct or ischemic necrosis. In contrast to the current belief oriented to reduce or avoid expansion of an infarct, death due to a myocardial infarct is not related to its size. About half of these cases have a size less than 20% of the left ventricular mass (Tab. 1). The same table show that: a) infarct size is not related to severity of coronary atherosclerotic lumen reduction and number of main vessels with sever stenosis; b) long survival (interval from the beginning to death) prevails in large infarcts; c) extensive myocardial fibrosis, as expression of chronic disease, does not correlate with infarct size; and d) the frequency of an occlusive thrombus is significantly higher in the largest infarcts supporting its secondary formation [5].
Figure 1 Infarct necrosis. The first change is lost of contraction with stretching of the myocardium in flaccid paralysis, resulting in a very early elongation of sarcomeres and nuclei (A) already visible within 30 minutes in experimental infarction. B, polymorphonuclear leukocyte infiltration from the periphery of the infarct after 6–8 hours. In the largest infarcts this infiltration arrests, along a line (maximal myocardial stretching in central part of infarct?) with occasional abscess-like formation (C). This infiltration disappears by lysis of the leukocytes, without evidence of myocellular colliquation or destruction (D). The myocardial cells maintain they sarcomeric registered order even in terminal healing phase. The repair process is carried out by macrophagic digestion (E) – and not by granulation tissue – ending in a compact and dense scar (F).
Table 1 Acute myocardial infarct size (% left ventricular mass) versus coronary atherosclerotic obstruction, extensive myocardial fibrosis (> 20% LVM), occlusive thrombus, survival and death in 200 selected cases*
Infarct size % N. cases Coronary stenosis % Extensive fibrosis Survival days Occlusive thrombus
<69 ≥ 70 1 2 ≥ 3** < 2 3–10 11–30
≤ 20 97 7 90 39 37 14 97 45 26 25 24
> 20 103 10 93 38 34 21 103 26 48 30 58
Total 200 17 183 77 71 35 200 71 74 55 82
*No invasive diagnostic/therapeutic techniques, no fibrinolysis.
** Number of main arteries with severe stenosis (≥ 70)
2. The opposite pattern is seen when the myocytes stop in contraction or better in hypercontraction (Fig. 2). In less than 10 minutes the hypercontracted myocardial cells break down forming hypereosinophilic transverse bands constituted by hypercontracted, extremely short sarcomeres with highly thickened Z lines. This rhexis of the myofibrillar apparatus ends in coagulation of sarcomeres, till a total, granular disruption. The sarcolemmal membrane is preserved and penetrated by macrophages which digest the necrotic material leaving empty sarcolemmal tubes ("alveolar" pattern) which subsequently collagenized (Fig. 3). These changes suggest that the mechanical contraction of the normal surrounding myocardium causes the break of these rigid elements in tetanic paralysis. This lesion is plurifocal with foci formed by one or few cells to thousands and is the typical necrosis obtained experimentally by catecholamine infusion and present in the myocardium of patients with pheochromocytoma. It must be stressed that in these experimental and human conditions no infarct necrosis is seen. Variously called as, microinfarct, infarct-like, focal myocytolysis, Zenker necrosis, coagulative myocytolysis, myofibrillar degeneration, focal myocarditis and overall contraction band necrosis, the more appropriate term is catecholamine necrosis to indicate cause-effect relationship.
Figure 2 Coagulative myocytolysis or contraction band necrosis or catecholamine necrosis (CN). Pancellular lesion involving the whole myocardial cell. A) histological view of a CN focus. B) ultrastructural hypercontraction with extremely short sarcomeres and highly thickened Z lines and focal myofibrillar rhexis. C) rupture of a hypercontracted myocell. EM view of pathological bands (D) formed by segments of hypercontracted and coagulated sarcomeres (intravenous infusion of catecholamines in dogs).
Figure 3 Repair process of CN. A) early monocytes infiltration which later, becomes extensive especially in large necrotic foci (B). It can be misinterpreted as lymphocytic myocarditis. This macrophagic reaction results in empty sarcolemmal tubes with numerous macrophages often loaded with lipofuscin and normal intramural vessels (C). The end result is a focal or plurifocal or confluent fibrosis (D). Microfocal fibrosis as result of necrosis of few myocells (E) can be confused with proliferation of collagen matrix. F) wavyness of normal myofiber around hypercontracted elements. G) all stages of CN in human pheochromocytoma.
The term "contraction bands" needs a more precise definition. Apart from the changeable "physiological bands" in relation to the normal contraction cycle and beside catecholamine necrosis, other "pathological bands" must be considered:
a. Paradiscal bands, part of catecholamine necrosis as a unique band of 10–15 hypercontracted sarcomeres adjacent to an intercalated disc, in an otherwise normal cell. This band does not show any rhexis, macrophagic reaction or other changes and may involve two adjacent myocytes from both sides of the same disc and may appear as a clear or dark band (Fig. 4). Already visible after 5 minutes from catecholamine infusion, the paradiscal bands correspond to the "zonal lesion" described in experimental hemorrhagic shock and prevented by betablocker. It is unclear if this change is a reversible one since in our experimental and human material a reaction of repair process was never seen. The clear band could represent a rebuilding of few damaged sarcomeres in a normally working myocyte.
Figure 4 CN. Paradiscal lesion. Always associated with the pancellular lesion, is already visible in experimental intravenous infusion of catecholamines within 5 minutes (pancellular within 10 minutes). It is formed by a unique band of hypercontraction involving 10–15 sarcomeres adjacent to an intercalated disc. The major part of the myocell is normal and this lesion shows ultrastructurally (A) a clear aspect without rhexis, thin myofibrils and Z lines (rebuilding of normal sarcomeres?) or as a band with different grade of density (B-D), often involving two myocells (C). The dense band can be see histologically (E). An hypercontracted "center" (F) induces wavyness of normal adjacent myocells seen by EM.
b. Cutting edge lesion i.e. a 0.5 millimeter layer of hypercontracted sarcomeres along the cut margin of living myocardium (biopsy, surgical sample, heart excised at transplantation); an artefact not to be confused with catecholamine necrosis (Fig. 5A,B)
Figure 5 Cutting edge lesion which involves a layer of 0.2–0.5 mm along the cut margin of a living myocardium (biopsies, surgical samples, heart excised at transplantation). A) histological aspect in heart excised at transplantation and B) ultrastructural pattern in dog. C) reflow or reperfusion necrosis characterized by CN plus massive interstitial hemorrhage never seen in other human and experimental conditions.
c. Reperfusion injury. From an experimental model of a) temporary coronary occlusion followed by reflow or b) long lasting coronary occlusion, the "wave front phenomenon" has been proposed [6], namely the expansion of a primary infarct established within one hour after occlusion of left circumflex coronary artery and limited to subendocardial layer and posterior papillary muscle [7]. Such an expansion includes the initial infarct with stretched necrotic myocytes, surrounded first by a large zone of "contraction band necrosis" associated with massive hemorrhage and externally by macrophagic reaction and reparative process (Fig. 5C).
This model has been erroneously considered to mimic human infarct. In 200 fatal, acute infarct cases, without any attempt of revascularization, resuscitation and fibronolytic therapy, the ischemic/reperfusion changes were never observed and wavefront expansion was due to nonhemorrhagic catecholamine necrosis, always present both in continuity with the central ischemic necrosis and in normal surrounding myocardium as well as in myocardium not related to the occluded artery. By left circumflex permanent occlusion for 10,18,40 and 60 minutes and temporary occlusion far 10 minutes followed by a 5 minute reperfusion in dog, we tested location and extent (number of foci and necrotic myocytes × 100 mm2) of catecholamine necrosis. The latter was present with a similar extent in ischemic and non ischemic myocardium being independent from amount of flow calculated by radioactive microspheres. Both myonecrosis and frequently associated ventricular fibrillation were prevented by beta-blocker.
For a better understanding of the meaning of catecholamine necrosis in cardiology, its presence and extent were quantified in different conditions (Table 2). The catecholamine myotoxicity was significantly higher in conditions with an adrenergic overtone than in normal controls dead from accident. In the latters with a short survival some damage likely due to an agonal release of interstitial catecholamines (not seen in instantaneous death) was found.
Table 2 Catecholamine myocardial necrosis – Frequency and extent in different conditions
Source Sudden/unexpected death Brain hemorrhage Transplant heart AIDS Congestive heart failure* Cocaine Head trauma Electro cution Carbon monoxide intoxication
Coronary Changes
Number cases 25 34 27 46 38 144 26 45 21 26
no infarct+ survival ≤ 1 day > survival ≤ 1 hour > 1
21 4* 14 13 26 19
Catecholamine necrosis
Present 15 4 34 12 12 39 25 126 11 1 8 1 3
Foci 27 ± 10 29 ± 10 3 ± 1 16 ± 5 37 ± 14 36 ± 9 4 ± 2 2 ± 0.3 4 ± 1 0.5 12 ± 6 8 1 ± 0.5
Myocytes 185 ± 48 1717 ± 698 34 ± 16 26 ± 29 108 ± 134 262 ± 47 13 ± 5 11 ± 2 11 ± 4 35 21 ± 12 46 5 ± 2
Cross band 2 1 8 9 4 14 19 65 11 1 8 1 3
Alveolar 11 2 16 3 6 17 6 25 - - - - -
Healing 2 1 10 - 2 8 - 36 - - - - -
*Silent infarct in apparently normal subjects. **Hearts excised at transplantation.
Foci and myocells quantified × 100 mm2 (standard error).
The conclusion was that catecholamine necrosis is an important signal of adrenergic stress [3,4] particularly in in sudden coronary death, (too often interpreted as synonymous of infarct), in which the unique acute lesion found was catecholamine necrosis in about 80% of cases while in 20% a "silent" infarct associated with catecholamine necrosis was detected. These figures are in agreement with clinical studies in resuscitated people [3].
3. The third damage consists in a disappearance of myofibrils with increasing myocardial cell vacuolization, edema and small mitochondria without any reaction (macrophages, inflammatory elements). This change (colliquative myocytolysis) was seen in about 40% of acute infarct cases, around vessels and in subendocardium in myocardial layers preserved by ischemic necrosis (Fig. 6). Its maximal frequency and extent was in congestive heart failure independently from the underlying disease (Tab. 3). This damage indicates failure of the myocardium, when other rare causes of vacuolization are excluded.
Figure 6 Colliquative myocytolysis associated with acute myocardial infarct. The lesion is confined in layers of the subendocardial myocardium (A) or around functioning vessels (B). These layers are preserved by the infarct necrosis as shown (C) in a perivascular myocardial layer around a vessel in an old infarct without congestive heart failure.
Table 3 Frequency and grade of colliquative myocytolysis in different conditions
Source Sudden/Unexpected death Transplanted Congestive heart failure* Brain AIDS Cocaine Carbon Head Electrocution
Coronary Chagas hearts CHD DCM VPT hemorrhage monoxide trauma
Cases 25 34 46 63 63 18 27 38 26 26 45 21
Colliquative myocytolysis grade
0 19 28 33 1 3 1 26 33 26 26 45 21
1 6 6 11 14 28 2 - 3 - - - -
2 - - 1 39 28 14 1 2 - - - -
3 - - 1 9 4 1 - - - - - -
Grading: 0 = absent, 1 few or smalls group of "empty" myocytes, 2 less and 3 more the 50% of involved myocytes in the inner half of cardiac wall
As a matter of fact, the recognition of different forms of "functional" myonecrosis, which diverge totally in term of structural pathology and molecular/ion biology, denies the assemblage of acute coronary syndromes as a unique etiopathogenetic entity; and helps in interpreting the evolutive phase of each one syndrome as sequence of events and their own causes and mechanisms. For example, a recent consensus [8] included all types of necrosis (coagulation necrosis, contraction band necrosis, apoptosis) measleading our understanding on what a myocardial infarct is
Myocardial Disarray
In discussing the myocardial cause of cardiac arrest, myocardial disarray is another pattern to be considered. It consists of a structural disorganization of the myocardium in which myocytes, instead of their usual parallel arrangement for a correct cardiac pump function, assume a star-like disposition with elements oriented obliquely or perpendicular to each other and joined by short, generally hypertrophic myobridges with interconnecting myofibrils and increased interstitial fibrosis (Fig. 7). This architectonic disorder without evidence of myocellular primary damage is visible in some specific zones of normal hearts at the site of directional change (apex, interventricular septum) of myocardial bundles suggesting "junctional nodes" to help contraction. Furthermore myocardial disarray has been observed around scars, in congenital malformed hearts, lentiginosis, Friedreich's ataxia, Turner's syndrome, hyperthyroidism and overall in hypertrophic cardiomyopathy. Its correlation with the adrenergic system has been suggested by human and experimental data. We studied frequency, extent of myocardial disarray in zones normally uninvolved, in conditions with and without adrenergic hypertone (Table 4). A significant increase in frequency and extent of myocardial disarray was documented in "adrenergic overtone" conditions and it correlated with frequency and extent of catecholamine necrosis. An interesting observation was the absence of myocardial disarray in transplanted hearts of patients dead in the first week after surgery in contrast to its presence in longer survivors [9]. The conclusion was that myocardial disarray, more frequent than originally supposed, may be linked with adrenergic stress and should be diagnosed in time due to its asynergic and arrhythmogenic effect leading to ventricular fibrillation.
Figure 7 Myocardial disarray. Different aspects (A-D) with increased interstitial fibrosis.
Table 4 Frequency and number of sites with myocardial disarray in different conditions
Source Sudden/unexpected death Brain hemorrhage Transplanted hearts Congestive heart failure AIDS Cocaine Carbon monoxide Electtrocution Head trauma
Coronary Chagas <7 7–30 31–365 >365
Cases 25 34 27 9 10 13 14 144 38 26 26 21 45
Disarray sites
0 13 25 12 9 4 6 6 124 34 22 26 21 45
1 2 4 - - 1 - 2 10 - 4 - - -
2 - 4 4 - 1 - - 5 - - - - -
3 2 - 1 - - 1 1 2 2 - - - -
4 1 1 3 - 3 - 1 - 1 - - - -
5 2 - 1 - - 3 - 3 - - - - -
6 - - 1 - - - - - - - - - -
7 2 - - - - 2 1 - - - - - -
8 3 - 5 - - 1 3 - - - - - -
Sites = 8 myocardial samplings from anterior, lateral posterior left and right ventricles and anterior and posterior interventricular septum. Samples were taken from zones which normally do not have focal disarray.
Myocardial Asynergy – Cardiac Arrest
Asynergy or dissinergy means a permanent or temporary, global or zonal contractile dysfunction. It may happen in any condition (coronary heart disease, cardiomyopathies, myocarditis, congenital malformation, etc) with the impression, that, no matter what the underlying disease is, asynergy is linked mainly with the morphofunctional damages previously described. Accordingly, the two apparently opposite patterns of non-functioning but viable myocardium secondary one to chronic ischemia (hibernating myocardium which return to function following revascularization) and the other to reperfusion (stunned myocardium able to refunction after hours, day or weeks) could be explained as a reversible form of relaxed or contracted phase; an assumption derived by experimental permanent coronary occlusion with flaccid myocyte paralysis and catecholamine venous infusion with hypercontraction. Any clinico-pathological correlation is irrealistic since reversibility means no damage of structures which return to function; their temporary blockage is likely at a molecular/ionic level difficult to see histologically and detectable only by immuno-histochemical or more sophisticated techniques at one condition: to sample the dysfunctioning myocardium (serial sections) and discriminate unrelated terminal changes.
In the previous review on coronary collaterals we questioned the existence of chronic ischemia and in the present one we question the existence of reperfusion necrosis in human pathology; suggesting possible alternative etiopathogenetic mechanisms, in which the autonomic nervous system may play an essential role. Agreement exists that in coronary heart disease (CHD), the starting point is zonal hypokinesis (denervation?). May the latter aggravate (akinesis-paradoxical bulging by increased intraventricular pressure) with consequent block by compression of vessels within the non-functioning myocardium, ending in infarct necrosis? Increased contractility by nervous reflexes of surrounding, normal myocardium to compensate the loss of contractility of infarcted myocardium, may result in catecholamine necrosis and ventricular fibrillation (cardiac arrest). In sudden coronary death catecholamine necrosis seems the trigger of ventricular fibrillation. However, in pheochromocytoma in man and in experimental infusion of catecholamines with widespread myocardial lesions ventricular fibrillation does not occur. Only when injected in one coronary artery (unpublished data), noradrenaline produces its typical myocardial necrosis and ventricular fibrillation. The question, therefore, is whether medial neuritis (i.e. lympho-plasmacellular inflammation involving nerves of the tunica in media in CHD) may be the trigger of local noradrenaline release. Similarly in sudden non coronary death in cases with myocarditis associated with catecholamine necrosis (as, for instance, in silent Chagas disease) we should investigate if myocarditis involves intramyocardial innervation. Other possibilities exist in relation to toxic substances or a direct brain/heart relationship with a release in excess of noradrenaline within the myocardium. On the other site, colliquative myocytolysis, not seen in sudden cardiac death, may indicate an acute or subacute or chronic congestive heart failure following an acute infarct or any other cardiac disease.
Target of Ultrasound Diagnosis: Present and Future
Information on composition and structure of myocardial tissue could be of major importance to better characterize the onset and progression of several myocardial diseases in both clinical and research setting. The use of ultrasounds techniques for this purpose is not new, since first applications date back to 40 years ago [10]. Its theoretical background is represented by the fact that ultrasound interacts differently with abnormal as compared to normal myocardium. However, an ideal technique is still far from ready for clinical use. Different methods, using both qualitative and quantitative approaches have been suggested during the last decades.
- Qualitative methods
The direct identification of specific abnormalities by the visual inspection of both M-mode and B-mode echocardiograms is the simplest technique used to study the characteristics of myocardial tissue. An increased intensity of the echocardiographic signal has been reported some weeks following anteroseptal myocardial infarction [10]. The same authors were also able to demonstrate a strong correlation between intensity of the signal and presence of scar tissue on surgical or post-mortem evaluation [11,12]. Color encoded digital processing of images proved to further improve the dynamic range of echocardiographic information [13]. In addition, the simple combination of increased acoustic reflectance and reduced end-diastolic thickness has been shown to represent a simple and reliable predictor of the scarred, asynergic myocardial segments which do not improve in function after revascularization [14]. In particular, it has very recently been confirmed that a diastolic wall thickness of ≤ 0.6 cm on baseline echocardiography can exclude the presence of significant viability with a negative predictive accuracy similar to that of dobutamine stress echocardiography [15].
- Quantitative methods
The biological basis of quantitative methods that have been introduced for the ultrasound tissue characterization is represented by the possibility that individual structural components of the myocardium can influence its acoustic properties in different physiologic and pathologic conditions [16]. These methods are essentially represented by radiofrequency (integrated backscatter) [17,18] and echocardiographic gray level (videodensitometry) analysis [19,20]. In particular, integrated backscatter analyzes the unprocessed radiofrequency signal returning from the myocardium, whilst videodensitometry bases on the conversion of analogic conventional ultrasonic images into a digitized form which allowing quantitative analysis of the ultrasonic myocardial texture. Both techniques [21,22] have been used in experimental or stress-induced myocardial ischemia to detect changes in the ultrasound property of the myocardium. In the setting of acute myocardial infarction, a dramatic reduction of the cyclic variation of integrated backscatter has been demonstrated in the infarct area [23]. Moreover, it was found that myocardial infarcts show an increase in integrated backscatter values and a loss of the cardiac cycle dependent variation in backscatter [24]. This characteristics may be of help in differentiating them from viable tissue [25] that shows preserved cyclic variation of the backscatter signal despite the reduction in wall motion [26]. Differentiation of viable from nonviable tissue has been recently attempted using wavelet transform analysis [27,28], a technique based on breaking up a signal into shifted (translation) and scaled (stretching or compressing) version of a mother wavelet signal [29], to calculate texture energy.
Despite promising preliminary remarks, the pathophysiological background as well as the effective clinical value of ultrasound tissue characterization remain to be defined. In particular, it is expected that more standardized approaches, that will be available in the very near future from digitized technologies, can be of help in allowing comparison of the results from different laboratories.
- Tissue Doppler Imaging
During a prolonged coronary artery occlusion, myocardial necrosis progresses from endocardium toward epicardium as a wave-front phenomenon [30]. Anatomic-pathological studies revealed the great heterogeneity of the reperfused myocardium that contains a variable amount of necrosis surrounded by a viable but transiently stunned epicardium [31]. This structural and functional heterogeneity complicates the interpretation of wall motion abnormalities by conventional echocardiography. Tissue Doppler imaging is a relatively recent ultrasound technique enabling quantification of intramural myocardial velocities by detection of consecutive phase shifts of the ultrasound signal reflected from the contracting myocardium [32]. Main interest of the technique for the myocardial tissue characterization is associated with its ability to differentiate transmural from nontransmural myocardial infarction and thus to assess myocardial viability [33].
Further improvements in both qualitative and quantitative imaging techniques are expected in the near future. This may provide a powerful tool to make information on biochemical composition and physiological state of the myocardial tissue easily available in clinical practice.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Prof. Giorgio Baroldi contributed to the conception and organization of this review and to the final comments.
Dr. Riccardo Bigi and Dr. Lauro Cortigiani summarized the use of ultrasound techniques in atherosclerotic plaque imaging
==== Refs
Baroldi G Different types of myocardial necrosis in coronary heart disease. A pathophysiological review of their functional significance Am Heart J 1975 89 742 746 1130267 10.1016/0002-8703(75)90189-1
Baroldi G Silver MD, Gotlieb AI, Schoen FJ Myocardial cell death including ischemic heart disease and its complications Cardiovascular Pathology 2001 3 Churchill Livingstone, New York 198
Baroldi G Silver MD The etiopathogenesis of coronary heart disease – heretical theory based on morphology Medical Intelligence Unit 2004 2 Eurekah-Com, RG Landes, Georgetown, Texas
Baroldi G Mittleman RE Parolini M Silver MD Fineschi V Myocardial contraction band. Definition, quantification and significance in forensic pathology Int J Legal Med 2001 115 142 151 11775016 10.1007/s004140100229
Baroldi G Bigi R Cortigiani L Ultrasound imaging versus morphopathology in cardiovascular diseases: coronary collateral and myocardial ischemia Cardiovasc Ultrasound 2005 3 6 doi:10.1186/1476-7120-3-6 15740620 10.1186/1476-7120-3-6
Reimer KA Lowe JE Rasmussen MM Jennings RB The wavefront phenomenon of ischemic cell death. 1. Myocardial infarct size vs duration of coronary occlusion in dogs Circulation 1977 56 786 794 912839
Jennings RB Sommers HM Herdson PB Kaltenbach JP Ischemic injury of myocardium Ann N Y Acad Sci 1969 156 61 78 5291154
Alpert JS Thygesen K Antman E Bassand JP Myocardial infarction redefined. A consensus document of the joint European Society of Cardiology/American College of Cardiology Committee for the redefinition of myocardial infarction J Am Coll Cardiol 2000 36 959 969 10987628 10.1016/S0735-1097(00)00804-4
Fineschi F Silver MD Parolini M Baroldi G Myocardial disarray: an architectural disorganization possibly linked with adrenergic stress? Int J Cardiol 2005 99 277 282 15749187 10.1016/j.ijcard.2004.01.022
Skorton DJ Miller JG Wickline A Barzilai B Collins S Perez JE Marcus ML, Shlebert HR, Skorton DJ, Wolf GL Ultrasonic characterization of cardiovascular tissue Cardiac Imaging 1991 Philadelphia, WB Saunders Co 538 556
Rasmussen S Corya BC Feigenbaum H Knoebel SB Detection of myocardial scar tissue by M-mode echocardiography Circulation 1978 57 230 237 618609
Fraker TD Nelson AD Arthur JA Wilkerson RD Altered acoustic reflectance on 2-dimensional echocardiography as an early predictor of myocardial infarct size Am J Cardiol 1984 53 1699 1702 6731315 10.1016/0002-9149(84)90605-2
Parisi AF Nieminen M O'Boyle JE Moynihan PF Khuri SF Kloner RA Folland ED Schoen FJ Enhanced detection of the evolution of tissue after acute myocardial infarction using color encoded two-dimensional echocardiography Circulation 1982 66 764 770 7116594
Faletra F Crivellaro W Pirelli S Parodi O De Chiara F Cipriani M Corno R Pezzano A Value of transthoracic two-dimensional echocardiography in predicting viabilità in patients with healed Q-wave anterior wall myocardial infarction Am J Cardiol 1995 76 1002 1006 7484851 10.1016/S0002-9149(99)80284-7
Zaglavara T Pillay T Karvounis H Haaverstad R Parharidis G Louridas G Kenny A Detection of myocardial viability by dobutamine stress echocardiography: incremental value of diastolic wall thickness measurement Heart 2005 91 613 617 15831644 10.1136/hrt.2003.028316
Kerut EK Given M Giles TD Review of methods for texture analysis of myocardium from echocardiographic images: a means of tissue characterization Echocardiography 2003 20 727 736 14641378 10.1111/j.0742-2822.2003.01126.x
Miller JG Perez JE Sobel BE Ultrasonic characterization of myocardium Prog Cardiovasc Dis 1985 28 85 110 3898225 10.1016/0033-0620(85)90020-9
Lattanzi F Picano E Mazzarisi A Landini L Benassi A Masini M Distante A L'Abbate A In vivo radiofrequency ultrasound analysis of normal human heart structures J Clin Ultrasound 1987 15 371 375 3134434
Skorton DJ Collins SM Nichols J Pandian NG Bean JA Kerber RE Quantitative texture analysis in two-dimensional echocardiography: application to the diagnosis of experimental myocardial contusion Circulation 1983 68 217 223 6851048
Skorton DJ Melton HE Pandian NG Detection of acute myocardial infarction in closed-chest dog by analysis of regional two-dimensional echocardiographic gray-level distributions Circ Res 1983 52 36 44 6848208
Picano E Faletra F Marini C Paterni M Danzi GB Lombardi M Campolo L Gigli G Landini L Pezzano A Increased echodensity of transiently asynergic myocardium in humans: a novel echocardiographic sign of myocardial ischemia J Am Coll Cardiol 1993 21 199 207 8417062
Barzilai B Vered Z Mohr GA Wear KA Courtois M Sobel BE Miller JG Perez JE Myocardial ultrasonic backscatter for characterization of ischemia and reperfusion: relationship to wall motion Ultrasound Med Biol 1990 16 391 398 2396328 10.1016/0301-5629(90)90068-N
Milunski MR Mohr GA Perez JE Vered Z Wear KA Gessler CJ Sobel BE Miller JG Wickline SA Ultrasonic tissue characterization with integrated backscatter. Acute myocardial ischemia, reperfusion, and stunned myocardium in patients Circulation 1989 80 491 503 2766504
Wickline SA Verdonk ED Wong AK Shepard RK Miller JG Structural remodeling of human myocardial tissue after infarction. Quantification with ultrasonic backscatter Circulation 1992 85 259 268 1728457
Marini C Picano E Varga A Marzullo P Pingitore A Paterni M Cyclic variation in myocardial grey level as a marker of viability in man: A videodensitometric study Eur Heart J 1996 17 472 479 8737224
Sagar KB Pelc LR Rhyne TL Komorowski RA Wann LS Warltier DC Role of ultrasonic tissue characterization to distinguish reversible from irreversible myocardial injury J Am Soc Echocardiogr 1990 3 471 477 1703768
Mojsilovic A Popovic MV Neskovic AN Popovic AD Wavelet image extension for analysis and classification of infracted myocardial tissue IEEE Trans Biomed Eng 1997 44 856 866 9282478 10.1109/10.623055
Neskovic AN Mojsilovic A Jovanovic T Vasiljevic J Popovic M Marinkovic J Bojic M Popovic AD Myocardial tissue characterization after acute myocardial infarction with wavelet image decomposition: A novel approach for the detection of myocardial viability in the early post-infarction period Circulation 1998 98 634 641 9715855
Aldroubi A Aldroubi A, Unser M The wavelet transform: A surfing guide Wavelets in Medicine and Biology 1996 New York: CRC Press 3 36
Reimer KA Jennings RB The "wavefront phenomenon" of myocardial ischemic cell death, II: transmural progression of necrosis within the framework of ischemic bed size (myocardium at risk) and collateral flow Lab Invest 1979 6 633 644 449273
Kloner RA Ellis SG Lange R Braunwald E Studies of experimental coronary artery reperfusion: effects of infarct size, myocardial function, biochemistry, ultrastructure, and microvascular damage Circulation 1983 68 I-8 I-15 6861331
Sutherland GR Steward MJ Grounstroem KWE Moran CM Fleming A Guell-Peris FJ Riemersma RA Fenn LN Fox KAA Mc Dicken WN Color Doppler myocardial imaging: a new technique for the assessment of myocardial function J Am Soc Echocardiogr 1994 7 441 458 7986541
Derumeaux G Loufoua J Pontier G Cribier A Ovize M Tissue Doppler Imaging Differentiates Transmural From Nontransmural Acute Myocardial Infarction After Reperfusion Therapy Circulation 2001 103 589 596 11157727
|
16209715
|
PMC1262736
|
CC BY
|
2021-01-04 16:38:31
|
no
|
Cardiovasc Ultrasound. 2005 Oct 6; 3:32
|
utf-8
|
Cardiovasc Ultrasound
| 2,005 |
10.1186/1476-7120-3-32
|
oa_comm
|
==== Front
Cerebrospinal Fluid ResCerebrospinal Fluid Research1743-8454BioMed Central London 1743-8454-2-81619754410.1186/1743-8454-2-8ResearchThe circumventricular organs participate in the immunopathogenesis of experimental autoimmune encephalomyelitis Schulz Martina [email protected] Britta [email protected] Theodor Kocher Institute, University of Bern, CH-3012 Bern, Switzerland2 Kerckhoff Institute, Department of Vascular Biology, Bad Nauheim, Germany3 Max-Planck Institute for Molecular Biomedicine, Münster, Germany2005 30 9 2005 2 8 8 15 7 2005 30 9 2005 Copyright © 2005 Schulz and Engelhardt; licensee BioMed Central Ltd.2005Schulz and Engelhardt; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
During inflammatory conditions of the central nervous system (CNS), such as in multiple sclerosis or in its animal model, experimental autoimmune encephalomyelitis (EAE), immune cells migrate from the blood stream into the CNS parenchyma and into the cerebrospinal fluid (CSF) spaces. The endothelial blood-brain barrier (BBB) has been considered the most obvious entry site for circulating immune cells. Recently, the choroid plexus has been considered as an alternative entry site for circulating lymphocytes into the CSF. The choroid plexus, belongs to the circumventricular organs (CVOs) localized in the walls of the ventricles. Other CVOs, which similar to the choroid plexus lack an endothelial BBB, have not been considered as possible entry sites for immune cells into the CNS parenchyma or the CSF. Here we asked, whether CVOs are involved in the recruitment of inflammatory cells into the brain during EAE.
Methods
We performed an extensive immunohistological study on the area postrema (AP), the subfornical organ (SFO), the organum vasculosum of the lamina terminalis (OVLT) and the median eminence (ME) in frozen brain sections from healthy SJL mice and mice suffering from EAE. Expression of cell adhesion molecules, the presence of leukocyte subpopulations and the detection of major histocompatibility complex antigen expression was compared.
Results
Similar changes were observed for all four CVOs included in this study. During EAE significantly increased numbers of CD45+ leukocytes were detected within the four CVOs investigated, the majority of which stained positive for the macrophage markers F4/80 and Mac-1. The adhesion molecules ICAM-1 and VCAM-1 were upregulated on the fenestrated capillaries within the CVOs. A considerable upregulation of MHC class I throughout the CVOs and positive immunostaining for MHC class II on perivascular cells additionally documented the immune activation of the CVOs during EAE. A significant enrichment of inflammatory infiltrates was observed in close vicinity to the CVOs.
Conclusion
Our data indicate that the CVOs are a site for the entry of immune cells into the CNS and CSF and consequently are involved in the inflammatory process in the CNS during EAE.
==== Body
Background
In multiple sclerosis and in its animal model, experimental autoimmune encephalomyelitis (EAE), inflammatory cells obtain access to the central nervous system (CNS) parenchyma and the cerebrospinal fluid (CSF) and initiate the events leading to signs of paralysis. The endothelial blood-brain barrier (BBB) has been considered the obvious place for entry for circulating lymphocytes into the CNS. Therefore most investigations have focused on defining the molecular mechanisms involved in leukocyte recruitment from the circulating blood across the endothelial BBB. The adhesion molecules, intercellular adhesion molecule 1 (ICAM-1) and vascular cell adhesion molecule 1 (VCAM-1), both members of the immunoglobulin superfamily, are upregulated on the endothelial cells of cerebral vessels during EAE and actively involved in the recruitment of inflammatory cells across the BBB (summarized in [1]).
Trafficking pathways for the entry of immune cells into the CSF remain unknown to date. The CSF of healthy individuals contains between 150,000 cells and 500,000 cells. During multiple sclerosis this number increases dramatically. Neither in the healthy individual nor during multiple sclerosis does the cellular composition of the CSF reflect that of the peripheral blood, suggesting a stringent control for leukocyte entry into the CSF at all times [2]. Recently it was considered that leukocytes enter the CSF using a direct pathway through the choroid plexus. The microvessels within the choroid plexus are different to those in brain parenchyma, the most significant of which is that the endothelial cells allow free movement of molecules via fenestrations and intercellular gaps (reviewed in [3]). Instead, the barrier is located at the level of the choroid plexus epithelial cells, which form tight junctions inhibiting paracellular diffusion of water soluble molecules [4].
Migration of leukocytes through the choroid plexus into the CSF has been suggested by the finding that fluorescently labeled splenocytes are present in the choroid plexus stroma two hours after intravenous injection in mice [5]. The adhesion molecules ICAM-1 and VCAM-1, which are required for leukocyte entry into the CNS, are expressed on the choroid plexus epithelium [6], become upregulated during EAE, and can mediate lymphocyte binding in vitro [7]. These observations suggest that the choroid plexus is involved in the communication of the immune system with the CNS probably by allowing the entry of immune cells directly into the CSF spaces.
Besides the choroid plexus there are additional structures in the CNS of mammals lacking an endothelial BBB. These areas fulfill neurohemal and neurosecretory functions, in that the neurons monitor hormonal stimuli and other substances within the blood or secrete neuroendocrines into the blood, and are commonly referred to as the circumventricular organs (CVOs; reviewed in [8,9]). CVOs are localized at strategic points close to the midline of the brain within the ependymal walls lining the 3rd and the 4th ventricle. Because they lack an endothelial BBB they lie within the blood milieu and thus form a blood-CSF barrier. At the cellular level the barrier between the CVOs and the neuropil is established by specialized epithelial cells called tanycytes. The median eminence (ME) belongs to the purely neuroendocrine CVOs also including the pineal gland and the neurohypophysis. The ME is localized between the stem of the infundibular or pituitary stalk of the hypophysis and the hypothalamus at the base of the 3rd ventricle. The ME contains the terminations of axons from hypothalamic neurons and specialized glial cells. Neuroendocrine functions of the ME include the release of gonadotropin-releasing hormone by which the ME influences reproductive functions [10].
The sensory CVOs such as the subfornical organ (SFO), the organum vasculosum of the lamina terminalis (OVLT) and the area prostrema (AP) are characterized by a high density of small neurons, astrocytes and microglial cells [11]. The SFO and the OVLT are structures in the anterior wall of the 3rd cerebral ventricle. The SFO bulges from the roof of the 3rd ventricle into its lumen at the level of the interventricular foramina, whereas the OVLT is localized in the lamina terminalis between the chiasma opticum and the anterior commissure. The AP can be found at the caudal end of the fossa rhomboidea in the floor of the 4th ventricle. The sensory CVOs monitor the changes in osmotic, ionic and hormonal composition of the blood and are therefore involved in regulating thirst and fluid metabolism [12].
The strategic localization of the CVOs in the wall of the ventricles prompted us to ask whether the CVOs might, similar to the choroid plexus, be involved in the recruitment of immune cells into the brain during EAE. We have performed an extensive immunohistological study localizing adhesion molecules, MHC class I and II antigens and leukocyte subpopulations within the subfornical organ (SFO), the median eminence (ME), the organum vasculosum of the lamina terminalis (OVLT) and the area postrema (AP) of the SJL/N mouse. During EAE similar changes were observed within all four CVOs included in this study. Specifically, the presence of significantly increased numbers of CD45+ leukocytes in the CVOs suggested the recruitment of inflammatory cells into the the parenchyma of the CVOs probably mediated via endothelial ICAM-1 and VCAM-1 which were both upregulated on the microvessels within the CVOs. These changes were accompanied by a significant upregulation of MHC class I throughout the CVOs and induction of MHC class II on perivascular cells. Our observations indicate an active communication between the CVOs and the CNS parenchyma and thus the involvement of the CVOs in the inflammatory process of the CNS during EAE.
Methods
Induction of EAE
Female SJL/N mice were obtained from Bomholdgård Breeding, Ry, Denmark and used for experiments at the age of 10 weeks. Active EAE (aEAE) was induced by immunization with syngeneic spinal cord homogenate as described in detail before [13,14]. Briefly, SJL/N mice were immunized with 100 μg of spinal cord homogenate (SCH) from syngeneic mice in complete Freund's Adjuvant (CFA) (Gibco Laboratories, Grand Island, NY), containing 60 μg/ml Mycobacterium tuberculosis H37Ra and 10 μg/ml Mycobacterium butyricum (Difco Laboratories, Detroit, MI) subcutaneously. Heat-killed Bordetella pertussis organisms (3 × 109; kindly provided by Dr. Kolbe, Behring Werke, Marburg, Germany or Dr. Kohler, Berna Biotech, Bern, Switzerland) were injected in 0.5 ml of PBS on days 1 and 3 post immunization. Animals were checked daily and clinical severity was documented as follows: 0.5 = limp tail; 1 = weak hindlimbs, unsteady gate; 2 = paraplegic, 3 = paraplegic plus incontinent and weakness. Clinical disease occurred approximately 14 days post immunization with spinal cord homogenate. A total of 46 mice with a clinical severity of EAE of 1 to 2 during their first clinical episode are presented in this study. As control, 10 untreated littermates, which were kept in the same cages – were sacrificed and investigated in the same way. All animal experiments were performed in accordance with the German and the Swiss legislation on the protection of animals and approved by the respective government authorities (permission numbers B2/127 and 55/04).
Monoclonal Antibodies
Mec13.3 (anti-mouse PECAM-1/CD31), MK2.7 (anti-mouse VCAM-1), 3C4 (anti-mouse ICAM-2), RB40.34/4 (anti-mouse P-selectin), Lyt 2 (anti-mouse CD8), 145-2C11 (anti-mouse CD3), F4/80 (anti-mouse macrophages), FD441.8 (anti-mouse LFA-1), Mel-14 (anti mouse L-selectin) were purchased form BD Pharmingen, Germany, where detailed information on the antibodies can be obtained . C363 (anti mouse CD3ε) was purchased from Southern Biotechnology Associates, Birmingham, AL, USA, . ER-TR2 (rat- anti mouse MHC class II) and ER-MP42 (rat anti-mouse MHC class I) were purchased from Dianova, Hamburg, Germany, . The hybridomas PS/2 (anti-mouse α4-integrin), M1/9 (anti-mouse CD45), M1/70 (anti-mouse Mac-1), B220 (anti-mouse CD45R), GK1.5 (anti-mouse CD4), MECA-367 and MECA-89 (anti- mouse MAdCAM-1) and Hermes-1 (= 9B5, anti-human CD44, used as an isotype-matched control) were obtained from ATCC (Rockville, MA, USA; ). 25ZC7 (anti-mouse ICAM-1) and 9DB3 (anti-mouse VCAM-1) were kindly provided by D. Vestweber (Münster, Germany; [13]. UZ 4 and UZ 7 (both anti-mouse E-selectin) and MECA 32 were kindly provided by R. Hallmann (Münster, Germany; [14,15]).
Immunohistochemistry
Animals were anesthetized using isoflurene anesthesia (Abbott, Wiesbaden, Germany), and were perfused with either phosphate buffered saline (PBS) or 1% formaldehyde (PFA) in PBS through the left ventricle of the heart. Tissue was removed, embedded in Tissue-tec, (OCT, Miles Inc., Vogel, Giessen, Germany), and snap-frozen in a 2-methylbutane (Merck, Darmstadt, Germany) bath at -80°C. Cryostat sections (6 μm) were air dried overnight, acetone fixed, and stained using a three-step immunoperoxidase technique. Sections were incubated sequentially with primary mAbs, biotinylated secondary goat-anti-rat IgG, (Vector, Boehringer Ingelheim Bioproducts, Heidelberg, Germany), and horseradish peroxidase-conjugated Streptavidin (Vector) for 30 min each step in a humidified chamber, with PBS washes in between the single steps. Sections were developed with 0.07% amino-ethylcarbazol (AEC, Sigma, Germany) and 0.009% hydrogen peroxide in 0.01 M acetate buffer (pH 5.2) for 10 min. Sections were counterstained with Hematoxylin (Gill's formula, Merck), coverslipped with Aquatex (Merck, Darmstadt, Germany) and immediately analyzed.
To detect presence of luminal antigens in cerebral blood vessels mice were injected intravenously with 250 μg primary antibody, anesthetized 15 or 30 minutes later, perfused first with PBS to remove unbound antibody then with 1% PFA to fix bound antibody. Tissue was processed as described above and immunohistology was performed accordingly leaving out the first antibody. This study summarizes results from a total of 49 individual immunostainings where CVOs from 46 EAE mice and 10 healthy mice were compared.
Results
In the mouse the panendothelial antigen MECA-32 becomes specifically downregulated in brain endothelia during the maturation of the blood-brain barrier [15]. As a consequence, the MECA-32 antigen was absent on the mature cerebral endothelium, whereas it remained present on vessels outside of the CNS and the microvessels within the choroid plexus and the circumventricular organs (CVOs, Figures 1, 2, 3, 4, 5, 6, 7, 8).
Figure 1 Immunohistology of the subfornical organ (SFO) in the healthy SJL mouse. The top panel shows MECA-32+ capillaries within the SFO. Panels below demonstrate that ICAM-1 and VCAM-1 can not be detected on the MECA-32+ capillaries within the SFO. Single CD45+ perivascular cells can be detected within the SFO (open arrows, lower panel). Note the positive immunostaining for MECA-32 on choroid plexus capillaires and for ICAM-1 and VCAM-1 on the choroid plexus epithelium visible at the left and right margins (closed arrows). Immunoperoxidase, hematoxylin counterstain. Bar = 100 μm. This result was observed 10 times.
Figure 2 Immunohistology of the area postrema (AP) in the healthy mouse. The V-shaped area of the AP is outlined by the MECA-32+ capillaries (top panel). ICAM-1 and VCAM-1 can not be detected on the MECA-32+ capillaries within the AP. Single CD45+ perivascular cells are present in this CVO (e.g. arrows, lower panel). Immunoperoxidase, hematoxylin counterstain. Bar = 100 μm. This result was observed 10 times.
Figure 3 CD45+ cells in the organum vasculosum of the lamina terminalis (OVLT). Some CD45+ cells, which are not exclusively associated with the MECA-32+ capillaries can be detected within the OVLT of healthy SJL mice (lower two panels; Ø = healthy control). During EAE there is an increased number of CD45+ cells present within the OVLT (upper two panels; EAE). Immunoperoxidase, hematoxylin counterstain. Bar = 100 μm. This result was observed in at least 10 individual stainings.
Figure 4 Immunohistology of the subfornical organ (SFO) during EAE. During EAE, MECA-32+ capillaries within the SFO stain positive for ICAM-1 and VCAM-1 (left panels). A large number of CD45+ cells preferentially localized in close vicinity to the microvessels can be detected within the SFO (second column top panel). Whereas expression of MHC class I is upregulated throughout the SFO parenchyma seen as diffuse staining, MHC class II is solely induced on perivascular cells (second column middle and lower panel). The majority of CD45+ cells within the SFO during EAE are macrophages/microglial cells as determined by positive immunostaining for the β2-integrin Mac-1 and for F4/80 (third column top and middle panel). Note, whereas staining for Mac-1 can be detected on round and ramified cells throughout the SFO, staining for F4/80 remains restricted to perivascular cells. Besides macrophages, a low number of CD3+ T lymphocytes of the CD4 and CD8 (data not shown) subclass and B220+ B cells can be detected (right panels). Most of the immune cells present within the SFO stain positive for LFA-1, but negative for α4-integrins (not shown). Immunoperoxidase, hematoxylin counterstain. Bar = 100 μm. This result was observed in at least 10 individual stainings.
Figure 5 Immunohistology of the area postrema (AP) during EAE. MECA-32+ capillaries within the AP stain positive for ICAM-1 and VCAM-1 during EAE. A large number of CD45+ cells can be detected within the AP, which are associated with the microvessels (e.g. filled arrows) or localized within the parenchyma (e.g. open arrows). The majority of those are macrophages/microglial cells as determined by positive immunostaining for F4/80 and Mac-1. Note, whereas staining for Mac-1 can be detected on cells throughout the AP, staining of F4/80 is only seen on perivascular cells. Immunoperoxidase, hematoxylin counterstain. Bar = 100 μm. This result was observed in at least 10 individual stainings.
Figure 6 Immunohistology of the median eminence (ME) during EAE. MECA-32+ capillaries within the ME stain positive for ICAM-1 and VCAM-1. Positive staining for CD45+ can be detected throughout the ME, staining perivascular cells (examples marked by filled arrow) or parenchymal cells (example marked by open arrow). MHC class I is detectable within the entire ME, with parenchymal localization (closed arrows) as well as vessel associated (open arrow). MHC class II is only present on perivascular cells. Immunoperoxidase, hematoxylin counterstain. Bar = 100 μm. This result was observed in at least 10 individual stainings.
Figure 7 Detection of luminal ICAM-1 and VCAM-1 in the capillaries of the SFO during EAE. Mice suffering from EAE were perfused intravascularly with antibodies directed against MECA-32, ICAM-1 or VCAM-1; an irrelevant rat IgG was perfused as control. Bound antibody was detected by immunohistology. Only a subpopulation of capillaries within the SFO stain positive for luminal ICAM-1 and VCAM-1. Immunoperoxidase, hematoxylin counterstain. Bar = 100 μm. This result was observed 3 times.
Figure 8 Detection of luminal ICAM-1 and VCAM-1 in the capillaries of the ME during EAE. Mice suffering from EAE were perfused intravascularly with antibodies directed against MECA-32, ICAM-1 or VCAM-1; an irrelevant rat IgG was perfused as control. Bound antibody was detected by immunohistology. Only ICAM-1, not VCAM-1, could be detected on the luminal surface of a subpopulation of capillaries in the ME during EAE. Immunoperoxidase, hematoxylin counterstain. Bar = 100 μm. This result was observed 3 times.
In the present study we investigated four CVOs, namely the subfornical organ (SFO), the area postrema (AP), the median eminence (ME) and the organum vasculosum of the lamina terminalis (OVLT), which all lack an endothelial blood-brain barrier (BBB) and demonstrate fenestrated MECA-32+ microvessels.
The circumventricular organs in healthy SJL/N mice
In general, characteristics of the 4 CVOs analyzed were similar. As expected in the healthy SJL/N mice we detected MECA-32+ microvessels within the SFO (Fig. 1), in the AP (Fig. 2) and the ME and OVLT (Table 1). Additionally, CVO microvessels stained positive for PECAM-1 and some of them stained for ICAM-2 (Table 1). Adhesion molecules such as ICAM-1 and VCAM-1 (Figs. 1, 2 and Table 1) or MAdCAM-1, E- and P-selectin (Table 1) could not be detected on the microvessels within the CVOs of healthy mice. A low number of CD45-positive perivascular cells was regularly observed in all CVOs (Figs. 1, 2, 3 and Table 1). In contrast, in only 1 or 2 out of 10 immunostainings performed for each CVO, were individual perivascular cells staining positive for macrophage markers such as F4/80 or Mac-1 or MHC class I or II and LFA-1 detected in CVOs of healthy mice (Table 1). This suggests a variable and rarely detectable presence of individual perivascular macrophages within the CVOs during health. In contrast, parenchymal staining for these molecules was never detected in the CVOs of healthy mice. Similarly, staining for CD3, CD4, CD8 or B220 was always negative in the CVOs of healthy mice (Table 1).
Table 1 Summary of the immunohistochemical characterization of CVOs in health and EAE
Molecule Healthy1 EAE2
Microvessels
MECA-32 + on all vessels + on all vessels
PECAM-1 + on all vessels + on all vessels
ICAM-2 + on some vessels + on some vessels
ICAM-1 No staining + on most vessels luminally and abluminally/perivascular
VCAM-1 No staining + on few vessels luminally and abluminally or perivascular no luminal VCAM-1 detected in ME
MAdCAM-1 No staining no staining
E-selectin no staining no staining
P-selectin no staining no staining
Immune cells
CD45 + on few perivascular cells + on many cells within the CVOs
CD3 no staining + on few round cells ca. 4 cells/section in SFO ca. 10 cells/section in AP
CD4 no staining + on few round cells
CD8 no staining + on few round cells
B220 no staining + on few round cells
F4/80 only rarely observed on single perivascular cells* + on many exclusively perivascular cells
Mac-1 (αM-integrin) only rarely observed on single perivascular cells* + on round and branched cells located perivascularly and throughout the CVOs
LFA-1 (αL-integrin) only rarely observed on single perivascular cells* + on round and branched cells located perivascularly and throughout the CVOs
α4-integrin no staining very faint staining on few round cells – not regularly detected
MHC class I only rarely observed on few perivascular cells§ considerable staining throughout the CVOs; negative on endothelium
MHC class II only rarely observed on single perivascular cells* + on many perivascular cells
1All four CVOs were analysed in a total of 10 healthy mice. Results therefore summarize 10 immunostainings. 2A total of 46 mice has been analyzed. Staining for each marker in each CVO was observed at least 10 times. If not stated differently, stainings were repeatedly observed in all CVOs analysed. *Single immunopositive cells were observed in only 1 or 2 of 10 immunostainings. §Few immunopositive cells were observed in 2 out of 10 immunostainings.
The circumventricular organs during EAE in SJL/N mice
During EAE the changes observed within the four CVOs analysed were overall very similar. The microvessels within the CVOs retained their characteristic staining for the MECA-32 antigen, PECAM-1 and ICAM-2 (Figs. 3, 4, 5, 6, 7, 8, Table 1). In contrast to healthy mice, a high number of CD45+ cells could be detected within the CVOs (Fig. 3, 4, 5, 6), with the most prominent increase of CD45+ cells detected within the OVLT (Fig. 3). CD45+ cells within all four CVOs were localized in close proximity to microvessels and were also found within the CVO parenchyma. Besides the intense CD45 staining on cells of round appearance, a fainter immunostaining for CD45 was detected on cells with a more ramified appearance, indicating that, in addition to the recruitment of CD45+ cells into these areas, CD45 might also be upregulated on the resident microglial cells within the CVOs (Fig. 4, 5, 6).
In order to further characterize the CD45+ cells detected within the CVOs during EAE we performed immunostaining for T cell, B cell, and macrophage markers. A large number of F4/80+ and Mac-1+ macrophages or microglial cells were detected within the CVOs (Figs. 4, 5). Whereas F4/80+ cells were exclusively detected adjacent to CVO microvessels, Mac-1 positive cells could also be found within the parenchyma of the different CVOs (Fig. 4 and 5). Mac-1 positive cells presented either as round cells or cells with many processes, suggesting that Mac-1 was detected on macrophages recruited into the CVOs and also upregulated on resident microglial cells. Besides macrophages, few B220+ B lymphocytes as well as few CD3+ T lymphocytes could regularly be detected within the CVOs during EAE, especially within the SFO (Fig. 4). Most of the CD3 positive T cells recruited into the SFO were CD4+ T helper cells as indicated by the detection of mostly CD4+ but rarely CD8+ cells within this CVO during EAE (Fig. 4 and Table 1).
The adhesion molecules from the integrin family, namely leukocyte function-associated antigen (LFA)-1 (αLβ2-integrin) and α4-integrins can be involved in inflammatory cell recruitment across the BBB during EAE. Therefore we investigated, whether CD45+ cells localized within the CVOs also express these adhesion molecules. Whereas a high number of LFA-1+ cells was detected in the CVOs (Fig. 4), α4-integrin positive cells were detected only rarely (Table 1).
Next we asked whether the potential endothelial ligands for Mac-1, LFA-1 and α4-integrins, namely intercellular adhesion molecule (ICAM)-1 and vascular cell adhesion molecule (VCAM)-1, are expressed on the microvascular endothelial cells within the CVOs. We found prominent induction of ICAM-1 and to a lesser degree of VCAM-1 (Figs. 4, 5, 6), but not of mucosal addressin cell adhesion molecule (MAdCAM)-1, E-selectin or P-selectin (Table 1) on CVO microvessels during EAE. In order to evaluate, whether ICAM-1 and VCAM-1 are available on the luminal surface of the CVOs capillary endothelial cells for the recruitment of circulating Mac-1+, LFA-1+ and α4-integrin+ leukocytes into the CVOs, we injected anti-ICAM-1 or anti-VCAM-1 antibodies into live animals suffering from EAE and investigated their binding to CVO microvessels by immunohistology. Whereas ICAM-1 was detected on the luminal surface of the capillaries of all CVOs investigated (Figs. 7, 8 and data not shown), VCAM-1 although present on the luminal surface of microvessels in the SFO, OVLT and AP (Fig. 7 and data not shown) could not be detected on the luminal surface of the capillaries within the ME (Fig. 8). Taken together, these observations suggest that recruitment of inflammatory cells into the CVOs mainly depend on LFA-1/ICAM-1 or Mac-1/ICAM-1 rather than VCAM-1/α4-integrin-mediated interactions.
In addition to the increased number of immunocompetent cells present in the CVOs during EAE, we observed a significant upregulation of major histocompatibility complex (MHC) antigens class I on cells throughout the entire parenchyma of the CVOs, where expression was absent before in healthy mice (Figs. 4, 6 and Table 1). In contrast, expression of MHC class II was solely induced on perivascular cells, a significant number of which could now regularly be found within all four CVOs during EAE (Fig. 4, 6 and Table 1).
Finally, besides the leukocyte infiltration of the CVOs proper, we observed the preferential localization of massive inflammatory cuffs within the brain parenchyma or the ventricular walls localized in close vicinity or directly adjacent to CVOs, suggesting an involvement of the CVOs in immune cell recruitment into these distinct areas (Fig. 9).
Figure 9 Localization of inflammatory cell infiltrates in the brain of mice suffering from EAE are enriched in close vicinity to CVOs. During EAE, CD45+ cellular infiltrates in the brain were found to be enriched in close vicinity to CVOs. The top panel shows the SFO with perivascular CD45+ cells present within the CVO. Massive CD45+ infiltrates are prominent within the ventricular walls and the brain parenchyma outside but in close vicinity to SFO. The bottom panel shows the AP with perivascular and parenchymal CD45+ cells present within this CVO. Many CD45+ cells have accumulated at the border between the AP and the 4th ventricle. These pictures are representative for the finding that CD45+ inflammatory cell infiltrates in the brain were found to be enriched in number and size in close vicinity to CVOs during EAE. Immunoperoxidase, hematoxylin counterstain. Bar = 100 μm. This result was observed in at least 10 individual stainings.
Discussion
In the present study we provide evidence that immune cell invasion into the CNS during EAE is accompanied by a prominent immune activation of the CVOs. Changes within the CVOs are characterized by the presence of an increased number of CD45+ cells, the considerable induction of MHC class I and MHC class II molecules within the CVOs as well as microglial activation documented by increased expression of Mac-1. Upregulated expression of adhesion molecules on the CVO microvessels suggest their involvement in the recruitment of macrophages, T- and B cells into the CVOs during EAE. Taken together, these observations indicate an active participation of the CVOs in the immunopathogenesis of EAE.
The involvement of CVOs in the communication of the immune system with the nervous system has been considered before. Due to their fenestrated capillaires the CVOs are often referred to as "windows of the brain" and have been thought to serve as entry points for pro-inflammatory cytokines into the CNS [16]. In fact, receptors for inflammatory cytokines and bacterial fragments are constitutively expressed in cells within the sensory CVOs and the release of pro-inflammatory cytokines such as interleukin-1 or TNF-α outside the CNS has been demonstrated to deliver signals into the CNS via the CVOs leading to neuroendocrine responses such as elevations in adrenocorticotropic hormone (ACTH) in the plasma, development of fever and the activation of the hypothalamic-pituitary-adrenal (HPA) axis [17]. Additionally, CVOs play an active role in the development of brain dysfunction during sepsis. Induction of experimental endotoxin shock by injection of lipopolysaccaride induces the sequential expression of molecules involved in innate immune responses such as CD14 or Toll-like receptors first in the CVOs and subsequently in the brain parencyhma (reviewed by [18]). Furthermore, involvement of CVOs in the establishment of an acute phase response during peripheral immune stimuli was demonstrated [19]. Finally, a role for CVOs in sending signals into the brain during systemic inflammation is supported by observations that lesions of CVOs block several components of brain-controlled illness responses (i.e. fever or neuroendocrine modifications) [20].
An active participation of CVOs in the auto-immunopathogenesis in EAE has not been reported to date. In the present study, we found that CVOs in the healthy SJL/N mouse presented a low number of perivascular CD45+ cells, probably macrophages, similar to other regions of the CNS. We only occasionally detected expression of MHC class I or MHC class II on individual perivascular cells suggesting that MHC-positive cells rarely reside within CVOs of healthy mice. Adhesion molecules associated with inflammation could not be detected. In contrast, during clinical EAE at the time when immune cells have penetrated the BBB and perivascular infiltration of T cells and macrophages is observed within the CNS parenchyma, the CVOs demonstrated a dramatic increase in the presence of CD45+ cells. As CD45 is upregulated on activated microglial cells, which reside in high numbers in CVOs [21], the increased number of CD45+ cells detected within the CVOs during EAE is most probably due to both the recruitment of circulating CD45+ immune cells from the periphery into these areas and additionally to the upregulation of CD45 on microglial cells residing within the CVOs.
Most of the CD45+ cells also expressed the macrophage marker F4/80 or Mac-1, which can be expressed on macrophages or activated microglial cells. Based on the immunological and morphological criteria, these cells most probably can be classified as perivascular macrophages or ramified microglial cells, as classical parenchymal microglial cells, as well as scattered rounded macrophage-like cells. Activation of microglial cells by upregulated expression of CD45 and induction of Mac-1 therefore parallels the events observed within the CNS parenchyma protected by the BBB during EAE.
Microglial activation within the CVOs during EAE is further supported by our observation that MHC class I antigens were induced throughout the BBB-deficient parenchyma in all CVOs. In addition, MHC class II molecules could now regularly be detected on perivascular microglial cells or macrophages within the CVOs, which may function as a cellular barrier against blood-borne pathogens. Thus, increased expression of MHC class I and II molecules on microglial cells and perivascular macrophages, which are known characteristic immunopathological changes observed in the CNS parenchyma during EAE, are similarly observed in the BBB-deficient areas of the CVOs [22].
Besides the immune activation of the CVOs, recruitment of a low number of CD3+ T cells and B220+ B cells was detected in the parencyhma of CVOs during EAE, demonstrating that immune cells are invading these areas during EAE. It has been shown that endothelial VCAM-1 and ICAM-1 are involved in lymphocyte recruitment across brain endothelium in vitro and in vivo [13,23-25]. The upregulated expression of ICAM-1 and VCAM-1 on the microvessels of the CVOs during EAE suggests that the same molecules mediating immune cell entry across the BBB might be involved in guiding inflammatory cells into the CVOs. This notion was confirmed by our findings that both molecules are available on the luminal surface of CVO microvessels, with the exception of the median eminence, where luminal VCAM-1 could not be detected. Inflammatory cells within the CVOs were found to stain positive for Mac-1+ and LFA-1+, therefore resemling those, present in the inflammatory cuffs surrounding brain venules forming a barrier [26]. In contrast, whereas during EAE inflammatory cells within the brain express high levels of α4-integrins [26] only few α4-integrin+ leukocytes were found within the CVOs. These observations indicate that inflammatory cell recruitment into the CVOs is not necessarily mediated by the same mechanisms as those regulating leukocyte recruitment across the inflamed BBB, which during EAE was shown to be mainly dependent on α4-integrin and VCAM-1[1,25].
Whether immune cells enter the brain or CSF via the CVOs cannot be determined by this study. It is, however, tempting to speculate that CVOs have an active role in determining inflammatory cell recruitment into the brain as we observed a significantly increased number of inflammatory cuffs and infiltrating cells in close vicinity to the CVOs as compared to regions of the brain with a greater distance to the CVOs. These observations suggest that during EAE the inflamed CVOs deliver cellular or molecular signals into the CNS, which may influence immune cell entry into the CNS and CSF.
Conclusion
Taken together our observations demonstrate that the CVOs are involved in the immunopathogenesis of EAE. Considering their strategic localization within the walls of the ventricles, they may be involved in immune cell entry from the blood into the CNS parenchmyma and also from the blood into the CSF. Therefore, the development of neuroprotective strategies will require consideration of the molecular changes in the CVOs during CNS inflammation.
List of abbreviations
AP area postrema
BBB blood brain barrier
CD cluster of differentiation
CNS central nervous system
CVOs circumventricular organs
EAE experimental autoimmune encephalomyelitis
ICAM-1 intercellular adhesion molecule-1
LFA-1 leukocyte function associated antigen-1
MAdCAM-1 mucosal addressin cell adhesion molecule-1
ME median eminence
MHC major histocompatibility complex
OVLT organum vasculosum of the lamina terminalis
PECAM-1 platelet endothelial cell adhesion molecule -1
SFO subfornical organ
VCAM-1 vascular cell adhesion molecule -1
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
The entire experimental work described in this paper was performed by MS. Evaluation of the data as well as the documentation was performed by both authors. BE designed and supervised the study, and wrote the manuscript. Both authors have read and approved the final manuscript.
Acknowledgements
We owe great thanks to Hiltrud Hölzinger for her expert assistance in photodocumentation. Most of this research was funded by the Max-Planck Society, Germany and the Deutsche Forschungsgemeinschaft (DFG), Germany. Martina Schulz was funded by the DFG.
==== Refs
Engelhardt B Ransohoff R The ins and outs of T cell trafficking to the central nervous system:
Anatomic sites and molecular mechanisms Trends Immunol 2005 16039904
Kivisakk P Mahad DJ Callahan MK Trebst C Tucky B Wei T Wu L Baekkevold ES Lassmann H Staugaitis SM Campbell JJ Ransohoff RM Human cerebrospinal fluid central memory CD4+ T cells: evidence for trafficking through choroid plexus and meninges via P-selectin. Proc Natl Acad Sci U S A 2005 100 8389 94. Epub 2003 Jun 26. 12829791 10.1073/pnas.1433000100
Betz LA Goldstein GW Katzman R Siegel GJ Blood-Brain-Cerebrospinal Fluid Barriers Basic Neurochemistry: Molecular, cellular, and medical aspects 1989 4th Ed. New York, Raven Press 591 606
Wolburg H Wolburg-Buchholz K Liebner S Engelhardt B OSP/claudin-11, claudin-1 and claudin-2 are present in tight junctions of choroid plexus epithelium of the mouse. Neurosci Lett 2001 13 77 80 11427304 10.1016/S0304-3940(01)01927-9
Carrithers MD Visintin I Viret C Janeway CSJ Role of genetic background in P selectin-dependent immune surveillance of the central nervous system. J Neuroimmunol 2002 129 51 57 12161020 10.1016/S0165-5728(02)00172-8
Engelhardt B Wolburg-Buchholz K Wolburg H Involvement of the choroid plexus in central nervous system inflammation Microsc Res Tech 2001 52 112 129 11135454 10.1002/1097-0029(20010101)52:1<112::AID-JEMT13>3.0.CO;2-5
Steffen BJ Breier G Butcher EC Schulz M Engelhardt B ICAM-1, VCAM-1, and MAdCAM-1 are expressed on choroid plexus epithelium but not endothelium and mediate binding of lymphocytes in vitro American Journal of Pathology 1996 148 1819 1838 8669469
Johnson AK Gross PM Sensory circumventricular organs and brain homeostatic pathways. FASEB J 1993 7 678 686 8500693
Leonhardt H Oksche, A. and Vollrath L Ependym und zirkumventrikuläre Organe Handbuch der mikroskopischen Anatomie des Menschen 1980 Berlin, Springr Verlag 177 666
Daftar SS Gore AC IGF-1 in the brain as a regulator of reproductive neuroendocrine function. Exp Biol Med (Maywood) 2005 230 292 306 15855296
McKinley MJ McAllen RM Davern P Giles ME Penschow J Sunn N Uschakov A Oldfield BJ The sensory circumventricular organs of the mammalian brain. Adv Anat Embryol Cell Biol 2003 172:III-XII 1 122
Gebke E Muller AR Jurzak M Gerstberger R Angiotensin II-induced calcium signalling in neurons and astrocytes of rat circumventricular organs. Neuroscience 1998 85 509 520 9622248 10.1016/S0306-4522(97)00601-5
Engelhardt B Laschinger M Schulz M Samulowitz U Vestweber D Hoch G The development of experimental autoimmune encephalomyelitis in the mouse requires alpha4-integrin but not alpha4beta7-integrin Journal of Clinical Investigation 1998 102 2096 2105 9854045
Engelhardt B Vestweber D Hallmann R Schulz M E- and P-selectin are not involved in the recruitment of inflammatory cells across the blood-brain barrier in experimental autoimmune encephalomyelitis Blood 1997 90 4459 4472 9373256
Hallmann R Mayer DN Berg EL Broermann R Butcher EC Novel mouse endothelial cell surface marker is suppressed during differentation of the blood-brain barrier Developmental Dynamics 1995 202 325 332 7626790
Elmquist JK Scammell TE Saper CB Mechanisms of CNS response to systemic immune challenge: the febrile response. Trends Neurosci 1997 20 565 570 9416669 10.1016/S0166-2236(97)01138-7
Lee HY Whiteside MB Herkenham M Area postrema removal abolishes stimulatory effects of intravenous interleukin-1beta on hypothalamic-pituitary-adrenal axis activity and c-fos mRNA in the hypothalamic paraventricular nucleus. Brain Res Bull 1998 46 495 503 9744286 10.1016/S0361-9230(98)00045-8
Sharshar T Hopkinson NS Orlikowski D Annane D Science review: The brain in sepsis--culprit and victim. Crit Care 2005 9 37 44. Epub 2004 Sep 8. 15693982 10.1186/cc2951
Buller KM Role of circumventricular organs in pro-inflammatory cytokine-induced activation of the hypothalamic-pituitary-adrenal axis. Clin Exp Pharmacol Physiol 2001 28 581 589 11458886 10.1046/j.1440-1681.2001.03490.x
Roth J Harre EM Rummel C Gerstberger R Hubschle T Signaling the brain in systemic inflammation: role of sensory circumventricular organs. Front Biosci 2004 9 290 300 14766367
Pedersen EB McNulty JA Castro AJ Fox LM Zimmer J Finsen B Enriched immune-environment of blood-brain barrier deficient areas of normal adult rats. J Neuroimmunol 1997 76 117 131 9184641 10.1016/S0165-5728(97)00038-6
Raivich G Banati R Brain microglia and blood-derived macrophages: molecular profiles and functional roles in multiple sclerosis and animal models of autoimmune demyelinating disease. Brain Res Brain Res Rev 2004 46 261 281 15571769 10.1016/j.brainresrev.2004.06.006
Lyck R Reiss Y Gerwin N Greenwood J Adamson P Engelhardt B T cell interaction with ICAM-1/ICAM-2-double-deficient brain endothelium in vitro: the cytoplasmic tail of endothelial ICAM-1 is necessary for transendothelial migration of T cells. Blood 2003 102 3675 83. Epub 2003 Jul 31. 12893765 10.1182/blood-2003-02-0358
Laschinger M Engelhardt B Interaction of alpha4-integrin with VCAM-1 is involved in adhesion of encephalitogenic T cell blasts to brain endothelium but not in their transendothelial migration in vitro J Neuroimmunol 2000 102 32 43 10626664 10.1016/S0165-5728(99)00156-3
Yednock TA Cannon C Fritz LC Sanchez-Madrid F Steinman L Karin N Prevention of experimental autoimmune encephalomyelitis by antibodies against alpha 4 beta 1 integrin Nature 1992 356 63 66 1538783 10.1038/356063a0
Steffen BJ Butcher EC Engelhardt B Evidence for involvement of ICAM-1 and VCAM-1 in lymphocyte interaction with endothelium in experimental autoimmune encephalomyelitis in the central nervous system in the SJL/J mouse Am J Pathol 1994 145(N1) 189 201 7518194
|
16197544
|
PMC1262737
|
CC BY
|
2021-01-04 16:37:38
|
no
|
Cerebrospinal Fluid Res. 2005 Sep 30; 2:8
|
utf-8
|
Cerebrospinal Fluid Res
| 2,005 |
10.1186/1743-8454-2-8
|
oa_comm
|
==== Front
CytojournalCytoJournal1742-6413BioMed Central London 1742-6413-2-151616475410.1186/1742-6413-2-15ResearchEfficiency of an inexpensive liquid-based cytology performed by cytocentrifugations: a comparative study using the histology as reference standard Garbar Christian [email protected] Corinne [email protected] Véronique [email protected] Department of Pathology, CHU de Charleroi (Université Libre de Bruxelles), 1 Boulevard Zoé Drion, 6000 Charleroi, Belgium2 Laboratory of Molecular Virology, Pasteur Institute, 642 rue Engeland, 1180 Brussels, Belgium2005 15 9 2005 2 15 15 26 7 2005 15 9 2005 Copyright © 2005 Garbar et al; licensee BioMed Central Ltd.2005Garbar et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Although liquid-based cytology (LBC) is now recommended for cervical cancer screening, it requires expensive automated devices and materials. To evaluate the efficiency of inexpensive LBC methods relying on an inexpensive fixative liquid, Easyfix®, we compared the results obtained by the liquid-based cytology (LBC) diagnoses performed by cytocentrifugations (Papspin® and Turbitec®) with those obtained by histology. Furthermore, we evaluated the efficiency of the fixative liquid, Easyfix®, to preserve HPV DNA in the collected samples.
Method
266 LBC were compared with 174 colposcopies and 91 Loop Electrosurgical Excision Procedure (LEEP). Among the LBC, 51 were performed using the Papspin® system and 215 were performed using the Turbitec® system. To control the quality of the preservation liquid, Easyfix®, we correlated the results of HCII assays with those of HPV PCR.
Results
For Papspin® and Turbitec® systems, the sensitivities were respectively 82.6% (95% CI: 61.2–95.0%, p < 0.001) and 75.0% (95% CI: 64.4–89.8%, p < 0.001) and the specificities were 92.6% (95%CI: 76.5–99.1%, p < 0.001) and 96.2% (95% CI: 91.3–98.7%, p < 0.001). We find no statistical difference between the results of the both systems (p = ns). The sensitivity of the HCII was 86.4% (95% IC: 77.4–92.8%, p < 0.001) and the specificity was 39.4% (95% CI: 31.2–48.1%, p < 0.001). The comparison between HCII and HPV-PCR shows a good correlation: the kappa was 0.89.
Conclusion
LBC performed by cytocentrifugations are inexpensive, reduce inadequate smears, show excellent efficiency and allow HPV detection by molecular biology.
==== Body
Background
For more than ten years, liquid-based cytology (LBC) has been developed for cervical cancer screening. Unlike to Conventional cervical Smears (CS), cells are scattered in a fixative liquid to produce a thin layer of cells on slides. The main advantages of this technique are to reduce the number of inadequate smears and to provide enough cells for the detection of infectious agents such as human papillomavirus (HPV) through molecular biology techniques [1-3]. At the moment, the majority of these techniques are using expensive automated devices leading to a significant increase in the price of LBC [4,5].
The LBC performed by cytocentrifugations exist since the 1970s and have been developed for the automated reading of cervical cancer cytology. Preliminary studies show that they significantly reduce the cost of LBC [6-13]. In Central Europe, the LBC performed by cytocentrifugation consist mainly in the Papspin® system (ThermoShandon Inc, Pittsburgh, the USA), in the CytoSCREEN® system (Seroa, Monaco, Monaco) and in the Turbitec® system (Labonord, Templemars, France). These last two techniques use the Hettich cytocentrifuge (Andreas Hettich Corp, Tuttligen, Germany).
The first purpose of this study was to evaluate the performance of two LBC performed by centrifugations: the Papsin® and the Turbitec® systems. In addition, we evaluated the efficiency of the fixative liquid, Easyfix® (Labonord, Templemars, France) to preserve HPV DNA.
Materials and methods
Patients
A total of 268 LBC were collected from 177 women (mean of age: 37.3 +/-10.8 years) from January 2002 to December 2004 in a routine gynecologic setting at the Department of Gynecology (CHU de Charleroi, Belgium). Indications for gynecological consultation were based on previous abnormal conventional cervical smears in a population-based screening. According to the histological diagnosis, the age groups were quite homogeneous, except for the women in the CIN3 group who were slightly younger than those of the other groups (Table 1, p = 0.04).
Table 1 Patients age distribution and samples distribution among the Papspin® and Turbitec® LBC according to the histological diagnosis.
Histology diagnosis Papspin® Age (Years)* Turbitec® Age (Years)*
WNL 12/51 (23.5 %) 45.1 +/- 15.6 86/215 (39.9%) 38.3 +/- 11.9
CIN 1 16/51 (31.3%) 42.6 +/- 8.9 46/215 (21.1%) 37.3 +/- 11,6
CIN 2 4/51 (7.8%) 38.6 +/- 2.8 15/215 (6.9%) 39.6 +/- 13.2
CIN 3 19/51 (37.2%) 33.5 +/- 7.9 70/215 (32.2%) 34.3 +/- 7.8
174 colposcopies associated with biopsy (30 for Papspin® and 144 for Turbitec®) and 92 Loop Electrosurgical Excision Procedure (LEEP) (21 for Papspin® and 71 for Turbitec®) were correlated with all LBC. All the cells were collected using a Cervexbrush® (Rovers, Oss, Nederlands), immersed in 15 ml of a non-buffered alcoholic fixative liquid containing 30% ethanol (Easyfix®, Labonord Corp, Templemars, France). 2 LBC corresponding to a cytological and histological diagnosis of cervical adenocarcinoma were discarded. All the patients were informed of the purpose of this study and agreed voluntarily to take part in it.
Methods
The Easyfix® cell fixative solutions, containing the Cervexbrush® and the cells, were homogenized by mechanical agitation (Vortex®) for at least 30 seconds.
For the Turbitec® technique, 4 drops albumin (StickOn®, Labonord) were put on a polylysined slide placed in the centrifuge chamber (Hettich Centrifuge®). The centrifuge chambers had a final volume of 6 ml; the surface of cells projection on the slide was 240 mm2. The cellularity was estimated using a photoelectric cell analyzer (Labonord) according to the dilution of the Easyfix® sample: 200 μl, 500 μl, 1 ml, 3 ml or 5 ml of Easyfix® solution. This volume was placed in the centrifuge chamber and diluted with an alcoholic fixative liquid containing polyethylene glycol (Cytofix®, Labonord), in order to obtain a finale volume of 6 ml. After 10 minutes centrifugation at 2000 revolutions per minute (rpm), the liquid was discarded.
The Papspin® technique was performed as described previously in the literature with the exception of the use of the Easyfix® and the Cytofix® fixative liquids [14]. The surface of cells projection was 300 mm2. The volume of dilution was respectively 500 μl, 1 ml and 2 ml and the finale volume was 3 ml, according to the turbidity of the sample. The chambers (Megafunnel® – ThermoShandon) were centrifuged 5 minutes at 1250 rpm (CytoSpin® – ThermoShandon) [9-12].
For cytology diagnosis, we used the Bethesda system 2001: within the normal limits (WNL), atypical squamous cells cannot exclude high grade squamous intraepithelial lesion (ASC-H), atypical squamous cells of undetermined significance (ASC-US), low and high grade squamous intraepithelial lesions (Lg-SIL and Hg-SIL) [14]. The histology diagnoses were classified in 4 groups: WNL, Cervical Intraepithelial Neoplasia grade 1 (CIN 1), grade 2 (CIN 2) and grade 3 (CIN3). For the cytology diagnosis and the HCII results, the reference standard was the CIN2 and above (CIN2+). No cytology or histology diagnoses were revised.
225 HPV tests performed by the Hybrid Capture II® (Digene Corp, Beltsville/HCII) were also realized (38 for Papspin® and 187 for Turbitec®). The HCII assays were performed on 5 ml of the residual liquid-based samples. After 5 min. centrifugation at 2000 rpm, the supernatant was discarded. The cellular pellet was washed once with 1 ml PBS, resuspended in 100 μl of Cervical Sample ® (Digene Corp, Beltsville) and denatured in an alkaline solution. Classical hybridization, detection and calibration were made according to the HCII kit's instructions [15,16]. We only used the probes against the high-risk HPV: 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68. The HCII results were expressed as positive or as negative depending on the relative light unit of 1 pg/ml of HPV DNA.
We also performed PCR techniques to detect HPV DNA in 72 samples fixed in Easyfix® solutions for 3.2 +/- 0.9 months. Total cellular DNA were extracted by a freeze/defreeze method as previously described [17] or were purified using the QIAamp blood minikit (Qiagen). The quality of the extracted DNA was evaluated by PCR using beta-globin specific primers, as described [18]. HPV-specific PCR were performed using the general primers, as previously described [19].
Statistics
A Chi 2 or a Fisher exact, ANOVA and kappa of Cohen tests were performed. A p-value less than 0.05 was considered statistically significant. The results are expressed as a mean +/- standard deviation or as a percentage.
We use the Epi Info 2002 revised 2003 (Centers of Disease Control and Prevention – W.H.O, Atlanta, USA) and the Analyse-it® 1.7 (Analyse-it Software, Leeds, England) programs.
Results
The microscopic reading of the LBC, performed by cytocentrifugations, was virtually the same as for other classical LBC techniques, especially the Prepstain® system, although a higher background (cell debris, inflammatory cells, lactobacillus, blood...) was observed but did not influenced the lecture (figures 1 to 4).
Figure 1 Lg-SIL : cellularity is satisfactory (Papanicolaou stain, high magnification – 20x objective, Papspin® system). HCII was positive.
Figure 2 Same case a high magnification : see cytoplasmic and nuclear details (Papanicolaou stain, high magnification – 100x objective, Papspin® system).
Figure 3 Hg-SIL : presence of inflammatory cells which did not influenced the lecture (Papanicolaou stain, high magnification – 20x objective, Turbitec® system). HCII was positive.
Figure 4 Same picture a high magnification : see cytoplasmic and nuclear details (Papanicolaou stain, high magnification – 40x objective, Turbitec® system).
According to Bethesda 2001, only 2 smears (0.9%) were classified as inadequate for cytology interpretation because of too low cellularity. These two smears were performed by Turbitec®. Indicative minor quality criteria were absence of cells junction for 11/51 Papspin® (21.5%) vs 24/215 for Turbitec® (11.1%), low cellularity for 1/51 (1.9%) vs 4/215 (1.8%), hemorrhage for 2/51 (3.8%) vs 6/215 (2.7%), severe inflammation for 1/51 (1.9%) vs 1/215 (0.4%) and severe cytolysis for 2/51 (3.8%) vs 2/215 (0.9%) of LBC. However, we can not draw any statistical conclusion because of the too low number of LBC.
In order to evaluate the performance of the LBC performed by cytocentrifugations from ours clinical samples preserved in the Easyfix® liquid, we compared the results of 2 methods of LBC centrifugations with the histological diagnosis.
According to the histology diagnosis, the samples were homogeneously distributed between the Turbitec® and the Papspin® cytocentrifugation systems (Table 1). There was a slight significant difference between the distribution of the histology diagnosis of the Turbitec® and Papspin® systems only for the group of the WNL (Table 1, p = 0,02).
The comparison between the results obtained by the two cytocentrifugation LBC and those obtained by histology is shown in Table 2. There was no statistical difference with caution of the small number of Papspin® sample. The global tendency was very similar except for ASC-H which are more frequently cited for Turbitec® (1/51 or 1.9% vs 13/215 or 6.0%, p = ns). This is probably due to the fact that Papspin® LBC were performed after Turbitec® LBC when ours cytologists and pathologists were more used to the Bethesda system 2001.
Table 2 Comparison of cytological and histological diagnosis of Papspin® and Turbitec®.
Histological diagnosis Cytological Diagnosis Papspin®
WNL ASC-US ASC-H Lg-SIL Hg-SIL
WNL 5/6 (83.3%) 2/5 (40.0%) - 4/18 (27.8%) -
CIN 1 1/6 (16.7%) 3/5 (60.0%) 1/1 (100%) 9/18 (50.0%) 2/21 (9.5%)
CIN 2 - - - 2/18 (5.5%) 3/21 (14.3%)
CIN 3 - - - 3/18 (16.7%) 16/21 (76.2%)
Turbitec®
WNL 37/42 (88.1%) 17/29 (58.6%) 6/13 (46.1%) 22/63 (34.9%) 2/68 (2.9%)
CIN 1 5/42 (11.9%) 7/29 (24.1%) 1/13 (7.8%) 30/63 (47.6%) 3/68 (4.4%)
CIN 2 - - - 2/63 (3.3%) 13/68 (19.1%)
CIN 3 - 5/29 (17.3%) 6/13 (46.1%) 9/63 (14.3%) 50/68 (73.6%)
38 samples from the Papspin® and 187 from the Turbitec® methods were also analyzed in the HCII assay. The frequency of HPV performed by HCII was respectively for Papspin® and Turbitec® of 50.0% vs 30.0% for WNL, 60.0% and 51.8% for ASC-US, 100% vs 100% for ASC-H, 87.5% vs 71.4% for Lg-SIL and 100% vs 85.0% for Hg-SIL. Those results suggested that HPV DNA was more detected among the ASC-US than among the WNL and that the HPV DNA is more frequently detected for ASC-H, Lg-SIL and Hg-SIL. However, because of the small cases in each group, no conclusion can be made.
The sensitivity and the specificity of Papspin® and Turbitec® methods were calculated using the CIN2+ histology detection as the reference standard for the cytological diagnosis of Hg-SIL. Respectively for Papspin® and Turbitec®, the sensitivity was 82.6% (95%CI: 61.2–95.0%, p < 0.001) and 75.0% (95% CI: 64.4–89.8%, p < 0.001) and the specificity was 92.6% (95%CI: 76.5–99.1%, p < 0.001) and 96.2% (95%CI: 91.3–97.7%, p < 0.001). There was no statistical difference between the 2 LBC systems (p = ns).
The false negatives LBC (with CIN2+ as reference standard) consist of 4 Papspin®, all Lg-SIL and all positive for HPV/HCII (4/4) and of 22 Turbitec®, 11 Lg-SIL (8/11 positive for HCII), 5 ASC-US (4/5 positive for HCII) and 6 ASC-H (6/6 positive for HCII). The false positives LBC consist of 2 for Papspin® and 4 for Turbitec®, all positive for HCII (2/2 and 4/4) and all with a colposcopy with only one small biopsy which is no representative of the whole cervix.
We also evaluated the quality of the Easyfix® fixative liquid to preserve HPV DNA in the residual materials of 72 LBC performed by cytocentrifugations. Even after 3.2+/-0.9 months, at room temperature, beta-globin or HPV DNA could still be amplified by PCR in those samples. The comparison between HPV detection using either general primer PCR or HCII tests showed a Kappa test of 0.89 (p < 0.001). Three cases were positive for the HPV PCR and negative for the HCII and one case was negative for the HPV PCR and positive for the HCII. This last case was also negative for the beta-globin PCR explaining that HPV could not be detected because of the poor quality of the extracted cellular DNA. The stability of DNA in the Easyfix® medium after 3 months was in general excellent.
Discussion
In this study, we demonstrated the feasibility and the efficiency of inexpensive LBC performed by cytocentrifugations: the Papspin® and the Turbitec® systems. We use colposcopy with biopsy or LEEP as reference standard. The diagnosis threshold was set up to the high-grade cervical intraepithelial neoplasia (CIN2+). This could have led to a selection bias of a high-risk patient population. Therefore, ours results are only acceptable in the framework of diagnosis but must be interpreted with some caution in the framework of cervical cancer screening in which the number of healthy patients should usually be as high as possible.
The adequacy of the LBC has been already described in the literature: Weynand et al. (2003) [12] described 0.7% of inadequate samples with the Papspin® system and Bergeron et al. (2003) [13] found 0.14% with the CytoScreen system (which are technically very similar of the Turbitec® system). These authors also demonstrated the superiority of the quality of LBC in comparison with those of CS. With cautions of small number of our samples and using the norms of the Bethesda system 2001, we also found only 0.9% of inadequate cytology [20-24]. Chhieng et al. (2004) also described a similar rate of 0.81% of unsatisfactory sample following the implementation of Bethesda 2001 [20]. Within a comparative study of Thinprep®, Autocyte PREP® and the new manual LBC of Digene (DNACITOLIQ®, Digene), Alves et al. (2004) also concluded that in spite of the different methodologies, the 3 methods adequately preserved cellular structure for morphologic evaluation [21]. Nam et al. (2004) comparing Thinprep® and a manual LBC, called MonoPrep2®, drew similar conclusions [22] All of these authors concluded that manual LBC are cost-effective and provide an alternative method to the currently automated technique of LBC.
Nevertheless, rare are the studies describing the accuracy, in terms of sensibility and specificity, of these manual LBC and particularly for cytocentifugation methods. In our study, to detect the CIN2+, the efficiency of ours LBC, performed by cytocentrifugations, Papspin® and Turbitec®, are quite similar. Indeed, their sensitivity are respectively 82.6 % and 75% (p = ns) and their specificity are respectively 92.6 % and 96.2% (P = ns). The small number of Papspin® LBC in our study is due to the subsequent choice for the Turbitec® method which is more cost-effective. By comparison with published results our LBC efficiency performed by cytocentrifugation seems better than CS. A sensitivity of 68% and the specificity of 79% has been indeed reported for the CS and similarly a sensitivity of 76% and a specificity of 86% was reported by others for the Thinprep® [23-26]. Recently, a sensitivity and a specificity quite similar to ours was also reported for the manual LBC of Digene, respectively 75.3% and 86.4% [27].
One main advantage of LBC is that they allow ancillary techniques such as those used in immunocytochemistry or molecular biology [28-32].
The Easyfix® fixative fluid, used in this study for both LBC, is not yet accredited for the use of HCII. With a Kappa test of 0.89 between DNA/HPV PCR and HCII, and with the comparison with the cytological or histological diagnosis, we can conclude that this liquid is efficient for molecular biology and for the HCII technique. Similar conclusions have recently been reached by Leduc et al.(2004) who compared 250 cervical samples which have been fixated with both Easyfix® and Cervical Sampler liquid® (Digene Corp). They found a Kappa of 0.74 [15]. In our study, the sensitivity of HCII, compared with CIN2+, is 86.4 % and the specificity is 39.4 %. These results are similar to those found by Lee et al. (2004) who used Cervical Sampler liquid® (Digene) and found a sensitivity of 94.2% and a specificity of 52.4%[33]. Howard et al. (2002) had similar results with a 81.8% sensitivity and a 51.5% specificity [34]. In our study, the slight difference of specificities is likely been due to the patients selection with high risk HPV infection. Indeed, 50% (41/81) of these HCII false positives showed a CIN1 at the biopsy. Among the others without histological lesion, only 7/40 had no cytological lesion (WNL).
Actually, it is generally accepted that the HCII test is more sensitive and less specific than cytology to identify Hg or Lg-SIL. However, in the case of ASC-US, a combination of HPV DNA and Papanicolaou smears can certainly increase the sensibility of the cervical cancer screening [35-39]. Likely, a cost-effective LBC associated of HPV DNA will probably have a place to reduce cervical cancer in under-developed countries or small laboratories which cannot invest expensive equipment.
We demonstrated by this study that LBC performed by cytocentrifugation are efficient and also allow the HPV DNA preservation in the Easyfix® as no-buffered alcoholic fixative liquid.
Inexpensive LBC performed by cytocentrifugations can be performed by small laboratories which cannot invest in expensive automated equipments.
Acknowledgements
Drs Antoine M, Bouchar S, Bouche MJ, Briers MP, Chaikh A, Chef R, Cornut P, Crousse P, Dawagne MP, Debroux H, Demogue M, Frecourt N, Evrard MC, Girard J, Hautecoeur W, Keuller H, Lancelle F, Marechal M, M'Barek L, Miserque S, Oberweis D, Regnard C, Vranckx P, Walravens F and Zouaoui Boudjeltia K for their collaboration.
Co-editors of CytoJournal Vinod B. Shidham, MD, FRCPath, FIAC and Barbara F. Atkinson, MD thank the academic editor Prabodh Gupta, M.D., FIAC, University of Pennsylvania Medical Center, Cytopathology & Cytometry Lab. Founders Pavilion-6th Floor, 34th & Spruce Sts. Philadelphia, PA 19104 for organizing and completing the peer-review process for this manuscript.
==== Refs
Sherman ME Shiffman MH Lorenz AT Manos M Scott DR Kurman R Toward objective quality assurance in cervical cytopathology: correlation of cytopathology diagnosis with detection of high risk human papillomavirus types Am J Clin Path 1994 102 182 87 8042586
Sherman ME Mendoza M Lee KR Ashfaq R Birdsong GG Corkill ME Mc Intosh KM Inhorn SL Zahniser DS Baber G Barber C Stoler MH Performance of liquid-based thin layer cervical cytology: correlation with reference diagnose and human papillomavirus testing Mod Path 1998 11 837 43 9758363
Darwin LH Cullen AP Arthur PM Long CD Smith KR Girdner JL Hook EW 3rdQuinn TC Lorincz AT Comparison of Digene Hybrid Capture 2® and conventional culture for detection of Chlamydia trachomatis and Neisseria gonorrhoeae in cervical specimens J Clin Microbiol 2002 40 641 4 11825985 10.1128/JCM.40.02.641-644.2002
Nanda K McCrory DC Myers ER Bastian LA Hasselblad V Hickey JD Matchar DB Accuracy of the Papanicolaou test in screening for and follow-up of cervical cytology abnormalities: a systematic review Ann Intern Med 2000 132 810 9 10819705
Obwegeser JH Brack S Does liquid-based technology really improve detection of cervical neoplasia? A prospective, randomized trial comparing the ThinPrep Pap Test with conventional pap test, including follow-up of HSIL cases Acta Cytol 2001 45 709 14 11575648
Leif RC Nordqvist S Clay S Ingram D Cameron BF Bobbitt D Gaddis R Leif SB Cabanas A A procedure for dissociating Ayre scrape samples J Histochem Cytochem 1977 25 525 37 894001
Leif RC Ingram D Clay C Bobbitt D Gaddis R Leif SB Nordqvist S Optimization of the binding of dissociated exfoliated cervico-vaginal cells to glass microscope slides J Histochem Cytochem 1977 25 538 43 894002
Van driel-Kulker AM Ploem-Zaaijer JJ van der Zwan M Tanke HJ A preparation technique for exfoliated and aspirated cells allowing different staining procedures Anal Quant Cytol 1980 2 243 6 6162408
Khalbuss WE Rudomina D Kauff ND Chuang L Melamed MR SpinThin, a simple, inexpensive technique for preparation of thin-layer cervical cytology from liquid-based specimens. Data on 791 cases Cancer 2000 90 135 42 10896326 10.1002/1097-0142(20000625)90:3<135::AID-CNCR1>3.0.CO;2-L
Johnson T Maksem J Belsheim B Roose E Klock R Eatwell L Liquid-based cervical-cell collection with brushes and wooden spatulas: a comparison of 100 conventional smears from high-risk women to liquid-fixed cytocentrifuge slides, demonstrating a cost-effective, alternative monolayer slide preparation method Diagn Cytopathol 2000 22 86 91 10649517 10.1002/(SICI)1097-0339(200002)22:2<86::AID-DC5>3.0.CO;2-4
Maksem J Endocervical cells collection using cytobrush, liquid-fixation, and cytocentrifugation: A feasibility study using 455 hysterectomy specimens Diagn Cytopathol 1999 21 419 26 10572276 10.1002/(SICI)1097-0339(199912)21:6<419::AID-DC11>3.0.CO;2-2
Weynand B Berlière M Haumont E Massart F Pourvoyeur A Beri P A new liquid-based cytology technique: ThermoShandon's Papspin Acta Cytol 2003 4 149 53 12685180
Bergeron C Fagnani F Performance of a new liquid-based cervical screening technique in the clinical setting of a large French laboratory Acta Cytol 2003 47 753 61 14526674
Solomon D Nayar R The Bethesda system for reporting cervical cytology 2004 Second Springer-Verlag, New York 1994
Leduc F Bonniere M Farre I Staelen P Human papillomavirus detection by the hybrid capture II assay with the liquid cell preservation solution Easyfix Ann Biol Clin 2004 62 687 90
Terry G Ho L Londesboraugh P Cuzick J Mielzynska-Lohnas I Lorincz A Detection of high-risk HPV types by the Hybrid Capture 2 test J Med Virol 2001 65 155 162 11505458 10.1002/jmv.2015
Van den Brule AJC Meijer CJLM Bakels V Kenemans P Walboomers JMM Rapid detection of human papillomavirus in cervical scrapes by combined general primer-mediated and type-specific polymerase chain reaction J Clin Microbiol 1990 28 2739 2743 2177751
Saiki RK Gelfand DH Stoffel S Scharf SJ Higuchi R Horn GT Mullis KB Erlich H A. Primer-directed enzymatic amplification od DNA with a thermostable DNA polymerase Science 1988 239 487 491 2448875
Gravitt PE Peyton CL Alessi TQ Wheeler CM Coutlee F Hidersheim A Schiffman MH Scott DR Apple RJ Improved amplification of genital human papillomaviruses J Clin Microbiol 2000 38 357 361 10618116
Chhieng DC Roberson J Gidley J Eltoum I Bethesda 2001. Impact on the reporting of gynecologic cytology Acta Cytol 2004 48 355 62 15192951
Alves VA Bibbo M Scmitt FC Milanezi F Longatto Filho A Comparison of manual and automated methods of liquid-based cytology. A morphologic study Acta Cytol 2004 48 187 93 15085750
Nam JH Kim HS Lee JS Choi HS Min KW Park CS A comparison of modified MonoPrep2 of liquid-based cytology with ThinPrep Pap test Gynecol Oncol 2004 94 693 8 15350360 10.1016/j.ygyno.2004.05.047
Karnon J Peters J Platt J Chilcott J McGoogan E Bewer N Liquid-based cytology in cervical screening: an updated rapid and systematic review and economic analysis Health Technol Assess 2004 8 1 78
Abulafia O Pezzullo JC Sherer DM Performance of ThinPrep liquid-based cervical cytology in comparison with conventionally prepared Papanicolaou smears: a qualitative survey Gynecol oncol 2003 90 137 44 12821354 10.1016/S0090-8258(03)00176-8
Bergeron C Bishop J Lemarie A Cas F Ayivi J Huynh B Barrasso R Accuracy of thin-layer cytology in patients undergoing cervical cone biopsy Acta Cytol 2001 45 519 24 11480712
Utagawa ML Pereira SM Makabe S Maeda MY Marques JA Santoro CL Di Loreto C Aguira LS Pitolli JE Das Dores GB Castelo A Longatto Filho A Pap test in a high-risk population comparison of conventional and liquid-base cytology Diagn Cytopathol 2004 31 169 72 15349987 10.1002/dc.20118
Mattosinho de Castro Ferraz Mda G Nicolau SM Stavale JN Focchi J Castelo A Dores GB Mielsynska-Lohnas I Lorincz A Rodrigues de Lima G Cervical biopsy-based comparison of a new liquid-based thin-layer preparation with conventional Pap smears Diagn Cytopathol 2004 30 220 6 15048954 10.1002/dc.10409
Bibbo M Klump WJ DeCecco J Kovatich AJ Procedure for immunocytochemical detection of P16INK4A antigen in thin-layer, liquid-based specimens Acta Cytol 2002 46 25 9 11843554
Graflund M Sorbe B Bryne M Karlsson M The prognostic value of a histologic grading system, DNA profile, and MIB-1 expression in early stages of cervical squamous cell carcinomas Int J Gynecol Cancer 2002 12 149 57 11975674 10.1046/j.1525-1438.2002.01088.x
Bianchi A Moret F Desrues JM Champenois T Dervaux Y Desvouas O Oursin A Quinzat D Dachez R Bathelier C Ronsin C PreservCyt transport medium used for ThinPrep Pap test is a suitable medium for detection of Chlamydia trachomatis by the COBAS Amplicor CT/NG test: results of preliminary study and future implications J Clin Microbiol 2002 40 1749 54 11980955 10.1128/JCM.40.5.1749-1754.2002
Guo M Hu L Baliga M He Z Hughson MD The predictive value of p16INK4a and hybrid capture 2 human papillomavirus testing for high-grade cervical intraepithelial neoplasia Am J Clin Pathol 2004 122 894 901 15539381 10.1309/0DGG-QBDQ-AMJC-JBXB
Luzzatto L van Haaften C Boon ME Proliferation patterns of cervical cells as visualized in Leiden liquid cytology slides Diagn Cytopathol 2004 31 5 9 15236256 10.1002/dc.20011
Lee K Lee J Saw H Can human papillomavirus DNA testing substitute for cytology in the detection of high-grade cervical lesions? Arch Pathol Lab Med 2004 128 298 302 14987159
Howard M Sellers J Kaczoroswski J Optimizing the hybrid capture II human papillomavirus test to detect cervical intraepithelial neoplasia Obstet Gynecol 2002 100 972 80 12423863 10.1016/S0029-7844(02)02315-3
Lytwyn A Sellors J Mahony J Daya D Chapman W Howard M Roth P Lorincz A Gafni A Walter S Adjunctive Human papillomavirus testing in the 2-year follow-up women with low-grade cervical cytology abnormalities Arch Pathol Lab Med 2003 127 1169 75 12946228
Levi AW Kelly DP Rosenthal DL Ronnett BM Atypical squamous cells of undetermined significance in liquid-based cytologic specimens: results of reflex human papillomavirus testing and histologic follow-upo in routine practice with comparison of interpretive and probabilistic reporting methods Cancer 2003 99 191 7 12925979 10.1002/cncr.11507
Manos MM Kinney WK Hurley LB Sherman ME Shieh-Ngai J Kurman RJ Ransley JE Fetterman BJ Hartinger JS McIntosh KM Pawlick GF Hiatt RA Identifying women with cervical neoplasia: using human papillomavirus DNA testing for equivocal Papanicolaou results JAMA 1999 281 1605 10 10235153 10.1001/jama.281.17.1605
Bollmann R Mehes G Torka R Speich N Schmitt C Bollmann M Detremination of features indicating progression in atypical squamous cells with undetermined significance: human papillomavirus typing and DNA ploidy analysis from liquid-based cytologic samples Cancer 2003 99 113 7 12704691 10.1002/cncr.11010
Sherman ME Wang SS Tarone R Rich L Schiffman M Histopathologic extend of cervical intraepithelial neoplasia 3 lesions in the atypical squamous intraepithelial lesion triage study: implications for subjet safety and lead-time biais Cancer Epidemiol Biomarkers Prev 2003 12 372 379 12692113
|
16164754
|
PMC1262738
|
CC BY
|
2021-01-04 16:36:22
|
no
|
Cytojournal. 2005 Sep 15; 2:15
|
utf-8
|
Cytojournal
| 2,005 |
10.1186/1742-6413-2-15
|
oa_comm
|
==== Front
Environ HealthEnvironmental Health1476-069XBioMed Central London 1476-069X-4-201620212810.1186/1476-069X-4-20ResearchInter-individual variations of human mercury exposure biomarkers: a cross-sectional assessment Berglund Marika [email protected] Birger [email protected]örnberg Karolin Ask [email protected] Brita [email protected] Östen [email protected] Marie [email protected] Department of Metals and Health, Institute of Environmental Medicine, Karolinska Institutet, Box 210, SE-171 77 Stockholm, Sweden2 Analys Modul Sweden AB, Tingsvägen 19, SE-191 61 Sollentuna, Sweden2005 3 10 2005 4 20 20 26 7 2005 3 10 2005 Copyright © 2005 Berglund et al; licensee BioMed Central Ltd.2005Berglund et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Biomarkers for mercury (Hg) exposure have frequently been used to assess exposure and risk in various groups of the general population. We have evaluated the most frequently used biomarkers and the physiology on which they are based, to explore the inter-individual variations and their suitability for exposure assessment.
Methods
Concentrations of total Hg (THg), inorganic Hg (IHg) and organic Hg (OHg, assumed to be methylmercury; MeHg) were determined in whole blood, red blood cells, plasma, hair and urine from Swedish men and women. An automated multiple injection cold vapour atomic fluorescence spectrophotometry analytical system for Hg analysis was developed, which provided high sensitivity, accuracy, and precision. The distribution of the various mercury forms in the different biological media was explored.
Results
About 90% of the mercury found in the red blood cells was in the form of MeHg with small inter-individual variations, and part of the IHg found in the red blood cells could be attributed to demethylated MeHg. THg in plasma was associated with both IHg and MeHg, with large inter-individual variations in the distribution between red blood cells and plasma. THg in hair reflects MeHg exposure at all exposure levels, and not IHg exposure. The small fraction of IHg in hair is most probably emanating from demethylated MeHg. The inter-individual variation in the blood to hair ratio was very large. The variability seemed to decrease with increasing OHg in blood, most probably due to more frequent fish consumption and thereby blood concentrations approaching steady state. THg in urine reflected IHg exposure, also at very low IHg exposure levels.
Conclusion
The use of THg concentration in whole blood as a proxy for MeHg exposure will give rise to an overestimation of the MeHg exposure depending on the degree of IHg exposure, why speciation of mercury forms is needed. THg in RBC and hair are suitable proxies for MeHg exposure. Using THg concentration in plasma as a measure of IHg exposure can lead to significant exposure misclassification. THg in urine is a suitable proxy for IHg exposure.
==== Body
Background
People are exposed to different forms of mercury (Hg), which differ with respect to toxicology. The target organ for methylmercury (MeHg) toxicity is the brain, which is especially susceptible during development [1]. Target organs for elemental mercury vapour (Hg0) are the brain and kidney and the target organ for inorganic Hg compounds (IHg, Hg2+) is the kidney [1]. Both MeHg and Hg0, but not IHg, readily passes the blood-brain and placental barriers [1]. Exposure to MeHg occurs almost exclusively via consumption of seafood, especially predatory fish and large marine mammals, while food in general contains low concentrations of both MeHg and IHg [2-4]. Dental amalgam fillings, releasing Hg0, are the major source of Hg0 exposure in the general population [5].
In the gastrointestinal tract MeHg is absorbed to approximately 95% [6,7], Hg2+ to approximately 7% [8] and elemental Hg to less than 0.01% [9,10]. The absorption of Hg0 in the lung is about 80% [11]. Within tissues, MeHg is slowly demethylated to Hg2+ [12,13]. In the blood, Hg0 is readily oxidized to Hg2+ by catalase [14].
The total mercury concentration (THg) in blood is often used as a proxy measure of MeHg exposure in individuals eating fish with the assumption that the IHg exposure, and thereby the IHg concentration in blood, is much lower [15-18]. In the blood, more than 90% of MeHg is bound to haemoglobin in the red blood cells (RBC), while IHg is more evenly distributed between RBC and plasma [7,19]. Therefore, total Hg in RBC is also sometimes used as a proxy measure of MeHg exposure [20-23] and total Hg in plasma is used as a proxy measure of IHg exposure (Hg2+ and Hg0; [3,22,24-26]).
The concentration of total Hg in hair (H-THg) is often used as a measure of MeHg exposure, assuming that > 80% of Hg in hair is in the form of MeHg [27]. Mercury is incorporated in hair during formation in the hair follicle, and mercury in hair is associated with the concentration of MeHg in blood [19]. It has been proposed that H-THg reflects inorganic mercury exposure at low MeHg exposure in populations with no or low fish consumption [1].
The total Hg concentration in urine is used as a measure of IHg exposure as MeHg is excreted primarily via the bile (as glutathione complex) and faeces (about 90%; as IHg) and only to a limited extent (about 10%) in urine (as IHg; [1,3,28]).
The mercury biomarkers are frequently used for estimation of exposure and risks of health effects, but the inter-individual variations are not well known. The available information on Hg kinetics is based on 25–35 years old experimental studies, sometimes with high exposure levels, involving a limited number of volunteers. The aim of the present study was to investigate the robustness of some of the statements forming the basis for the biomarkers frequently used, and to explore the inter-individual variations. In order to do so, we have improved the traditional Magos' mercury speciation method [29,30] and developed an automated analytical system for speciation of mercury in whole blood, RBC, plasma, hair and urine.
Methods
Sampling
In 2003 we recruited 28 volunteers, 23 women and 5 men, 28–60 years of age (mean 48 years) for measurement of Hg biomarkers. Sampling comprised venous blood from the cubital vein (5 mL, Venoject II, EDTA(K2), VP-050SDK), red blood cells (RBC) and plasma (5 mL, Venoject II, EDTA(K2), VP-050SDK; Terumo Corp., Leuven; Belgium), hair (a hair sample was tied with a cotton thread, cut close to the scalp from the back of the head and put into a plastic bag), and urine (a spot sample collected in acid washed plastic containers). Information regarding fish intake (usual number of meals/month) and number of dental amalgam fillings was collected via self-reported questionnaires. A usual number of 0–22 fish meals/months and a total number of dental amalgam fillings between 0–15 were reported. For evaluation of Hg distribution in hair, and blood to hair ratio, we also used data previously collected in a study of women with a high fish intake (N = 145, 20–50 years of age; [31]). The study was approved by the Ethics Committee of the Karolinska Institutet, Stockholm.
Sample treatment
Whole blood, RBC, plasma and urine samples (1.0 mL) were treated with 1.0 mL L-cysteine (0.012 M), 1.5 mL NaOH (11 M) and 0.5 mL deionised water, and stored in the dark over night at room temperature to complete the solubilisation. Hair samples (3 cm from the scalp end; approximately 20 mg) were treated with 2.0 mL L-cysteine (0.083 M), 4.0 mL NaOH (11 M) and 14 mL NaCl (0.17 M). The mixture was heated to 90–95°C for 20 minutes to complete the solubilisation.
Analyses
Total mercury (THg) and inorganic Hg (IHg) were analyzed in whole blood, RBC, plasma, hair, and urine using cold vapour atomic fluorescence spectrophotometry (CVAFS, Merlin, PSA 10.023; P.S. Analytical Ltd., Orpington, Kent, UK), following reduction to Hg0 in a reaction tower, using an automatic multiple-injection analysis (MIA) system, with a Tefzel® 13-channel selector valve (Analys Modul Sweden AB; Figure 1). In this system, a motor-driven pump (Microlab 900, Hamilton Bonaduz AG, Switzerland) is connected to the central port of the selector valve, which opens to one of 13 peripheral ports at a time. The pump dispenses volumes with low variation which enables low volumes of chemicals to be used. Deionised water is used as an extended syringe piston and washing medium. No chemicals or samples reach the syringe at the base of the tube loop (Figure 1). The sampler, the reaction tower and the reagent bottles are connected to the peripheral ports by Tefzel® tubes. The sampler and the selector, with an AD converter, are controlled by a PC using a control program (EASYLAB, Analys Modul Sweden AB). The program executes the commands from a command list for THg or IHg in a sequential or logical order. In order to reduce blank values, all reagents are initially mixed in the reaction tower to eliminate any Hg impurities. Any Hg0 formed in this initial cleaning step is transported with argon (Ar) gas through the detector.
Figure 1 The automatic mercury analysis system. The automatic multiple-injection analysis (MIA) system, with a Tefzel 13-channel selector valve (Analys Modul Sweden AB) for mercury analysis.
For determination of IHg, 800 μl L-cysteine solution (0.1% w/v L-cysteine in 1.5% w/v NaCl), 200 μl 8 M H2SO4 (p.a.) with 0.4% antifoaming agent (Antifoam 204; Sigma Chemical Co., S:t Louis, MO, USA; soluble in acid but not in alkaline solution), 2000 μl 11.25 M NaOH, 100 μl deionised water, and 100 μl 10% w/v SnCl2 in 2.4 M H2SO4 were delivered to the reaction tower for elimination of Hg impurities, after which 500 μl sample solution and 200 μl cysteine solution, followed by 1200 μl of deionised water (for rinsing) were added. For determination of THg, 800 μg L-cysteine solution, 600 μl 8 M H2SO4 with 0.4% antifoaming agent, 2000 μl 11.25 M NaOH, 200 μl cysteine solution, 1000 μl deionised water, and 100 μl of a mixture of 10% CdCl2 and 50% SnCl2 in 8 M H2SO4 were delivered to the reaction tower for elimination of Hg impurities. Then 500 μl 8 M H2SO4 (supra pure) was added (in order to increase the temperature), followed by 500 μl sample solution and 500 μl deionised water.
The Hg0 released from the sample was transported by Ar gas (0.087 L/min) through a moisture trap, chilled with ice, followed by a tubular permeable membrane (Perma Pure mini-dryers, model MD-125-12S, Perma Pure Products, Inc, Farmingdale, USA) before reaching the AFS detector. The signal was stored by the PC and also recorded on paper for process control (Perkin Elmer Model 56 recorder). The area under the curve was integrated by the computer and used for evaluation of the amount (ng) Hg in the sample. The shield gas flow (Ar) for the detector was 0.099 L Ar/min. The standard solutions (0.1–2.0 ng Hg/mL) were made in 0.1% L-cysteine. Furthermore, a MeHg standard of 0.4 ng/mL was included in the standard curve in order to control the degree of demethylation in the IHg analysis and for recovery in the THg analysis. The time interval between reactions was 15 minutes. All samples were analyzed in duplicates.
The concentrations of the organic mercury fraction (OHg) were calculated by the subtraction of the IHg concentrations from that of the THg concentrations. OHg is assumed to be mainly MeHg as the only other known exposure sources of organic mercury compounds in Sweden are a few vaccines containing thiomersal (ethylmercurithiosalicylate), a seldom-used preservative containing ethylmercury. Hg concentrations in urine were adjusted to specific gravity (1.019 μg/mL; urine specific gravity refractometer, Uricon-Ne, Atago Co., Ltd., Tokyo, Japan) and creatinine (analyzed at the Department of Clinical chemistry, Karolinska University Hospital, Stockholm). Haematocrit and haemoglobin (Hb) concentrations were measured (Department of Clinical chemistry, Karolinska University Hospital, Stockholm).
Evaluation of MeHg demethylation during sample treatment and analysis
For evaluation of potential demethylation during solubilisation, a purified (> 99%) radiolabeled methylmercury-chloride (203Hg, Amersham Laboratories, Amersham, UK) solution was solubilised according to the method described above, at room temperature and at 88°C for one hour. The total Hg concentration was measured by gamma-counting (Searle 1195, Searle Analytical Inc.). The loss of total Hg during solubilisation was < 1%, i.e. not detectable. An aliquot (10 g) of each of the solubilised solutions was acidified by 3 mL 6 M HCl, stored over night at +4°C, extracted with chloroform three times (30+20+10 mL) to separate MeHg and IHg [32]. The water phase, containing the IHg from MeHg demethylation, was then measured by gamma-counting to calculate the degree of demethylation during solubilisation.
In order to quantify the degree of demethylation in the reaction tower during the analytical step, samples with and without addition of IHg and MeHg were solubilised and then acidified and extracted with chloroform, following the procedure described above. THg and IHg were then measured by the analytical method described above, and IHg from demethylation of MeHg was calculated as the percentage of the initial amounts.
Analytical quality control
The blood sampling material was tested for Hg contamination. Simulated blood sampling using a weak acid (0.03 M HNO3) was performed. The acid solutions were analyzed for Hg content by the method described above. The material was found to be essentially free from Hg contamination (all acid solutions were below the limit of detection, LOD, i.e. 3 standard deviations of the mean of the chemical blanks). All other materials used for analysis were acid washed. Appropriate reference materials for Hg in blood, serum, urine and hair were analyzed in each analytical run, respectively (see Results and Table 1).
Table 1 Results of analytical quality control. Results (Mean; standard deviation, SD; coefficient of variation, CV%; and number, n) of repeated analyses of total mercury (THg) and inorganic mercury (IHg) in reference materials analyzed together with collected samples of whole blood, red blood cells (RBC), plasma, urine (μg/L) and hair (mg/kg).
Media Reference material THg IHg
Mean (SD) Mean (SD)
CV% (n) CV% (n)
Blood/RBC Seronorm 404107X
Rec. value THg: 3 μg/L 2.3 (0.18) 0.54 (0.051)
Range: 2.2–3.3 μg/L 8.1 (7) 9.4 (7)
Seronorm 404108
Rec. value THg: 8 μg/L 8.1 (0.52) 6.4 (0.32)
Range: 6.7–8.4 μg/L 6.3 (7) 5.0 (7)
Plasma Seronorm 201405
Rec. value THg: 0.96 μg/L 1.0 (0.081) 0.70 (0.017)
Range: 0.87–1.06 μg/L 8.0 (5) 2.4 (5)
Urine Seronorm 2524
Rec. value THg: 0.21 μg/L 0.20 (0.0068) 0.11 (0.011)
95% CI: 0.17–0.25 μg/L 3.4 (4) 10 (4)
Seronorm 2525
Rec. value THg: 40.3 μg/L 38 (1.9) 38 (0.85)
95% CI: 37.7–42.9 μg/L 5.0 (6) 2.2 (6)
Hair IAEA086
Rec. value THg: 0.573 mg/kg 0.58 (0.028) 0.27 (0.022)
95% CI: 0.534–0.612 4.8 (5) 8.1 (5)
IMM-hair1
Rec. value THg: 4.8 mg/kg 4.7 (0.14) 0.50 (0.026)
SD: 0.3 mg/kg 3.0 (5) 5.1 (5)
1 Björnberg et al. 2003
Statistics
We used spearman correlation (rs) test to test for associations between parameters, and linear regression analysis for evaluation of association between parameters when the requirements for normally distributed residuals were met. Statistical analyses were conducted using SigmaStat® (Version 2.03 for Windows (Systat Software GmbH, Erkrath, Germany). Statistical significance was set to p < 0.05.
Results
Analytical method
The accuracy of the Hg speciation method, as evaluated by repeated analyses of reference materials, was satisfactory (Table 1). There are no commercially available reference materials for IHg. However the obtained values for IHg in blood were well in agreement with our results from previous analytical runs of the same Seronorm sample [31,33]. The analytical variability, as calculated by coefficients of variation (CV%) of duplicate analysis of collected samples and reference materials, was low (Table 2). The detection limits (LOD) were lowered significantly by the introduction of the cleaning step of the reagent chemicals in the reaction tower prior to the sample addition (Table 3). As a result, very few samples of whole blood (n = 3), plasma (n = 2), and RBC (n = 3) had IHg concentrations below LOD.
Table 2 Precision of duplicate analyses. Precision (Coefficient of variation, CV%) of duplicate analyses of total mercury (THg) and inorganic mercury (IHg) in blood, red blood cells (RBC), plasma, urine (μg/L) and hair (mg/kg), in collected samples and reference materials.
THg IHg
CV (%) No. of duplicate analyses Concentration range CV (%) No. of duplicate analyses Concentration range
Blood 4.3 32 0.34–8.5 6.4 30 0.079–6.8
RBC 5.9 34 0.38–14 12 24 0.061–6.7
Plasma 5.3 29 0.048–1.3 7.3 26 0.060–1.1
Hair 2.6 32 0.081–4.9 9.0 31 0.01–0.52
Urine 3.2 35 0.19–6.2 2.5 34 0.094–6.3
Table 3 Concentrations of total, inorganic and organic mercury in various biological media. Concentrations of total mercury (THg), inorganic mercury (IHg) and organic mercury (OHg) in whole blood, plasma, red blood cells (RBC), urine and hair in 28 individuals, and limits of detection (LOD, i.e. 3 × standard deviation of mean of chemical blank/solubilisation solution; the number of samples was 5–10) of the two chemical runs.
THg IHg OHg
Whole blood (μg/L) Mean ± SD 2.2 ± 1.4 0.35 ± 0.23 1.8 ± 1.3
Median 2.0 0.35 1.6
Range 0.34–7.3 0–0.94 0.26–6.9
LOD 0.05/0.09 0.03/0.06
Plasma (μg/L) Mean ± SD 0.65 ± 0.30 0.39 ± 0.26 0.26 ± 0.16
Median 0.63 0.37 0.22
Range 0.07–1.3 0–1.1 0.05–0.70
LOD 0.04 0.02/0.05
RBC (μg/L) Mean ± SD 4.1 ± 2.6 0.29 ± 0.18 3.8 ± 2.5
Median 4.0 0.26 3.6
Range 0.40–14 0–0.70 0.25–13
LOD 0.03/0.04 0.05/0.05
Urine (μg/L; adjusted to density 1.019) Mean ± SD 1.4 ± 1.2 1.4 ± 1.2 0.012 ± 0.073
Median 1.0 1.0 0.015
Range 0.27–6.1 0.18–6.3 0–0.11
LOD 0.03/0.05 0.03/0.02
Urine (μg/g creatinine) Mean ± SD 1.9 ± 2.0 1.9 ± 2.1 0.013 ± 0.12
Median 1.3 1.2 0.018
Range 0.12–10 0.12–11 0–0.23
Hair (mg/kg) Mean ± SD 0.76 ± 0.40 0.062 ± 0.030 0.69 ± 0.37
Median 0.71 0.060 0.66
Range 0.08–2.0 0.010–0.12 0.072–1.9
LOD 0.01 0.01
Demethylation of MeHg to IHg takes place during the solubilisation of samples, and in the reaction tower, during the analysis of the solubilised samples. After solubilisation at room temperature, acidification and extraction with chloroform, the percentage of excess IHg from demethylation of MeHg was 0.9 ± 0.1% (n = 4). The demethylation after solubilisation at 88°C was 2.6 ± 0.5% (n = 4). The acidification step, which is necessary in order to perform the extraction, probably also, increases to a small extent the degree of demethylation. In the reaction tower, the percentage of excess IHg (from further demethylation of MeHg) was calculated to 3.0 ± 0.3% (n = 4).
Biomarker concentrations and correlations
A summary of the concentrations of Hg species in whole blood (B), red blood cells (RBC), plasma (P), urine (U) and hair (H) is given in Table 3. The correlations between the Hg species in the various media as well as the exposure variables fish consumption (number of meals per month; range 0–22 meals per month) and number of dental amalgam fillings (range 0–>15), are given in Table 4. Fish consumption was positively correlated with THg in blood (rs = 0.74, p < 0.001), RBC, and hair, and with OHg in blood, RBC, plasma and hair (Table 4). Fish consumption was also correlated with IHg in hair. Number of dental amalgam fillings was positively correlated with THg in plasma (rs = 0.46, p = 0.01) and urine (rs = 0.49, p = 0.009), and with IHg in blood, plasma and urine (Table 4). Mercury levels in blood (THg, IHg and OHg in whole blood, RBC or plasma) were not associated with haemoglobin and haematocrit.
Table 4 The Spearman correlation coefficients between mercury species in different media and exposure vavariables. Spearman correlation coefficients of inorganic mercury (IHg) and organic mercury (OHg) species in whole blood (B; μg/L), red blood cells (RBC; μg/L), plasma (P; μg/L) and urine (U: μg/L, adjusted to density 1.019) and IHg, OHg and total mercury (THg) in hair (H; mg/kg) and the exposure variables fish consumption (number of meals per month) and number of dental amalgam fillings. The number of samples are 25–28. The significance level is indicated as * p < 0.05; ** p < 0.01; *** p < 0.001.
B-OHg RBC-IHg RBC-OHg P-IHg P-OHg H-THg H-IHg H-OHg U-IHg U-OHg Fish Amalgam
B-IHg 0.19 0.83 *** 0.25 0.91 *** 0.05 0.28 0.29 0.27 0.81 *** -0.11 0.07 0.48 *
B-OHg 0.38 0.96 *** 0.07 0.82 *** 0.87 *** 0.79 *** 0.87 *** -0.07 0.27 0.82 *** 0.09
RBC-IHg 0.45 * 0.70 *** 0.34 0.46 * 0.42 * 0.43 * 0.73 *** 0.04 0.37 0.27
RBC-OHg 0.13 0.77 *** 0.82 *** 0.81 *** 0.81 *** 0.03 0.22 0.76 *** 0.14
P-IHg -0.13 0.16 0.16 0.14 0.74 *** -0.14 -0.04 0.49 *
P-OHg 0.77 *** 0.74 *** 0.75 *** -0.003 0.29 0.82 *** 0.06
H-THg 0.86 *** 0.99 *** 0.18 0.32 0.75 *** 0.28
H-IHg 0.83 *** 0.28 0.28 0.63 *** 0.32
H-OHg 0.17 0.31 0.74 *** 0.30
U-IHg 0.05 -0.03 0.49 **
U-OHg 0.24 -0.007
Fish 0.13
Mercury in blood
The distribution of OHg and IHg in whole blood between RBC and plasma was calculated as the percentage of total OHg (or IHg) in whole blood according to equation 1 and 2 (below), using individual haematocrit values (B-EVF, %). The range of B-EVF was 35–47% (mean 42%). Data below LOD were not included because of their uncertainty.
1) RBC-OHg * (B-EVF/B-OHg) * 100
2) P-OHg * ((1-B-EVF)/B-OHg) * 100
On average 87% of OHg in whole blood was localized in the RBC (95% CI of mean ± 3.7; range 76–104%; n = 20) and 9.6% in plasma (95% CI of mean ± 1.6; range 5.1–20%; n = 22). On average 34% of IHg in whole blood was localized in RBC (95% CI of mean ± 4.0; range 15–54%; n = 22) and 64% in plasma (mean; 95% CI of mean ± 5.4; range 30–81%; n = 22). The distribution of OHg or IHg between RBC and plasma did not change with increasing concentrations of the respective Hg form.
The concentration of IHg in RBC was positively correlated with both the concentration of OHg in RBC and the concentration of IHg in plasma (Table 4) indicating that IHg in RBC is a function of both IHg and OHg exposures. RBC-IHg was on average 6.8% of RBC-THg (median; range 3.3–24%), and increased with increasing concentrations of RBC-OHg (rs = 0.46; p = 0.03) and increasing consumption of fish (rs = 0.60; p = 0.003), but not with increasing number of dental amalgam fillings. In a person with no dental amalgam fillings RBC-IHg was 4.6% of RBC-THg.
The average RBC to plasma ratio of IHg concentrations was 0.90 (range 0.25–2.5), or as evaluated by linear regression, 0.50 (RBC-IHg = 0.11+0.50*P-IHg; R2 = 0.56; Figure 2). The ratio increased with fish consumption (rs = 0.52; p = 0.008; n = 25), but not with the number of dental amalgam fillings (rs = -0.23).
Figure 2 The relationship between inorganic mercury in plasma and red blood cells. The relationship between inorganic mercury in plasma (P-IHg) and red blood cells (RBC-IHg) evaluated by linear regression (RBC-IHg = 0.11+0.50 P-IHg; R2 = 0.56).
The average RBC to plasma ratio of OHg concentrations was 14 (range 3.1–28). When evaluated by linear regression the ratio was also 14 (RBC-OHg = 0.20+14*P-OHg; R2 = 0.72). The ratio did not increase with fish consumption (or number of dental amalgam fillings).
Mercury in hair
The total mercury concentration in hair (H-THg) was positively correlated with B-OHg and P-OHg, as well as with fish consumption, but not with B-IHg or the number of dental amalgam fillings (Table 4). Speciation of Hg in hair showed that on average 91% of THg was OHg (CI of mean ± 1.2; range 79–95%; n = 28), and 8.9% was IHg (CI of mean ± 1.1; range 4.9–21%; n = 28). In our previous study of women with a high fish intake, the distribution was approximately the same, i.e. 91% of THg was OHg (CI of mean ± 1.2; range 82–97%) and 8.7% of THg was IHg (CI of mean ± 1.2; range 3.2–18%; n = 144). The percentage of IHg in hair was not associated with the number of dental amalgam fillings. The average percentage of IHg in hair was 8.3% (range 4.4–13%; n = 23) in individuals without dental amalgam fillings, and 8.8% (range 4.6–18%; n = 48) in individuals with 10 fillings or more (including data from our previous study). The difference was not statistically significant (Student's t-test, p = 0.4). The concentration of IHg in hair was highly correlated with OHg in hair, and with OHg in blood (B-OHg, RBC-OHg and P-OHg; Table 4), but not with IHg in blood (B-IHg, P-IHg or RBC-IHg; Table 4). The concentration of IHg in hair was also positively correlated with fish consumption, but not with number of dental amalgam fillings (Table 4).
The average hair to blood ratio, (H-THg (mg/kg) divided by B-THg (μg/L)) was 0.366 (median 0.373; 95th percentile 0.552; range 0.185 to 0.673). If H-THg was divided by B-OHg, the ratio was 0.465 (95th percentile 0.670). If we included the hair and blood mercury data from our previous study of women with a high fish consumption the average hair to blood ratio was 0.341 (median 0.330; 5th percentile 0.168; 95th percentile 0.563; range 0.066–0.824; n = 173). The hair to blood ratio seemed to decrease with increasing B-OHg (Figure 3).
Figure 3 The hair to blood ratio of total mercury versus organic mercury in blood. The hair to blood ratio of total mercury (H-THg/B-THg) as a function of organic mercury in blood (B-OHg).
The ratio as determined by linear regression of H-THg versus, B-THg was 0.264, i.e. H-THg = 0.179+0.264*B-THg (R2 = 0.83; p < 0.001; n = 28). Replacing B-THg with B-OHg only resulted in an increased intercept, to 0.282 (R2 = 0.80; n = 25). Inclusion of data from our previous study in the linear regression analysis resulted in H-THg = 0.169+0.254*B-THg (R2 = 0.62; p < 0.001; n = 173; Figure 4).
Figure 4 The relationship between total mercury in hair and blood. The relationship between total mercury in hair (H-THg) and blood (B-THg) evaluated by linear regression (H-THg = 0.169+0.254 B-THg; R2 = 0.62).
Mercury in urine
Essentially all Hg in urine (> 98%) was IHg. IHg in urine, adjusted to specific gravity (1.019 g/mL) or adjusted to creatinine (g creatinine/L urine) were highly correlated with IHg in blood, plasma and RBC, but not with OHg in the various media (Table 4). IHg in urine was moderately associated with the number of dental amalgam fillings, but not with fish consumption (Table 4).
Discussion
This study of Hg biomarkers was possible due to the improvements and modifications of the CVAFS method used for determination of total and inorganic Hg in blood, hair and urine. The method provides high sensitivity, low analytical variability, and high accuracy also in the low concentration range. The limits of detection (between 0.01 and 0.09) were about 2–10 times lower than those previously reported [18,22,23,30,34-37]. It can be concluded that our modified analytical method is suitable for the purposes of speciation of Hg in human biological media and for evaluation of the main exposure sources.
The total variability in the different biomarkers measured includes inter-individual differences in the Hg kinetics as well as demethylation of MeHg to IHg during sample treatment and analysis. We have determined the degree of demethylation in our analytical procedure and we conclude that the method results in about 5% demethylation of MeHg, half of it in the solubilisation step if samples are heated during solubilisation (as for hair), and the other half in the analytical step. If solubilisation takes place at room temperature, as for blood, RBC, plasma and urine, the overall demethylation is further reduced, to less than 4%. The acidification of the samples, which was a prerequisite for the extraction procedure and the separation of IHg from MeHg, may be responsible for some of the demethylation during the tests. The demethylation of MeHg in blood during sample preparation and analysis has previously been reported to be 2–3% using acidic digestion [38].
The distribution of OHg between RBC (87%) and plasma (9.6%) was in good agreement with the earlier observations that the major part of MeHg in blood is found in the RBC, bound to haemoglobin [6,7,19]. The inter-individual variation was relatively low (total range about 15%). The average RBC to plasma OHg ratio of 14 found in the present study was between the ratios of 10 [6,7] and 20 [19], previously reported. In those studies, a limited number of volunteers (3 to 15 male and female volunteers) were given oral doses of either a radioactive MeHg salt in solution (about 10 μg Hg; [7]), radioactive MeHg bound to fish muscle protein (about 10 μg Hg; [6]) or a meal of fish containing 18–22 μg Hg/kg b.w. (providing 1400 μg Hg/70 kg man; [19]). The RBC to plasma OHg ratio in the present study was not influenced by the mercury concentrations in blood. Thus, it seems as the distribution of MeHg between RBC and plasma is rather constant over a large range of exposures.
The distribution of IHg between RBC (34%) and plasma (64%) displayed a much larger inter-individual variation (total range about 40–50%) than that of OHg. The RBC-IHg was positively correlated with both P-IHg and RBC-OHg, but P-IHg was not correlated with RBC-OHg. It can be concluded that IHg in RBC is partly emanating from inorganic Hg exposure, mainly Hg0 via amalgam, and partly from MeHg exposure via fish, which has demethylated to IHg in the body and, to some extent, in the analysis (less than 4%). Thus, the variation in MeHg exposure from fish adds to the variation in RBC-IHg, which can partly explain the larger inter-individual variation measured in the distribution of IHg between RBC and plasma. Little is known about the mechanisms involved in the conversion of MeHg to IHg in the human body, and the inter-individual variability. Based on our data, there seems to be little demethylation taking place in the blood (a few percent).
Our data strongly indicates that the small fraction of IHg in hair (about 9%), with relatively small inter-individual variations (CV about 15%) is a result of MeHg exposure and demethylation of MeHg in blood or hair follicles (and in the analysis), rather than a result of IHg exposure. IHg in hair was positively correlated with fish intake, but not with dental amalgam fillings. It was also highly correlated with OHg in blood, RBC and plasma. The hypothesis is further supported by our results in non-fish eating individuals, which showed a positive correlation of MeHg in blood and hair, but no correlation of IHg in blood and hair, despite a very low MeHg exposure (B-MeHg below 1.0 μg/L; [4]). MeHg in hair has been shown to be stable over time [39,40], indicating that demethylation within the hair strand is very limited. However, it should be borne in mind that artificial waving and other hair treatments may reduce Hg concentrations within the hair strand [41]. It has previously been suggested that MeHg is demethylated to inorganic Hg in the cells of the hair follicle [27]. As the IHg fraction in hair was about 9% and since the demethylation of MeHg in the hair analysis is about 5%, the average degree of demethylation in the hair follicles would be on average 4%. Because of the demethylation, THg in hair is a better measure of MeHg exposure than MeHg in hair.
In humans, a frequently cited blood to hair ratio (B-THg:H-THg) evaluated by linear regression is 1:250, however with large inter-study variations (range 140–370; [2,28]). In the present study, the ratio as evaluated by linear regression was 1:254. When evaluating mercury blood to hair ratios by linear regression there is always a positive intercept. The intercept may reflect the different time frames of the integrated exposure as measured in hair and blood, and occasional high MeHg exposure. If the blood to hair ratio of 1:250 is used to calculate B-MeHg from H-THg, B-MeHg will always be underestimated due to the positive intercept. The inter-individual variation in the blood to hair ratio as determined by division was very large. The variability seemed to decrease with increasing B-OHg concentrations (Figure 3), most probably due to more frequent fish consumption and thereby blood concentrations approaching steady state.
Our data shows that IHg in urine reflects the IHg exposure as nearly all Hg in urine (> 98%) was IHg and as IHg in urine did not reflect fish consumption or the OHg concentration in various media. Experimental data report greater concentrations of Hg in kidneys in males than in females exposed similarly [42]. A higher excretion of IHg in urine in women (1.5 μg/L adjusted to specific gravity 1.019 g/mL and 2.1 μg/g creatinine) than in men (0.80 μg/L adjusted to specific gravity 1.019 g/mL and 0.75 μg/g creatinine; p = 0.03) was noted in the present study, despite a similar exposure to IHg as measured by IHg in plasma (0.4 μg/L). However, the sample size was too small to draw any conclusions from those data. Further studies are warranted on gender differences in Hg metabolism and toxicity.
Conclusion
As expected, fish consumption was positively correlated with THg in blood, RBC, and hair. The use of THg concentration in blood as a proxy for MeHg exposure will give rise to an overestimation of the MeHg exposure, small or large, depending on the exposure to IHg (Hg2+ and Hg0). In order to reduce the inter-individual variability it can be recommended to speciate the various forms of Hg in blood when evaluating exposure, dose and risk for health effects. The demethylation taking place during sample preparation and analysis with this method will lead to a small underestimation of the MeHg concentration and an overestimation of the IHg concentration in the sample.
The total Hg concentration in the RBC gives a good measure of the MeHg exposure at low IHg exposure levels. Most of the Hg found in the RBC is in the form of MeHg with small inter-individual variations. Part of the IHg in RBC is emanating from demethylated MeHg, leaving a small fraction of IHg that is the result of IHg exposure.
Using THg concentrations in plasma as a measure of IHg exposure can lead to significant exposure misclassifications. The total concentration of Hg in plasma is associated with both IHg and OHg, with large inter-individual variations in the distribution between RBC and plasma, depending on both the MeHg and IHg exposure.
The THg concentration in hair reflects MeHg exposure at all exposure levels. The small fraction of IHg in hair is most probably emanating from MeHg that was demethylated in the body and during the sample preparation and analysis. IHg in hair was also correlated with fish consumption. THg in hair seems to provide the best measure of long-term average MeHg exposure. THg in urine reflects IHg exposure, also at very low exposure levels. Number of dental amalgam fillings was highly positively correlated with THg in plasma and urine.
List of abbreviations
Ar Argon
B Blood
B-EVF Blood-Erythrocyte Volume Fraction (haematocrit)
b.w. Body weight
CI Confidence interval
CV Coefficient of variation
CVAFS Cold vapour atomic fluorescence spectrophotometry
H Hair
Hg0 Elemental mercury vapour
Hg2+ Inorganic mercury, ionic form
IHg Inorganic mercury
LOD Limit of detection
MeHg Methylmercury
MIA Multiple-injection analysis
n Number
OHg Organic mercury
P Plasma
RBC Red blood cell
Rec. value Recommended value
THg Total mercury
U Urine
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
MB participated in the design of the study, performed the data analyses and drafted the manuscript, BL participated in the development of the analytical method and revision of the manuscript, KAB participated in the design of the study and helped to draft the manuscript, BP carried out the mercury analyses, ÖE participated in the development of the analytical method, and MV participated in the design of the study and the critical revision of the manuscript. All authors read and approved the final manuscript.
Acknowledgements
The participants of this study, and Agneta Åkesson for skilful blood sampling, are gratefully acknowledged.
==== Refs
NRC Council NR Toxicological effects of methylmercury 2000 Washington, DC, National Academy Press 344
WHO (IPCS) IPCS Methylmercury Environmental Health Criteria 1990 101 Geneva, World Health Organization
WHO (IPCS) IPCS Inorganic mercury Environmental Health Criteria 1991 118 Geneva, World Health Organization
Lindberg A Bjornberg KA Vahter M Berglund M Exposure to methylmercury in non-fish-eating people in Sweden Environ Res 2004 96 28 33 15261781 10.1016/j.envres.2003.09.005
WHO Elemental mercury and inorganic mercury compounds: Human health aspects Concise international chemical assessment document (CICAD) 2003 50 Geneva, World Health Organization
Miettinen JK Miller MW and Clarkson TW Absorption and elimination of dietary (Hg2+) and methylmercury in man Mercury, Mercurial, and Mercaptans 1973 Springfield, C.C. Thomas 233 246
Aberg B Ekman L Falk R Greitz U Persson G Snihs JO Metabolism of methyl mercury (203Hg) compounds in man Arch Environ Health 1969 19 478 484 5822994
Rahola T Hattula T Korolainen A Miettinen JK Elimination of free and protein-bound ionic mercury (20Hg2+) in man Ann Clin Res 1973 5 214 219 4203781
Friberg L Nordberg G Miller MW and Clarkson TW Inorganic mercury - a toxicological and epidemiological appraisal Mercury, Mercurials and Mercaptans 1973 Springfield, C.C. Thomas 5
Bornmann G Henke G Alfes H Mollmann H [Intestinal absorption of metallic mercury] Arch Toxikol 1970 26 203 209 5451546 10.1007/BF00578041
Hursh JB Cherian MG Clarkson TW Vostal JJ Mallie RV Clearance of mercury (HG-197, HG-203) vapor inhaled by human subjects Arch Environ Health 1976 31 302 309 999343
Vahter ME Mottet NK Friberg LT Lind SB Charleston JS Burbacher TM Demethylation of methyl mercury in different brain sites of Macaca fascicularis monkeys during long-term subclinical methyl mercury exposure Toxicol Appl Pharmacol 1995 134 273 284 7570604 10.1006/taap.1995.1193
Dock L Rissanen RL Vahter M Demethylation and placental transfer of methyl mercury in the pregnant hamster Toxicology 1994 94 131 142 7801317 10.1016/0300-483X(94)90033-7
Halbach S Clarkson TW Enzymatic oxidation of mercury vapor by erythrocytes Biochim Biophys Acta 1978 523 522 531 656439
Weil M Bressler J Parsons P Bolla K Glass T Schwartz B Blood mercury levels and neurobehavioral function Jama 2005 293 1875 1882 15840862 10.1001/jama.293.15.1875
Grandjean P Weihe P Jorgensen PJ Clarkson T Cernichiari E Videro T Impact of maternal seafood diet on fetal exposure to mercury, selenium, and lead Arch Environ Health 1992 47 185 195 1596101
Grandjean P Weihe P White RF Debes F Araki S Yokoyama K Murata K Sorensen N Dahl R Jorgensen PJ Cognitive deficit in 7-year-old children with prenatal exposure to methylmercury Neurotoxicol Teratol 1997 19 417 428 9392777 10.1016/S0892-0362(97)00097-4
Schober SE Sinks TH Jones RL Bolger PM McDowell M Osterloh J Garrett ES Canady RA Dillon CF Sun Y Joseph CB Mahaffey KR Blood mercury levels in US children and women of childbearing age, 1999-2000 Jama 2003 289 1667 1674 12672735 10.1001/jama.289.13.1667
Kershaw TG Clarkson TW Dhahir PH The relationship between blood levels and dose of methylmercury in man Arch Environ Health 1980 35 28 36 7189107
Sakamoto M Kubota M Matsumoto S Nakano A Akagi H Declining risk of methylmercury exposure to infants during lactation Environ Res 2002 90 185 189 12477463 10.1016/S0013-9351(02)00011-7
Sakamoto M Kubota M Liu XJ Murata K Nakai K Satoh H Maternal and fetal mercury and n-3 polyunsaturated fatty acids as a risk and benefit of fish consumption to fetus Environ Sci Technol 2004 38 3860 3863 15298193 10.1021/es034983m
Skerfving S Mercury in women exposed to methylmercury through fish consumption, and in their newborn babies and breast milk Bull Environ Contam Toxicol 1988 41 475 482 3224165 10.1007/BF02020989
Hallgren CG Hallmans G Jansson JH Marklund SL Huhtasaari F Schutz A Stromberg U Vessby B Skerfving S Markers of high fish intake are associated with decreased risk of a first myocardial infarction Br J Nutr 2001 86 397 404 11570992
Halbach S Welzl G Kremers L Willruth H Mehl A Wack FX Hickel R Greim H Steady-state transfer and depletion kinetics of mercury from amalgam fillings Sci Total Environ 2000 259 13 21 11032131 10.1016/S0048-9697(00)00545-3
Bjorkman L Sandborgh-Englund G Ekstrand J Mercury in saliva and feces after removal of amalgam fillings Toxicol Appl Pharmacol 1997 144 156 162 9169079 10.1006/taap.1997.8128
af Geijersstam E Sandborgh-Englund G Jonsson F Ekstrand J Mercury uptake and kinetics after ingestion of dental amalgam J Dent Res 2001 80 1793 1796 11926235
Cernichiari E Toribara TY Liang L Marsh DO Berlin MW Myers GJ Cox C Shamlaye CF Choisy O Davidson P Clarkson T The biological monitoring of mercury in the Seychelles study Neurotoxicology 1995 16 613 628 8714867
Clarkson TW Hursh JB Sager PR Syversen TLM Clarkson TW, Friberg L, Nordberg GF and Sager PR Mercury Biological Monitoring of Toxic Metals 1988 New York and London, Plenum Press 199 246
Magos L Selective atomic-absorption determination of inorganic mercury and methylmercury in undigested biological samples Analyst 1971 96 847 853 5131698 10.1039/an9719600847
Magos L Clarkson TW Atomic absorption determination of total, inorganic, and organic mercury in blood J Assoc Off Anal Chem 1972 55 966 971 5071876
Bjornberg KA Vahter M Grawe KP Berglund M Methyl mercury exposure in Swedish women with high fish consumption Sci Total Environ 2005 341 45 52 15833240 10.1016/j.scitotenv.2004.09.033
Matsuo N Suzuki T Akagi H Mercury concentration in organs of contemporary Japanese Arch Environ Health 1989 44 298 303 2817960
Bjornberg KA Vahter M Petersson-Grawe K Glynn A Cnattingius S Darnerud PO Atuma S Aune M Becker W Berglund M Methyl mercury and inorganic mercury in Swedish pregnant women and in cord blood: influence of fish consumption Environ Health Perspect 2003 111 637 641 12676628
Mahaffey KR Mergler D Blood levels of total and organic mercury in residents of the upper St. Lawrence River basin, Quebec: association with age, gender, and fish consumption Environ Res 1998 77 104 114 9600803 10.1006/enrs.1998.3834
Morrissette J Takser L St-Amour G Smargiassi A Lafond J Mergler D Temporal variation of blood and hair mercury levels in pregnancy in relation to fish consumption history in a population living along the St. Lawrence River Environ Res 2004 95 363 374 15220070 10.1016/j.envres.2003.12.007
Muckle G Ayotte P Dewailly E Jacobson SW Jacobson JL Determinants of polychlorinated biphenyls and methylmercury exposure in inuit women of childbearing age Environ Health Perspect 2001 109 957 963 11673127
Oskarsson A Schultz A Skerfving S Hallen IP Ohlin B Lagerkvist BJ Total and inorganic mercury in breast milk in relation to fish consumption and amalgam in lactating women Arch Environ Health 1996 51 234 241 8687245
Bergdahl IA Schutz A Hansson GA Automated determination of inorganic mercury in blood after sulfuric acid treatment using cold vapour atomic absorption spectrometry and an inductively heated gold trap Analyst 1995 120 1205 1209 7771681 10.1039/an9952001205
al-Shahristani H Shihab KM Variation of biological half-life of methylmercury in man Arch Environ Health 1974 28 342 344 4829091
Phelps RW Clarkson TW Kershaw TG Wheatley B Interrelationships of blood and hair mercury concentrations in a North American population exposed to methylmercury Arch Environ Health 1980 35 161 168 7189997
Yasutake A Matsumoto M Yamaguchi M Hachiya N Current hair mercury levels in Japanese: survey in five districts Tohoku J Exp Med 2003 199 161 169 12703660 10.1620/tjem.199.161
Hultman P Nielsen JB The effect of dose, gender, and non-H-2 genes in murine mercury-induced autoimmunity J Autoimmun 2001 17 27 37 11488635 10.1006/jaut.2001.0521
|
16202128
|
PMC1262739
|
CC BY
|
2021-01-04 16:36:32
|
no
|
Environ Health. 2005 Oct 3; 4:20
|
utf-8
|
Environ Health
| 2,005 |
10.1186/1476-069X-4-20
|
oa_comm
|
==== Front
Environ HealthEnvironmental Health1476-069XBioMed Central London 1476-069X-4-211621909610.1186/1476-069X-4-21Case ReportThe effects of injection of bovine vaccine into a human digit: a case report O'Neill Jennifer K [email protected] Simon W [email protected] David M [email protected] Marc H [email protected] Orthopaedic Department, Princess Royal Hospital, Lewes Road, Haywards Heath, West Sussex RH16 4EX, UK2005 11 10 2005 4 21 21 27 5 2005 11 10 2005 Copyright © 2005 O'Neill et al; licensee BioMed Central Ltd.2005O'Neill et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The incidence of needlestick injuries in farmers and veterinary surgeons is significant and the consequences of such an injection can be serious.
Case presentation
We report accidental injection of bovine vaccine into the base of the little finger. This resulted in increased pressure in the flexor sheath causing signs and symptoms of ischemia. Amputation of the digit was required despite repeated surgical debridement and decompression.
Conclusion
There have been previous reports of injection of oil-based vaccines into the human hand resulting in granulomatous inflammation or sterile abscess and causing morbidity and tissue loss.
Self-injection with veterinary vaccines is an occupational hazard for farmers and veterinary surgeons. Injection of vaccine into a closed compartment such as the human finger can have serious sequelae including loss of the injected digit. These injuries are not to be underestimated. Early debridement and irrigation of the injected area with decompression is likely to give the best outcome. Frequent review is necessary after the first procedure because repeat operations may be required.
==== Body
Background
Calf diarrhoea is a contagious and often fatal disease of calves. Rotavirus, Coronavirus and E. Coli are three of the most important causal agents.
The combined bovine Rotovirus, Coronavirus and E.coli F5 (K99) vaccine is an inactivated vaccine for calf diarrhoea. It is presented in an oil based emulsion adjuvant to boost the efficacy and duration of the vaccine. Vaccination is carried out by a 2 ml intramuscular injection. A single injection is given to pregnant cows and heifers 3–12 weeks before calving is expected [1].
Injection of this vaccine into a closed compartment such as the human finger can have serious sequelae. The oil-based nature of the vaccine means that it has the potential to cause increased compartmental pressure [2]. The entire finger (especially the finger pulp and tendon sheaths) is at risk. The information leaflet supplied with the vaccine [1] and a WHO report [2] suggests the user is strongly advised to obtain prompt surgical attention after accidental self-injection; early debridement and irrigation of the injected area may be necessary.
Case Presentation
A 20-year-old herdsman accidentally injected his right little finger with the combined bovine vaccine. Approximately 1 ml of vaccine was injected under hand injecting pressure. One hour after injury he attended the Accident and Emergency department and was treated with oral Augmentin and a sling to elevate the hand. Four hours after the injury the patient returned as the finger was increasingly swollen and painful. On examination his finger was swollen, tense and tender. The capillary refill was less than two seconds. Sensation was decreased in the finger when compared to the little finger on the contralateral hand. A needle puncture site was noted at the base of the little finger on the palmar side over the proximal interphalangeal joint skin crease. Radiographs of the hand were normal. The white cell count was 16.2 × 109/l (normal range 4–11 × 109/l) with the differential showing both raised neutrophils at a count of 12.2 × 109/l (normal range 2–7.5 × 109/l) and monocytes at a count of 1 × 109/l (normal range 0.2–0.8 × 109/l).
Surgical decompression was undertaken through a Brunner incision. The flexor sheath was windowed in the A3 pulley. Lipid based fluid was released under high pressure. A further window was made at the A5 level and the flexor sheath was thoroughly irrigated. The skin incisions were left open. The hand was dressed and placed in a back-slab plaster in the 'position of safety'. The hand was elevated and neurovascular observations were performed regularly on the ward. The patient received intravenous Cefuroxine antibiotic. Forty-eight hours later secondary closure of the wound was performed in theatre and the patient was discharged four days after admission with a supply of oral Cephalosporin antibiotics.
Two days later the patient represented with a wound infection. The wound was discharging pus and he had tracking lymphangitis to the axilla. He was apyrexial and his white cell count was within the normal range but his CRP was raised to 11 (normal Value <9). In addition, a non-itchy rash with circular macules had developed over his chest and neck. The rash was thought to be a reaction to the Cephalosporin antibiotic. The arm was elevated in a Bradford sling and the antibiotics were changed to intravenous Flucloxacillin and Benzlypenicillin. The rash appeared to improve. Five days later the lymphangitis had receded and the wound was clean. The patient was discharged on oral Flucloxacillin and Penicillin antibiotics. Microbiology cultures of the fluid removed in theatre, the pus from the wound and blood cultures were all negative for organisms.
The patient was readmitted three days later with recurrence of lymphangitis. Although he was still apyrexial and his white cell count remained within the normal range, his CRP had now risen to 126. The wound was again debrided in theatre and oily white fluid was found in the flexor sheath. The central part of the wound was left open and packed with saline soaked ribbon gauze. Cultures still showed no growth of any organism.
The pain and swelling continued and a week later he was referred to plastic surgery. At operation, there was necrotic tissue around the middle and proximal phalanges of the little finger, extending into the tendons and around the head of the fifth metacarpal. The radial neurovascular bundle could not be located. Amputation of the finger at the metocarpophalangeal joint was performed. No histopathological examination was performed on the specimen but microbiological examination of the specimen did not reveal any pathogenic organisms.
Two months after the amputation the patient was seen at a follow up appointment complaining of some pain in the fifth metacarpal of the right hand a feeling of stiffness of the joints of the right hand. The story was complicated by a recent injury to the hand as he crushed it between a chair and a heavy table three days prior to the appointment. He also had a recurrence of the rash on his chest. A dermatologist opinion diagnosed the rash as pityriasis versicolor due to a malassezia furfur infection on the skin and this was proven by skin biopsies. It was considered by the dermatologist that an atypical mycobacteria infection of the hand (possibly from soil encountered on the farm) might have been missed on routine histopathology and culture. To treat both this possibility and the pityriasis, Tetralysal tablets were started. This treatment was continued for two months and good progress was made with resolution of symptoms.
Discussion
Self injection with veterinary vaccines is an occupational hazard for farmers and veterinary surgeons. A survey of veterinary surgeons reported the rate of needlestick injuries at 5.5/100 vets per year or 1 in 1000 cases of vaccine given [3]. The hand was the site of injury in 17% of cases.
In a closed compartment such as the flexor sheath of a digit injected oil adjuvant based vaccine may have disastrous consequences. In a similar way, high pressure injections of paint [4,5], grease or diesel oil [5,6] or dry cleaning solvents [7] may lead to severe and irreversible loss of function or amputation [8] due to increased pressure within the closed space [8] or subsequent infection [9]. The high pressure injecting devices used in industry [5] and those that deliver a fixed volume per injection [10] are thought to be particularly dangerous but in this case the patient was using a simple 2 ml syringe and 23 guage needle.
Oil based veterinary vaccines have previously been reported to cause a prolonged chronic granulomatous reaction with sterile abscess formation [11] that may result in significant morbidity requiring multiple operations for debridement [11].
Amputation of the dominant thumb at the metacarpophalangeal joint has been reported following an injection of a pig parvovirus vaccine [10] and another patient, following injection of fowl pest vaccine, lost the terminal phalanx of a finger [12]. Other long-term sequelae include neuralgic pain and cold intolerance at the site of the injection. These vaccinations are also delivered at hand pressure and the lack of a high pressure injecting system does not negate serious sequelae of these injections. The outcome appears to be related to the volume injected. These injuries are often under-estimated [9] and delayed diagnosis and debridement may occur. The use of systemic steroids to decrease the swelling is controversial (as it may increase the risk of infection) but should be considered if oedema is significant [7].
Conclusion
Our case involves injection of a small amount of bovine vaccine into the base of the little finger. The swelling and reaction caused by oil trapped within the flexor sheath was likely to be causing compression of the neurovascular supply to the finger. Despite decompression, further ischemia and necrosis of the digit occurred. In addition, there may have been an infection, even though a pathologic agent was never cultured by microbiological tests. Amputation was required. There are no clear guidelines available to guide practicing clinicians in deciding whether to explore and debride such a wound. It has been suggested that injections of a small amount of vaccine can be treated conservatively [4]. However, this case illustrates the seriousness of such an injury despite early surgical treatment. These injuries are not to be underestimated. Early debridement and irrigation of the injected area with decompression is likely to give the best outcome [13]. Frequent review is necessary after the first procedure because repeat operations may be required.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
Jennifer O'Neill carried out the literature searches and is the main author, Simon Richards was involved with writing and editing, David Ricketts was involved with editing and revising the paper and Marc Patterson was involved with editing and revising the paper.
Acknowledgements
No acknowledgements are necessary. The patient has given written consent for publication of this information.
==== Refs
Schering Plough Animal Health information leaflet for Rotavec Corona The combined Bovine Rotavirus, coronavirus and E. Coli F5 (K99) vaccine (inactivated)
Oil based veterinary vaccines WHO Drug Information 1988 2 30
Patterson CJ La Venture M Hurley SS Davis JP Accidental self-inoculation with Mycobacterium paratuberculosis bacterin (Johne's bacterin) by veterinarians in Wisconsin JAVMA 1988 192 1197 9 3391850
Templeman TM Borg DH Kon M Injury of the hand by a high pressure injection: often serious subcutaneous damage Ned Tijdschr Geneeskd 2004 148 2334 8 15587053
Flotre M High-pressure injections of the hand Am Fam Physician 1992 45 2230 4 1575117
Jebson PJ Sanderson M Rao VK Engber WD High-pressure injection injuries of the hand WMJ 1993 92 13 6 8424275
Gutowski KA Chu J Choi M Friedman DW High-pressure hand injection injuries caused by dry cleaning solvents : case reports, review of the literature and treatment guidelines Plast Reconstr Surg 2003 111 174 7 12496578 10.1097/00006534-200301000-00031
Rappold G Rosenmayr E High-pressure injection injuries of the hand. Pathogenesis, problems and therapy Handchir Mikrochir Plast Chir 2001 33 332 41 11600950 10.1055/s-2001-17765
Stoffelen D De Smet L Broos PL Delayed diagnosis of high-pressure injection injuries to the finger. A case report and review of the literature Acta Orthop Belg 1994 60 332 3 7992614
Couzens G Burke FD Veterinary high pressure injection injuries with inoculations for larger animals J Hand Surg 1995 20B 497 499
Jones DP Accidental self inoculation with oil based veterinary vaccines N Z Med J 1996 109 363 5 8890863
Duncan K Accidental self-inoculation with veterinary vaccine Br Med J 1996 312 1436 8664612
O'Neill AC Ismael TS McCann J Regan PJ Fish vaccine injection injuries of the hand Br J Plast Surg 2005 58 547 9 15897041
|
16219096
|
PMC1262740
|
CC BY
|
2021-01-04 16:36:32
|
no
|
Environ Health. 2005 Oct 11; 4:21
|
utf-8
|
Environ Health
| 2,005 |
10.1186/1476-069X-4-21
|
oa_comm
|
==== Front
Emerg Themes EpidemiolEmerging Themes in Epidemiology1742-7622BioMed Central London 1742-7622-2-101619754510.1186/1742-7622-2-10ReviewBiodemographic perspectives for epidemiologists Olshansky S Jay [email protected] Mark [email protected] Jacob [email protected] Bruce A [email protected] School of Public Health, University of Illinois at Chicago, Chicago, Illinois, USA2 Reynolds Department of Geriatric Medicine, University of Oklahoma Health Sciences Center, University of Oklahoma, Oklahoma City, Oklahoma, USA2005 30 9 2005 2 10 10 10 6 2005 30 9 2005 Copyright © 2005 Olshansky et al; licensee BioMed Central Ltd.2005Olshansky et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
A new scientific discipline arose in the late 20th century known as biodemography. When applied to aging, biodemography is the scientific study of common age patterns and causes of death observed among humans and other sexually reproducing species and the biological forces that contribute to them. Biodemography is interdisciplinary, involving a combination of the population sciences and such fields as molecular and evolutionary biology. Researchers in this emerging field have discovered attributes of aging and death in humans that may very well change the way epidemiologists view and study the causes and expression of disease. In this paper, the biodemography of aging is introduced in light of traditional epidemiologic models of disease causation and death.
==== Body
Introduction
The timing with which death occurs and the underlying causes that contribute to it in humans and other sexually reproducing species have been the subject of scientific inquiry for hundreds of years, occupying the minds of scientists from widely ranging scientific disciplines [1]. For example, actuaries (who study humans exclusively) focus on the practical use of death statistics, as in calculating premiums for insurance companies or forecasting survival in order to assess the solvency of age-entitlement programs [2-4]. Demographers have historically worked in much the same way as actuaries, also on a single species, by explaining and understanding the trends, causes, and consequences of mortality within and between subgroups of the population and across time. In modern times, population biologists and entomologists have begun to study the demography of death among species other than humans [5-7]. Unlike demographers, epidemiologists invoke a disease-specific approach that has its historical roots in the study of infectious disease epidemics, with a subsequent focus on chronic diseases and conditions. In contrast to the scientists who work at the level of populations, biogerontologists examine death for humans and other species from a micro perspective as they endeavor to explain and understand changes that occur across time in molecules, cells, tissues, and organs that eventually contribute to disease and death.
The biodemography of aging is a "new" scientific discipline [7-9], arising in the late 20th century as a product of efforts to merge the micro analysis of mortality explored by biologists with the macro analysis of scientists who do research at the population level (e.g., demographers and epidemiologists) in order to provide a biological rationale for the timing and causes of death in humans and other sexually reproducing species.
The earliest antecedents to the modern biodemography of aging date back to the French zoologist Georges Buffon, who demonstrated that "physical laws" regulate the duration of life in humans and other species [10]. These physical laws, according to Buffon, link the biological clocks that govern growth and development to similar clocks that he thought influenced duration of life. British actuary Benjamin Gompertz provided the first mathematical support for this view by demonstrating that for a significant portion of the human lifespan, the risk of death rises exponentially with age [11]. The linkage between Buffon's early life events and Gompertz's regularity in the timing of death appeared many years later with empirical evidence demonstrating that for some species duration of life is calibrated to the onset and length of the reproductive window [12]. These observations serve as the foundation for the biodemography of aging, and might profitably influence the traditional epidemiologic view of disease and death.
An epidemiological view of disease and death
Epidemiology developed from the long history of human experience with infectious disease epidemics. An empirically-based science arising during the 18th and 19th centuries, the driving force of the discipline is an effort to explain and understand in humans, non-random health-related attributes of small and large populations, ranging from the clustering of fevers to age shifts in populations across time [13]. The necessity for more and better data led epidemiologists to forge collaborative ties with policy makers and biostatisticians in order to develop reliable reporting systems for vital events, such as births, deaths, and the identification and distribution of specific diseases. These collaborations with scientists from other disciplines exemplify the important interdisciplinary background of the field [14].
The biology of association (with the exception of genetic heterogeneity) is neither the main focus of epidemiology, nor does it assume great prominence in epidemiologic theory or methods. "Biological plausibility" is widely used in epidemiologic practice as a pragmatic gauge of the possible relevance of observed associations. Epidemiologists strive to understand causal factors (genetic or otherwise) that contribute to age-associated diseases. While epidemiologists have an appreciation of the importance of the aging process in disease expression, aging per se is not a focus of research in the discipline. As such, answers to questions about the lifespan of humans, or any other species for that matter, are outside the usual purview of epidemiology.
During the 20th century, epidemiologists and others in the field of public health contributed to the first leap in human life expectancy, as infectious diseases yielded to improved sanitation, clean water, better diets, and increasingly more insulated living and working environments. Rapid declines in death rates at younger ages led to a redistribution of death from the young to the old, contributing to the accompanying 30-year rise in life expectancy that occurred over the century. Although the rise in life expectancy in the 20th century is a triumph of public health and modern medicine, the price paid for this great success is a developed world dominated by the dual demographic landmarks of population growth and population aging [15], and a shift in underlying causes of death from infectious and parasitic diseases to chronic degenerative diseases expressed at middle and older ages [16]. The field of epidemiology naturally shifted some of its attention away from infectious diseases to the causation and prevention of chronic degenerative diseases.
Although scientific studies based on theoretical and methodological principles of epidemiology have been instrumental in identifying risk factors for chronic degenerative diseases (e.g., smoking, fatty foods, obesity, stress, etc., and their effects on heart disease, stroke, and many cancers) [17,18], the relative effects of these interventions on life expectancy at birth are much smaller than those that resulted from the control of infectious diseases early in life [19]. This occurs, in part, because the number of person-years-of-life (PYL) added to the life table when a child is saved from death is almost always considerably greater than the PYL added when an older person has their life extended. In addition, while preventive measures have been shown to reduce death rates at middle and older ages among population subgroups that adhere to healthy behavioral practices [20], most of the population has not adopted these practices. Diminishing increases in life expectancy at birth in developed nations today, despite the intensive efforts by epidemiologists and other public health experts to modulate lifestyles, present a paradox that can also benefit from a biodemographic examination of aging and disease.
A biodemographic view of disease and death
Before examining epidemiology from a biodemographic perspective, it is important to understand the basic theoretical rationale supporting this new paradigm. The modern notion of the biodemography of aging arose in the early 1990s, as scientists from a broad range of scientific backgrounds began speculating about the biological forces responsible for the similar age patterns of death that they observed among sexually-reproducing species. The theoretical basis for the biodemography of aging is derived principally from evolutionary biology [21-24], but it has roots in historical efforts of scientists who speculated on what was referred to as a "law of mortality" [11,25-30] – an observation that a consistent age pattern of death is known to occur across species [12,31-34]. Modern biodemographers have theorized that sexually-reproducing species, including humans, experience a common age pattern of death because the aging of individuals, and by implication the aging of populations, is calibrated to life history traits whose own evolution was unrelated to either aging or duration of life – namely, the biology and timing of reproduction [8,12].
The link between death and reproduction is based on the evolutionary principle that the force of natural selection begins to decline rapidly once reproduction commences, approaching negligible levels at the end of the reproductive window (i.e., at menopause) [34]. The force of natural selection refers to the ability of selection to influence the distribution and frequency of alleles in the population – a force inherently linked to reproduction. As selection wanes, alleles with adverse health consequences expressed at progressively older ages can accumulate in the gene pool [21-24,35,36]. Evolution is blind to the health consequences of genes expressed in older regions of the lifespan because there is no selection to act upon them once they have been propagated. Empirical tests of this hypothesis have shown that the age trajectory of death is, in fact, a species-specific phenomenon that, as predicted from evolution theory, is calibrated to the onset and length of a species' reproductive window [12]. The reproductive window, in turn, is a genetically determined attribute that is established as part of a life history strategy that has been molded by the environment within which each species evolved. What this biodemographic perspective has taught us thus far is that for each species there is a fundamental link between the timing of sexual maturation, the length of the reproductive window, and the rate of increase in the death rate from biological causes of death. Thus, there is biology in the life table as originally anticipated by Benjamin Gompertz – a biology that Buffon [10] speculated on in the 18th century, and which is driven by evolutionary forces that operate through genetic mechanisms.
Natural selection, the very heart of Darwin's theory of evolution, was based on the biological consequences of departures from a norm. Darwin's observations about imperfections in the morphological characteristics of living things were inconsistent with the works of an intelligent designer, which led him to the idea that all forms of life have an evolutionary history based on change over time. The life history details such as growth, development, and maturation that emerge from this unique evolutionary history of every species are central to understanding the variations that exist between individuals and species (including humans) in the temporal distribution of age-determined diseases and the timing of death [37]. Finally, the implications that the biodemography of aging has for the degree to which chronic degenerative diseases can be influenced by risk factor modification may be of even greater relevance to epidemiologists – the goal of chronic disease epidemiology.
An implied perfection of the human body is a persistent concept that emerged from the major World religions and has appeared repeatedly in legends from almost every culture dating back to antiquity [38]. These images of perfection describe a distant past when humans were either immortal or extremely long-lived. The most common historical explanation for the loss of immortality, the lack of perfect health, and the steady decline in human longevity has been that each new generation has adopted increasingly more decadent lifestyles. Roger Bacon, an influential English philosopher and scientist of the 13th century, was the first to popularize this view [39]. However, he also believed that the trend toward shorter lifespans could be reversed by invoking the "secret arts" of the past – namely, the adoption of more austere lifestyles and the ingestion of foods and other substances believed to have life-extending properties. Thus, the perspective that aging and diseases are amenable to modification through changes in lifestyles, has its origins in thinking that extends back in time at least one thousand years.
This persistent belief in perfection and the consequences of a departure from it has spawned two other beliefs that continue to have a significant philosophical and practical influence on contemporary scientific views of mortality. The most important of these is the belief that aging and diseases are unnatural and are, therefore, somehow avoidable. The second is the notion that the health and longevity consequences associated with perfection can be reclaimed through human actions. Elements of this latter idea contribute to modern epidemiologic thought. These beliefs, and the quest for longer lives that arises from them, have been obsessions throughout human history [40], having become a central part of the paradigm of modern medicine and the effort of epidemiologists to understand how risk factors alter death rates.
In modern times, aging is described by some as a disease that can be reversed, slowed, or even eliminated by changes in lifestyles or by ingesting vitamins, minerals, anti-oxidants, and hormones – modern versions of the anti-aging remedies of the past [41,42]. On the surface, this philosophy of personal empowerment is seductive. However, a biodemographic perspective, which is based in part on an examination of the anatomical structures and functions of the human body, raises doubts about the validity of this perspective. For example, from a biodemographic perspective it is suggested that aging and many of the diseases that accompany it are not deviant departures from perfection, or even the sole consequence of moderately decadent lifestyles. Instead, they are primarily the consequence of operating our bodies beyond their biological warranty period [43] (i.e., beyond the time when parents can contribute to the reproductive fitness of their own offspring). Thus, the philosophy that people are empowered to control their own disease, aging, and longevity has become, in modern times, an ideology of personal blame. In effect, it has become common to blame people for many of the diseases and disorders that they experience as they age, and more importantly, some may be led to believe that aging and the diseases that accompany it are largely avoidable. From a biodemographic perspective, it is certainly true that some diseases and disorders are entirely preventable and that aging and death can be hastened by imprudent lifestyles, but once these harmful lifestyles are avoided, most of what is commonly recognized as aging and disease is an inevitable by-product of operating the machinery of life. Even though aging, disease, and death are not programmed into our genes, once the engine of life switches on, aging is inevitable.
Herein lies the link between epidemiology and biodemography. Modern epidemiology arose out of an infectious disease paradigm where treatments and prevention were shown to dramatically reduce the incidence and prevalence of communicable diseases. Once the epidemiologic transition from high to low mortality permitted most people to survive beyond the end of their reproductive window, chronic degenerative diseases appeared with rising frequency. Similar epidemiologic approaches to disease modification were then applied to these diseases expressed mostly at middle and older ages. Although the results of this effort have often been dramatically successful (e.g., established linkages between smoking and cancer; obesity and diabetes; and high blood pressure and stroke), the resulting behavioral modifications have not led to another quantum leap in life expectancy like that observed during the 20th century. Instead, what occurred was a steady decline in death rates accompanied by a diminishing rise in life expectancy at birth and at older ages. The phenomenon of diminishing longevity gains from lifestyle modification is referred to here as the Medawarian Paradox, named after Sir Peter Medawar [23] who suggested that aging is "...revealed and made manifest only by the most unnatural experiment of prolonging an animal's life by sheltering it from the hazards of its ordinary existence" (p.13).
The dramatic increase in the last century in the number of people living for seven decades or more [44] has revealed an entirely new set of "weak links" in the structure and function of the human body that are associated with living well beyond our reproductive years [45,46]. Extended survival into the post-reproductive period permits a number of anatomical and physiological features of the human body to reveal themselves as debilitating diseases and disorders such as Alzheimer's disease and osteoporosis. These weak links were not commonly known or thought of as such in the past because they were uncommon – people rarely lived long enough to experience them. We define them as "weak links" now because of the Medawarian Paradox of the unusual circumstance of survival into older ages.
It is important to emphasize that the concept of anatomical oddities and weak links in humans and other living things is not new. The idea began when Charles Darwin suggested that imperfections in the design and functioning of parts of living things are inevitable by-products of natural selection's blind eye to body design. According to evolution theory, selection does not operate with any particular goal in mind; it simply optimizes the perpetuation of DNA across time by constructing bodies capable of carrying the DNA and passing it successfully from one generation to the next. This idea has since appeared several times in the published literature in the 20th century, including the first detailed presentation of morphology and body design [47], in later publications by evolutionary theorists [48-51], and in modern discussion of reliability theory [33]. However, neither Darwin nor those who followed in his footsteps ever examined the morphology of living things from the perspective of an aging animal.
Epidemiology of degenerative diseases in an aging world
The theoretical and methodological basis for epidemiology arose out of a communicable disease model where a notion of avoiding and curing the common fatal diseases of the time was prevalent and enormously successful [35]. As a measure of that success, half of the gain in life expectancy at birth in the 20th century was achieved by 1920 – largely as a result of public health interventions. As life expectancy increased, chronic diseases became the overwhelming cause of death, with diseases of the heart, cancer, and cerebrovascular disease accounting for 70 percent of all mortality. These diseases do not have a single cause, but usually result from a complex web of causation. The concept of risk factors was introduced to facilitate understanding and prevention. Thus, smoking was shown to increase the risk of heart disease and cancers. Fatty foods, obesity, and lack of exercise clearly predisposed people to these conditions [52]. However, frustration arose over the years, as it was shown that these risk factors accounted at best for only half of the deaths from these three causes. It is now realized that such phenomena as environment, psychosocial factors, and genetics also play important roles [53].
In general, control of chronic diseases has been effective enough to produce a gradual increase in life expectancy, albeit at a diminishing rate. In developed countries, where life expectancy is about 75 years for males and 80 years for females, the classic indicators of health are becoming less useful. Infant mortality has been very low for more than 20 years and the concept of "premature death" is no longer a sensitive health indicator [54]. Instead, with each succeeding year lived at the tip of the exponential curve of mortality, numerous detrimental physiologic mechanisms and morbidities are activated. Although one role of the epidemiologist is to improve quality of life, often at best what can be accomplished is slowing the progression of diseases and conditions expressed in older regions of the lifespan by suppressing their symptoms. These include Alzheimer's disease and related dementias, vision and hearing loss, hip fracture, osteoporosis and osteoarthritis, incontinence, depression, social isolation, widowhood, and institutionalization [55]. In an aging world where the envelope of human survival is continuing to be extended, a biodemographic perspective leads to the realization that the expression of disease becomes more a by-product of extended survival rather than the end product of identifiable and modifiable risk factors.
Conclusion
When the focus of epidemiology shifted from communicable diseases to chronic degenerative diseases, the process of aging intruded into the traditional epidemiologic model that links behaviors to mortality risks and life expectancy determination. Now that human survival has been progressively extended deeper into the post-reproductive period of the lifespan, a biodemographic perspective reveals biochemical and biomechanical forces that have a profound influence on population frailty, disease expression, and duration of life. Biochemical constraints involve the inevitable accumulation of damage that occurs at all levels of biological organization, including the maintenance and repair processes themselves – a loss of biological fidelity that most biogerontologists suggest cannot as yet be modulated [44,56]. Biomechanical forces involve the progressive and currently immutable diminishment of structure and function in the very morphological features that give species their phylogenetic identity [40,46,49,50]. Anticipated advances in research (e.g., caloric restriction mimetics, embryonic stem cells, and progress in the replacement of body parts) will continue to erode the biochemical and biomechanical barriers to longevity and quality of life.
An important message in this paper is that biology reminds us that evolutionary success does not require living to old age, it only requires living long enough to reproduce. Our bodies fail over time not because they were designed to fall victim to aging and disease at a predetermined age [57], or even because of the acquisition of risk factors and decadent lifestyles, but because they were not designed for extended operation. The diseases and disorders we experience in the post-reproductive period of the lifespan are, therefore, not flaws from an evolutionary perspective, and unless proven otherwise, should not be attributed to personal failure or exclusive by-products of environmental risk factors. The biological consequences of aging are crucial factors for epidemiologists, whose concepts and methods for the pursuit of specific causes and risk factors are not entirely applicable to animals living long enough to experience the Medawarian Paradox.
Those of us alive today are the most recent recipients of an evolutionary legacy that includes a human body filled with both awe-inspiring complexity, and a host of anatomical oddities and weak links that are revealed with the passage of time. The diseases and disorders that arise from the negative side of this legacy are unfortunate and unanticipated by-products of the human ingenuity that has allowed our bodies to be operated far longer than nature has historically permitted. Had natural selection operated with a particular goal in mind, such as a healthy old age, there is reason to believe that the morphological structures and biochemical makeup of sexually-reproducing species would probably be far different from what is currently the case [46]. It is important to remember that in an aging world the expression of disease and how long we live as both individuals and populations is more a product of evolutionary neglect, not evolutionary intent. As such, traditional epidemiologic models should become more sensitive to the unique biological forces that come into play in a world where aging is common and biochemical and biomechanical forces have an important influence on disease expression and length of life.
Competing interests
The author(s) declare that they have no competing interests.
Acknowledgements
Funding for this work was provided by the National Institutes of Health/National Institute on Aging for S. Jay Olshansky (AG13698-01) and Bruce A. Carnes (AG00894-01).
==== Refs
Olshansky SJ Carnes BA Ever since Gompertz Demography 1997 34 1 15 9074828
Bennett NG Olshansky SJ Forecasting U.S. age structure and the future of social security: The impact of adjustments to official mortality schedules Pop Dev Rev 1996 22 F703 27
Lee RD Long-run global population forecasts: A critical appraisal Pop Dev Rev 2000 16 44 71
Lee RD Carter L Modeling and forecasting U.S. mortality J Am Stat Assoc 1992 87 659 671
Carey JR Liedo P Orozco D Vaupel JW Slowing of mortality rates at older ages in large medfly cohorts Science 1992 258 457 61 1411540
Curtsinger JW Hidenor HF Townsend DR Vaupel JW Demography of genotypes: Failure of the limited life-span paradigm in Drosophila melanogaster Science 1992 258 461 3 1411541
Wachter KW Finch CE Eds Between Zeus and the Salmon: The Biodemography of Longevity 1997 Washington, D.C.: National Academy Press
Carnes BA Olshansky SJ Evolutionary perspectives on human senescence Pop Dev Rev 1993 19 793 806
Olshansky SJ Carnes BA Demographic perspectives on human senescence Pop Dev Rev 1994 20 57 80
Buffon GLL Histoire Naturelle générale et particuliére 1747
Gompertz B On the nature of the function expressive of the law of human mortality and on a new mode of determining life contingencies Phil Trans Royal Soc of London 1825 115 513 85
Carnes BA Olshansky SJ Grahn D Continuing the search for a law of mortality Pop Dev Rev 1996 22 231 64
Beaglehole R Bonita R Kjellstrom T Basic epidemiology 1993 World Health Organization, Geneva 3 4
Mausner JS Kramer S Epidemiology-an introductory text Population Dynamics and Health 1985 Sanders, Philadelphia 239 262
Olshansky SJ Carnes BA Cassel C The aging of the human species Sci Am 1993 46 52 8446881
Omran AR The epidemiologic transition: A theory of the epidemiology of population change Mil Mem Fund Q 1971 49 509 38
Stallones RA The rise and fall of ischemic heart disease Sci Am 1982 243 53 59 7423181
Stamler S Stamler R Neaton JD Low risk-factor profile and long-term cardiovascular mortality and noncardiovascular mortality and life expectancy JAMA 1999 282 2012 18 10591383 10.1001/jama.282.21.2012
Olshansky SJ Carnes BA Cassel C In search of Methuselah: Estimating the upper limits to human longevity Science 1990 250 634 40 2237414
Manton KG Stallard E Tolley D Limits to human life expectancy Pop Dev Rev 1991 17 603 37
Hamilton WD The moulding of senescence by natural selection J Theor Bio 1966 12 12 45 6015424 10.1016/0022-5193(66)90184-6
Kirkwood TBL Evolution of aging Nature 1977 270 301 04 593350 10.1038/270301a0
Medawar PB An unsolved problem in biology 1951 London: Lewis
Williams GC Pleiotropy, natural selection and the evolution of senescence Evolution 1957 11 398 411
Brody S The kinetics of senescence J Gen Phys 1924 6 245 57 10.1085/jgp.6.3.245
Brownlee J Notes on the biology of a life-table J Royal Stat Soc 1919 82 34 77
Gompertz B On one uniform law of mortality from birth to extreme old age, and on the law of sickness J Inst Act 1872 16 329 44
Greenwood M Laws of mortality from the biological point of view J Hygiene 1928 28 267 94
Loeb J Northrop JH Is there a temperature coefficient for the duration of life? Proc Nat Acad Sci 1916 2 456 7 16586628
Makeham WM On the law of mortality J Instit Act 1867 13 325 58
Deevey ES Jr Life tables for natural populations of animals Q Rev Biol 1947 22 283 314 10.1086/395888
Finch CE Pike MC Witten M Slow mortality rate accelerations during aging in some animals approximate that of humans Science 1990 249 902 5 2392680
Gavrilov LA Gavrilova JS The biology of life span: A Quantitative Approach 1991
Charlesworth B Evolution in age-structured populations 1994 Cambridge: Cambridge University Press
Brody JA Prospects for and ageing population Nature 1985 315 463 6 3158823 10.1038/315463a0
Charlesworth B Williamson JA The probability of the survival of a mutant gene in an age-structured population and implications for the evolution of life-histories Genetical Research 1975 26 1 10 1218732
Carnes and Olshansky Heterogeneity and its biodemographic implications for longevity and mortality Experimental Gerontology 2001 36 419 430 11250115 10.1016/S0531-5565(00)00254-0
Grumman G A history of ideas about the prolongation of life Trans Am Phil Soc 1966 56 1 102
Bacon R Richard Browne ML The cure of old age, and preservation of youth 1683 College of Medicine, London
Olshansky SJ Carnes BA The quest for immortality: Science at the frontiers of aging 2001 Norton: New York
Chopra D Grow younger, live longer: 10 steps to reverse aging 2001 Harmony Books: New York
Klatz R Grow young with HGH: The amazing medically proven plan to reverse aging 1998 Harper Perennial Library
Carnes BA Olshansky SJ Grahn D Biological Evidence for Limits to the Duration of Life Biogerontology 2003 4 31 45 12652187 10.1023/A:1022425317536
Kannisto V Lauritsen J Thatcher AR Vaupel JW Reductions in mortality at advanced ages: Several decades of evidence from 27 countries Pop Dev Rev 1994 20 793 810
Olshansky SJ Carnes BA Grahn D Confronting the boundaries of human longevity Am Sci 1998 86 52 61 10.1511/1998.1.52
Olshansky SJ Carnes BA Butler R If Humans Were Built to Last 2001 284 Scientific American 50 55 11234506
Thompson DW On Growth and Form 1942 Cambridge University Press
Dawkins R The blind watchmaker 1996 W.W. Norton: New York. Switzerland: Harwood Academic Publishers
Gould SJ The Panda's thumb 1992 W.W. Norton: New York
Morgan E The scars of evolution: What our bodies tell us about human origins 1994 Oxford University Press
Nesse R Williams G Why We Get Sick: The new science of Darwinian medicine 1996 Vintage Books: New York
Olshansky SJ Passaro D Hershow R Layden J Carnes BA Brody J Hayflick L Butler RN Allison DB Ludwig DS A potential decline in life expectancy in the United States in the 21st century New England Journal of Medicine 2005 352 1103 1110 15784663 10.1056/NEJMsr043743
Brody JA Spicker SF, Ingman SR, Lawson IR Aging and dependency in the 21st Century Ethical dimensions of geriatric care 1987 Holland/Boston: D.Reidel Publishing Co 3 21
Schopper D Pereira J Torres A Estimating the burden of disease in one Swiss Canton: what do disability adjusted life years (DALY) tell us? Int J Epidemiol 2000 29 871 7 11034971 10.1093/ije/29.5.871
Brody JA Grant MD Age-associated diseases and conditions: Implications for decreasing late life morbidity Aging Clin Exp Res 2001 13 64 7
Finch CE Kirkwood TBL Chance, Development, and Aging 2000 Oxford:Oxford University Press
Hayflick L The Future of Ageing Nature 2000 408 267 70 11089985 10.1038/35041709
|
16197545
|
PMC1262741
|
CC BY
|
2021-01-04 16:38:27
|
no
|
Emerg Themes Epidemiol. 2005 Sep 30; 2:10
|
utf-8
|
Emerg Themes Epidemiol
| 2,005 |
10.1186/1742-7622-2-10
|
oa_comm
|
==== Front
Harm Reduct JHarm Reduction Journal1477-7517BioMed Central London 1477-7517-2-131614657710.1186/1477-7517-2-13CommentaryDrug use and harm reduction in Afghanistan Todd Catherine S [email protected] Naqibullah [email protected] Steffanie A [email protected] Division of International Health & Cross-Cultural Medicine, Department of Family & Preventive Medicine, University of California, San Diego, 9500 Gilman Drive, 0622 La Jolla, CA, USA, 92093-06222 National HIV/AIDS Control Program, Ministry of Public Health, Massoud Road, Kabul, Afghanistan2005 7 9 2005 2 13 13 23 11 2004 7 9 2005 Copyright © 2005 Todd et al; licensee BioMed Central Ltd.2005Todd et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Opium has been cultivated in Afghanistan since 1100 A.D., although production has steadily increased since 1979. Currently, Afghanistan produces three-quarters of the global opium supply, with injection drug use and HIV currently following the opium trade route through Central Asia. Although systematic studies are lacking, heroin use appears to be on the rise in Afghanistan. The purpose of this paper is to briefly provide historical background and current statistics for drug production and use in Afghanistan, to discuss the new government's policies towards problem drug use and available rehabilitation programs, and to assess Afghan harm reduction needs with consideration of regional trends.
==== Body
Introduction
Afghanistan is at a cross-roads; the country is emerging from more than twenty years of political and social unrest as the leading global producer of opium in a geographic region widely affected by drug use, particularly injection drug use, and blood-borne infections, including human immunodeficiency virus (HIV). Countries bordering Afghanistan (with the exception of Turkmenistan, for which there is no available data) are experiencing concentrated epidemics of HIV and hepatitis C in IDU populations [1-4]. Afghanistan is currently at risk for these potentially destabilizing events. Historically, countries slow to respond or instituting only punitive measures for ascending rates of drug use have experienced dramatic outbreaks of HIV and hepatitis among injection drug users (IDU), often with diffusion into the general population [5-7]. The rationale for this paper is to examine the current situation and policy of Afghanistan, as little is known about substance abuse in this country. We will briefly provide historical background and current statistics for drug production and use in Afghanistan, present the new government's policies towards problem drug use and available rehabilitation programs, and compare the situation in Afghanistan to that of the surrounding geographic region, much of which is experiencing the most rapid increase of HIV cases due to injection drug use.
Opium History in Afghanistan
We will focus on opium, the substance with greatest impact on risk of blood borne infections in Afghanistan. Information was obtained from electronic searches through PubMed and Google, with additional information obtained through site-specific searches, such as United Nations Office of Drugs and Crime (UNODC). Selected search words were: opium, Afghanistan, trafficking, Central Asia, and heroin. While we have chosen to focus only on opium, the same routes for trafficking opium are used to transport both other illicit substances, such as cannabis/hashish (also produced in Afghanistan) and amphetamines and licit drugs of abuse, such as pharmaceutical compounds (e.g. benzodiazepines, opioid analgesics), and volatile inhalants.
Afghanistan, along with Pakistan and Iran, form the Golden Crescent, an area known for opium and cannabis cultivation and trafficking from the time poppies were introduced from Europe by Arab traders along the Silk Road [8]. Opium production in Afghanistan did not reach large scale until the Russian invasion in 1979. The growth in production was attributed to direct loss of government controls on production and indirect market demand created by decreased production due to political disruption in Vietnam and Laos, formerly the chief suppliers to Europe and North America [9,10]. By this time, Iran had significantly decreased opium production due to blockage of trade routes and severe punishment for drug-related convictions by the new theocratic regime [10]. Restrictions on cultivation and refining in Pakistan in the mid to late 1990's led to the shift of these activities to Afghanistan, resulting in the creation of new trade routes into Pakistan and Central Asia [11]. Opium cultivation was further encouraged by warlord commanders in constant conflict with each other following the Russian retreat in 1989. These commanders required economic support for military actions in response to loss of United States funding to Najibullah's government in 1991 [12]. During the mujaheddin [freedom fighters] era, opium and heroin production rose steadily with Afghanistan becoming the leading global supplier, overtaking Burma in the mid-1990's [13]. Since this time, either Afghanistan or Burma have contributed the greatest percentage to the world's opium market, with Afghanistan being the single largest country producer for the last four years.
The rise of the Taliban regime was marked with steadily increasing opium production, despite their pledge to "cleanse Afghanistan of the poisoned poppies" [14]. Increased opium production has been attributed to economic realities faced by the Taliban, who received little external donor support due to international sanctions. The Taliban charged a 10% tax to opium farmers, netting $20 million or more each year, and controlled the opiate trade, with confiscated boxes bearing the words, "Not for use by Muslims" [15,16]. However, in 2000, the Taliban, now controlling the majority of Afghanistan, banned opium cultivation and enforced harsh punitive measures against drug use, which included maiming the hands of drug users. These steps, as well as severe drought in Afghanistan, were highly effective in reducing the amount of available opiates in the world market, resulting in drug shortages in Europe and a ten-fold increase in price [13,17]. Some believe the move was economically motivated to increase price, but this will remain open to debate as the Taliban were deposed in 2001 [18,19].
Within the year following removal of the Taliban regime, opium production recovered to near-record levels, with 3400 and 3600 metric tons produced in 2002 and 2003, respectively. In 2003, total income to opium farmers alone was equal to half of the legal gross domestic product and illustrated that, despite Hamid Karzai's declaration of a jihad [holy war] on opium, regional commanders continue to rely on opium production and trafficking to maintain their strongholds [13]. Opium cultivation has been revived in southern provinces and introduced in eastern and northern Afghan provinces, likely due to economic consideration as it is at least twelve times more profitable than wheat [13,20,21]. In 2004, a UNODC survey was performed to assess opium production within Afghanistan [21]. The study reports opium cultivation in all provinces with 2.9% of all arable land devoted to this purpose, though as much as 29% is cultivated in some provinces. The estimated crop for 2004 would have exceeded the record set in 1999 had drought and other plant stressors not compromised crop yields [21]. Despite these losses, Afghanistan produced 87% of the global opium supply last year; this supply increase may be impacting price as gross income per cultivated hectare decreased 64% and gross family income among opium farmers decreased 56%, based on study interviews [21]. The price per kilogram has decreased in all markets, though prices are markedly different between provinces with lowest prices noted in the northeastern areas [21]. UNODC posits that the declining price may also be due to declining quality (reduced opium content per gram in irrigated fields), competitive lower prices in Tajikistan, and the small number of traders that control the market.
The UNODC survey used satellite images as well as photos and GPS coordinates covering 16% of all arable land in 10 provinces; a survey of farmers was also performed by sampling approximately 8% of all villages in 21 provinces, representing 19% of the total cultivation area [21]. This is the largest study performed to date for estimation of opium cultivation in Afghanistan; however, significant regional differences may not have been adequately assessed in areas under-sampled by the survey. These estimates resulted in a wide confidence interval (109,000–152,000 hectares), though would still represent a 36% increase in cultivation at lowest estimate. Additionally, while opium production is believed to have increased, the study states that production is based on robust estimates as obtaining objective evidence on a crop that is not openly traded is not possible.
This increase in production and the portent of further production increases, indicated by the increasing number of farmers and hectares, has lead Antonio Maria Costa, the executive director of UNODC, to state that,
"Afghan annals will record 2004 as contradictory. Political progress towards democracy culminated in the near plebiscite election of President Karzai. For this splendid accomplishment we all salute President Karzai's courage and determination. Yet, opium cultivation, which has spread like wildfire throughout the country, could ultimately incinerate everything – democracy, reconstruction and stability."[21].
Current Opium Laws
As the government of Afghanistan develops, laws concerning opium production and use have been the subject of multiple decrees, often with external influence. The United Nations Security Council Resolution and the Bonn Agreement of 2001 stated that the new government of Afghanistan should respect international obligations and cooperate with the international community in the fight against terrorism, drugs and organized crime [22]. In 2002, Hamid Karzai, at the time the appointed interim leader of the Transitional Islamic State of Afghanistan (TISA), issued decrees banning cultivation, production, drug abuse and trafficking of narcotic drugs, and the simultaneous implementation of an eradication campaign by the government [22].
Use of opium products is illegal in Afghanistan; conviction results in a three-month prison sentence.
Opium Use in Afghanistan
Historically, opium has been used in Afghan communities as medication for different conditions, particularly pain and respiratory complaints. Opium use also has a traditional role in the societies of some groups [23]. There are few national estimates of opium use in Afghanistan; the highest regional use is noted in northeastern Badakhshan Province along the Tajik border, with 20–30% of the local population estimated to be addicted. High use rates have also been reported in districts of Herat and Farah Provinces [23]. In February 2001, UNODC conducted a study in five remote districts of four provinces. The estimated total adult population of these five districts (Khak-e-Jabar, Azro, Hesarak, Gardez, and Sayed Karam) is 120,000 people. According to key informants, there were at least 694 opium users, 164 heroin users, 8514 hashish users and 2556 persons using recreational pharmaceuticals [24]. However, because the interviews were with a limited number of drug users and key informants, these figures are only approximations; there is no official drug user registry in Afghanistan.
Recreational opium use appears to be common in Kabul, based on data from a recent study conducted by UNODC, interviewing 100 key informants and 200 drug users [25]. There are estimated to be at least 6,026 heroin users, 10,257 opium users, 26,415 hashish users, 15,526 pharmaceutical drugs addicts and 8,128 alcohol addicts within Kabul. However, due to the small numbers of drug users interviewed and inherent biases introduced from interview of key informants, these numbers are believed to represent conservative estimates. There are no reports for the number of drug users in other urban areas.
Although heroin is predominately used by men, multiple sources document opiate use starting in childhood and affecting both genders [24,25]. Based on these studies, the Counter Narcotic Department (CND), the highest drug control authority under the presidential office, estimates that there are approximately 500,000 people within Afghanistan addicted to different psychoactive substances (Personal Communication, Dr. M. Zafar, Drug Demand Reduction Officer, CND, October 29, 2004).
Heroin is easily accessible in Afghanistan and there is a disturbing trend towards injection of heroin alone and in combination with other substances, linked to returning refugees importing behaviors from other countries where injection use is common [25,26]. According to a drug user in Kabul: "Drugs are like vegetables here. Very cheap and infinitely available"[24]. In Kabul, single use doses of opium cost about 20–50 Afghanis ($0.50–1.00 US) whereas a typical dose of heroin costs about 40–50 Afghanis ($1 U.S.) [26]. However, prices are not stable and change with the seasonal availability of opium and heroin in the local market. Pharmaceutical opiates and other psychoactive substances can be easily obtained from the estimated 15,000 registered pharmacies or many unregistered pharmacies. People can obtain different psychoactive drugs, sedatives, pain killers and narcotics without a prescription and in unlimited quantities [26]. As in Pakistan and India, some pharmacies are reputed to sell buprenorphine (Temgesic) and some addicts report using it, though there is no documented evidence [12,27]. Needles and other injection paraphernalia are available over the counter, but their cost may be prohibitive to drug users who are most often unemployed. Pharmacies are likely to continue as a common source of drugs since the Ministry of Public Health (MOPH) does not currently have the capacity to monitor pharmacies.
Although problem drug use appears to be increasing in Afghanistan, addiction treatment remains limited. Medical services are provided to addicts through both public and private sectors, which, together, are not able to meet the demand for services. In the public sector, the National Mental Health Institutes, under direction of the MOPH, have functioning treatment and rehabilitation centers in several Afghan cities. The center in Kabul (Mental Health Institute) has only 30 treatment slots. (personal communication, Dr. Khaitab Khakar, Director, MoPH Kabul Mental Health Institute, June 30, 2005) In a few provinces, there are branches of the Mental Health Institute providing out-patient services, such as counseling, but these do not have an in-patient facility.
The private sector also has limited treatment resources, with only two non-government organizations (NGO) currently providing in-patient services. The Nejat Center has ten treatment beds and two outreach teams in each of their Kabul and Badakhshan locations. According to the Nejat Center director, Dr. Tareq Suleyman,"We have the capacity to treat just 20 addicts a month but we have 3,000 people on the waiting list "[28]. Between 2001 and 2003, 4335 drug addicts have been treated, with 956 treated at the Kabul Mental Health Institute and 1308 at the Nejat Center [28]. Another NGO, Welfare Association for Afghanistan (WADAN), has a fifteen bed facility for drug addicts in Gardez, Paktiya Province. The standard of care for rehabilitation in Afghanistan is a fifteen day in-patient stay, followed by continued counseling via outreach counselors in the home or return visits to the outpatient department. Methadone treatment has not yet been introduced, though several groups agree that substitution therapy is needed in this setting.
No data is available on relapse due to lack of a reliable, functioning follow-up system. Human resources are scarce for harm reduction activities, like drug demand reduction and rehabilitation, due to lack of trained staff and a severe shortage of female health workers and counselors. There are currently a small number needle exchange programs in Kabul, orchestrated through Zindagi Nawin drug counseling programs. (Personal communication, Dr. M. Ilyas Azami, German Technical Cooperation, August 16, 2005) NGO activities involved in harm reduction education are limited, with the majority of their activities conducted in Kabul city, though counseling and prevention activities are being conducted by Nejat in Kabul and German Technical Cooperation (GTZ) with NGO partners SHRO (Herat), Wadan, (Gardez and Kandahar) and KOR in Kabul and Faizabad.
Regional Opium Use and Influential Trends
The experiences and influence of other countries in the region are an important consideration for predicting future harm reduction needs and blood-borne infection rates in Afghanistan. Larger supplies of heroin are anticipated to be available in Afghanistan as production increases and spillover from new trafficking routes threatens to affect a larger number of people by reaching remote areas of the country.
Data for heroin production within Afghanistan is based on border seizures. Central Asian countries, particularly Tajikistan, are reporting record amounts of drug seized, with the disturbing trend of drug transition from opium to heroin as early as 2001 [16,29]. Security has increased at the Iranian border as a part of that country's response to rising drug use and violence associated with trafficking, but the heroin demand continues in Iran, driving trafficking activity [2,29]. Additionally, trafficking has increased to Central Asia and Pakistan, with the risks of transporting blood-borne pathogens intrinsic to trafficking activities [29]. Traffickers routinely test the quality of the substance with the dealer/distributor in the next country, often sharing injection equipment. These activities allow transmission of infection from areas of presumed higher prevalence to Afghanistan and could initiate or fuel the final component of the cycle related to heroin.
The concern for transmission of blood borne viruses in this context cannot be minimized. Both hepatitis B and C have measurable documented prevalence in injection drug users (IDUs) and the general populations of bordering countries Pakistan and Uzbekistan [3,30-34]. In Pakistan, hepatitis C prevalence ranges from 5.3 to 7% in the general population, [30-32] 22% in non-injecting heroin users,[34] and 89% in IDUs [3]. Rising prevalence of hepatitis B and C due to injection drug use have been noted in other Central Asian Republics [30,35]. Central and South Asia are experiencing a rapid increase in HIV cases introduced by injection drug use and the commercial sex trade [7,16,36-38]. The HIV prevalence among IDUs in neighboring countries is largely unknown. Recently, prevalences of 29.8% and 12.1% were reported among intravenous drug users in Dushanbe, Tajikistan and Tashkent, Uzekistan respectively; of all HIV cases in Iran, 65% are among IDUs [1,39,40]. Injection drug use appears to be increasing in Afghanistan, raising concerns that a concentrated epidemic of HIV will ensue, as IDU and HIV have been documented to follow overland heroin trafficking routes [6,19,41].
The epidemic of injection drug use in Central Asia has been attributed to the poor socioeconomic conditions and proximity to opium trafficking routes [42]. These factors may contribute to the increasing number of IDU in Afghanistan. However, Afghanistan has several other characteristics predisposing its populace to drug addiction and transition to injecting use. Previous studies have documented that refugees are at increased risk to adopt drug use, largely due to poor economic indicators and psychological changes leading to increased risky behavior [43,44]. An estimated 3.5 million Afghans have repatriated within the last four years, of whom a significant proportion remain internally displaced [45]. Two recent studies suggest importation of learned drug use and other risk behaviors by this vulnerable population [34,46]. New behaviors learned by Afghan refugees in Pakistan, and, to a lesser degree, Iran and the Central Asian Republics, where rates of both injection drug use and blood-borne infections are quickly rising, may be impacting drug use patterns [29,37]. Afghans may be disproportionately at risk for blood-borne infections resulting from injection drug use as displaced Afghan drug users exhibited less knowledge regarding HIV transmission and engage in high-risk behavior with greater frequency when compared to Pakistani drug users. A study done among IDU in Quetta, Pakistan revealed that, of 143 Afghans surveyed, none used condoms, only 4% had ever heard of HIV/AIDS, 18% injected drugs, and of those, 72% reported needle sharing, all of which displayed a significantly greater degree of risk than their Pakistani counterparts. Additionally, 41% of Afghan drug users stated they had engaged the services of commercial sex workers [46]. There have been efforts to increase awareness of blood-borne infection transmission among vulnerable groups in Kabul city by several non-government organizations, including ORA, Nejat Center, and GTZ, as well as by the Ministry of Public Health and the National HIV/AIDS Control Program (NACP). The outreach workers affiliated with these programs have established rapport with several marginalized risk groups, predominantly drug users. Preliminary findings from an on-going study of blood-borne infection prevalence among injection drug users in Kabul indicates that, of 67 surveyed, the majority report not sharing "works" and purchasing single use syringes from the pharmacy daily (cost 3 Afghanis = US$0.06). However, another study surveying high-risk and sentinel population groups in Kabul, Heart, Mazar-i-Sharif, and Kandahar notes that only approximately 40% of those surveyed, including drug users, had ever heard of HIV/AIDS. (Personal communication, John Foran, ActionAid Afghanistan, August 16, 2005) Prevention messages have also been disseminated to the general population. The NACP has engaged the religious community in dialogue about the risks of HIV to Afghanistan and their role in community preventive education in a particularly noteworthy program.
There have been few changes in the number or content of rehabilitation programs in Kabul city, though some NGOs wish to initiate substitution therapy following procurement of funding. (Personal communication, Wayne Bazant, German Technical Cooperation, July 6, 2005) UNODC is currently conducting a country-wide assessment of drug use, which may also provide compelling evidence for increasing both the available number and therapeutic options of rehabilitation programs. Additional in-depth studies of risky behavior, particularly before and after the introduction of a harm reduction program, would provide meaningful data.
Conclusion
Although Afghanistan is a major producer of heroin, injection drug use appears to be a relatively new phenomenon. Greater numbers of heroin users have been observed following the end of the Taliban regime and the return of Afghan refugees from neighboring countries [23]. Although few studies are available, high risk behaviors have been documented among Afghan IDUs along with low HIV/AIDS awareness and virtually no condom use [46]. The growing number of injection drug users, the availability of heroin, and small, geographically-limited number of harm reduction and drug treatment programs in Afghanistan place the country at great risk for epidemics of blood-borne infection. Further research on blood-borne infection risk behaviors and seroprevalence among drug users in Afghanistan would be helpful to better describe the current situation. Funding of programs to broaden education programs on HIV/AIDS and viral hepatitis, harm reduction, and drug treatment services should be an urgent priority.
Statement of Competing interests
The author(s) declare they have no competing interests.
Authors' contributions
CT researched and wrote the section on the history of opium cultivation and use in Afghanistan as well as the section on injection drug use trends in Central Asia. NS researched and wrote the section on the current Afghan situation, including law, government policy, and treatment services available. SS researched and wrote the summary statements and contributed to the section on regional influence. All authors read and approved the final manuscript.
==== Refs
United Nations Development Programme (UNDP) Iran: HIV/AIDS and Intravenous Drug Usage. YOUANDAIDS: the HIV/AIDS portal for Asia Pacific
IRIN News Tajikistan: Drug use, migration and ignorance fuel rise in HIV infections. UN Office for the Coordination of Humanitarian Affairs. IRINews.org February 24, 2004
United Nations Office for Drug Control and Crime Prevention (UN-ODCCP) and UNAIDS Baseline Study of the Relationship between injection drug use, HIV and hepatitis C among male injection drug users in Lahore. United Nations Office on Drugs and Crime
United Nations Childrens Fund (UNICEF) Turkmenistan: Statistics
Panda S Chatterjee A Bhattacharya SK Manna B Singh PN Sarkar S Naik TN Chakrabarti S Detels R Transmission of HIV from injecting drug users to their wives in India Int J STD AIDS 2000 11 468 473 10919490 10.1258/0956462001916137
Beyrer C Razak MH Lisam K Chen J Lui W Yu XF Overland heroin trafficking routes and HIV-1 spread in south and south-east Asia AIDS 2000 14 75 83 10714570 10.1097/00002030-200001070-00009
Rhodes T Ball A Stimson GV Kobyshcha Y Fitch C Pokrovsky V Bezruchenko-Novachuk M Burrows D Renton A Andrushchak L HIV infection associated with drug injecting in the newly independent states, eastern Europe: the social and economic context of epidemics Addiction 1999 94 1323 36 10615718 10.1046/j.1360-0443.1999.94913235.x
Chouvy PA Levinson D, Christensen K Golden Crescent Encyclopedia of Modern Asia 2002 1 Chicago: Chas Scribner & Sons 441
Girardet E In Afghanistan, drug trade is blooming Christian Science Monitor December 28, 1988
MacDonald S Afghanistan's Drug Trade Society 1992 29 61 67
Drug Enforcement Agency Drug Intelligence Brief. Pakistan: Country Brief. DEA Resources for Law Enforcement Officers
Reid G Costigan G eds Revisiting 'The Hidden Epidemic'-A Situation Assessment of Drug Use in Asia in the Context of HIV/AIDS 2002 Australia: The Centre for Harm Reduction, The Burnet Institute 20 26 12138788
United Nations Office for Drugs and Crime (UNODC) Afghanistan: Opium Survey 2003. UNODC-Crop Monitoring
Dynes M Holy army bolsters heroin trade. The Times London October 2, 1996
Peuche JC Central Asia: Charges Link Russian Military to Drug Trade. Radio Free Europe/Radio Liberty June 8, 2001
International Crisis Group Central Asia: Drugs and Conflict Asia Report No 25, Osh/Brussels 2001
Farrell G Thorne J Where have all the flowers gone?: evaluation of the Taleban crackdown against opium poppy cultivation in Afghanistan Int J Drug Policy
Staff Writer Afghanistan's opium fiends The Economist February 24, 2001
Beyrer C Human immunodeficiency virus (HIV) infection rates and drug trafficking: fearful symmetries Bull Narcotics 2002 54 103 16
UNODC, Counter-Narcotics Directorate, Government of Afghanistan Afghanistan: Farmers Intentions Survey Report
UNODC, Counter-Narcotics Directorate, Government of Afghanistan Afghanistan: Opium Survey 2004
ReliefWeb.org The situation in Afghanistan and its implications for international peace and security
UNDCP Community Drug Profile # 4: An assessment of the drug use in rural Afghanistan- the Great Azro Initiative target districts 2001 UNDCP Afghanistan Office, PanGraphics (Pvt) Ltd. Islamabad
UNODC Community Drug Profile #1: Problem Drug Use in Afghan Communities: An Initial Assessment UNDCP Program, Islamabad, 1999
UNODC Community Drug Profile #5: An assessment of problem drug use in Kabul city
IRIN-United Nations Office for Coordination of Humanitarian Affairs Bitter-Sweet Harvest: Afghanistan's New War Accessed August 5, 2005
Strathdee SA Zafar T Brahmbhatt H Baksh A ul Hassan S Rise in needle sharing among injection drug users in Pakistan during the Afghanistan war Drug Alcohol Depend 2003 71 17 24 12821202 10.1016/S0376-8716(03)00072-3
One World Afghan carpet weavers: Junkies from birth. E-Ariana: Today's Afghan News
UNODCCP Illicit drug situation in the regions neighboring Afghanistan and response of ODCCP
Ruzibakiev R Kato H Ueda R Yuldasheva N Hegay T Avazova D Kurbanov F Zalalieva M Tuichiev L Achundjanov B Mizokami M Risk factors and seroprevalence of hepatitis B virus, hepatitis C virus, and human immunodeficiency virus infection in Uzbekistan Intervirology 2001 44 327 32 11805437 10.1159/000050066
Khan AJ Luby SP Fikree F Karim A Obaid S Dellawala S Mirza S Malik T Fisher-Hoch S McCormick JB Unsafe injections and the transmission of hepatitis B and C in a periurban community in Pakistan Bull World Health Organ 2000 78 956 63 10994278
Luby SP Qamruddin K Shah AA Omair A Pahsa O Khan AJ McCormick JB Hoodbhouy F Fisher-Hoch S The relationship between therapeutic injections and high prevalence of hepatitis C infection in Hafizabad, Pakistan Epidemiol Infect 1997 119 349 56 9440439 10.1017/S0950268897007899
Khokhar N Gill ML Malik GJ General seroprevalence of hepatitis C and hepatitis B virus infections in population J Coll Physicians Surg Pak 2004 14 534 6 15353136
Kuo I ul-Hasan S Zafar T Galai N Ghanzafar I Sherman SG Strathdee SA Prevalence of HIV, HCV, and changes in drug supply and cost related to transition to injection among injection drug users (IDUs) in Pakistan 15th International AIDS Conference, July 11–16, 2004, Bangkok
Illiev SKh Gaipova MB Karmanova GA The epidemiological characteristics of HIV infection in Turkmenistan Zh Mikrobiol Epidemiol Immunobiol 1999 1 19 21 10096188
Parfitt T Drug addiction and HIV infection on rise in Tajikistan Lancet 2003 362 1206 14570036 10.1016/S0140-6736(03)14560-6
Joint United Nations Program on HIV/AIDS (UNAIDS) Report on the global AIDS epidemic. Geneva 2004
CDC/MMWR The global HIV and AIDS Epidemic, 2001 Morb Mortal Wkly Rep 2001 50 434 9
Sanchez JL Todd CS Bautista CT Botros B Khakimov MM Giyasova GM Yakubov SK Abdulaeva M Saad MD Graham RR Carr JK Earhart KC HIV prevalence and risk factors among injecting drug users in Tashkent, Uzbekistan, 2003–04 Drug Alcohol Depend
Stachowiak JA Tichonova F Strathdee SA Stibich MA Mogilnii V Beyrer C Marked ethnic differences in HIV prevalence and risk behaviors among injecting drug users in Dushanbe, Tajikistan, 2004 Drug Alcohol Depend
Westermeyer J The importance and difficulty of drug research in developing countries: a report from Kabul as timely reminder Addiction 2004 99 802 804 15200575 10.1111/j.1360-0443.2004.00777.x
UNDP HIV/AIDS in Central and Eastern Europe and the Commonwealth of Independent States: reversing the epidemic, facts and policy options 2004 RENESANS, Bratislava, Slovak Republic 20 24
Westermeyer J Lyfoung T Westermeyer M Neider J Opium addiction among Indochinese refugees in the United States: characteristics of addicts and their opium use Am J Drug Alcohol Abuse 1991 17 267 77 1928021
Carballo M Puvacic S Zeric D Implications of complex emergencies, uprooting and forced migration on risk of HIV/AIDS: The case of Bosnia and Herzegovina XII World AIDS Conference, Geneva, June 28–July 3, 1998
United Nations High Commissioner for Refugees (UNHCR) UNHCR's global refugee figure lowest since 1980 – Internal displacement, statelessness remain high June 17, 2005
Zafar T Brahmbhatt H ul Hassan S Strathdee SA A comparison of HIV knowledge and risk behaviors among Afghani and Pakistani drug users in Quetta, Pakistan J Acquir Immune Defic Syndr 2003 32 394 8 12640197
|
16146577
|
PMC1262742
|
CC BY
|
2021-01-04 16:36:49
|
no
|
Harm Reduct J. 2005 Sep 7; 2:13
|
utf-8
|
Harm Reduct J
| 2,005 |
10.1186/1477-7517-2-13
|
oa_comm
|
==== Front
Harm Reduct JHarm Reduction Journal1477-7517BioMed Central London 1477-7517-2-141616229710.1186/1477-7517-2-14CommentaryThe US war on harm reduction: fixing policy on intelligence and facts Wodak Alex [email protected] Director, Alcohol and Drug Service, St Vincent's Hospital, Sydney, Australia2 Conjoint Senior Lecturer, School of Public Health and Community Medicine, School of Medicine, The University of New South Wales, Sydney, Australia2005 15 9 2005 2 14 14 19 7 2005 15 9 2005 Copyright © 2005 Wodak; licensee BioMed Central Ltd.2005Wodak; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
==== Body
At a two-day private meeting in Tokyo in June 2005, some of Japan's most senior politicians and powerbrokers met to consider the steadily expanding HIV/AIDS epidemic. AIDS has recently become a matter of increasing concern in Japan following an HIV epidemic in several major Japanese cities among Japanese men having sex with men at sex-on-premises venues. The Japanese elites at the Tokyo meeting were shocked to learn that the United States has by far the highest annual AIDS incidence among OECD countries at 15/100,000 [1]. Spain, with an annual AIDS incidence of 3.3/100,000, has the second highest rate among industrialized countries, while Australia was well down the ranking with an incidence only one tenth that of the United States at 1.5/100,000 [1].
The pragmatic Japanese were stunned to learn that the high AIDS incidence in the United States was no accident: abstinence-only rather than explicit, peer-based sex education and tokenistic, rather than early and vigorous, needle syringe programmes have produced the expected public health outcomes. In 2002, needle syringe programmes in the United States actually declined from the previous year, exchanging and distributing 25 million needles and syringes [2] for a total population of about 290 million. In contrast, Australia, with a population of 20 million, exchanged and distributed 32 million needles and syringes in 1998/99 [3]. As Randy Shiltz recorded in 'And The Band Played On' [4], from the outset the United States responded to the greatest global public health crisis of the last half millennium with consistent and breath-taking denial. President Reagan failed to make any public comment on HIV/AIDS for the first six years of the epidemic. Three Presidents later, little has changed. It's still business as usual despite the United States failing to meet government declared targets for reducing the number of new HIV infections.
The HIV/AIDS epidemic was officially recognized almost a quarter century ago. In that time, the intelligence and facts about prevention of HIV have become well established. What is less well appreciated is the pivotal importance of political leadership in translating the intelligence and facts about prevention into evidence-based programmes established in time and on a scale commensurate with control of the epidemic. It was political leadership in Uganda, Thailand and Cambodia which changed the trajectory of three epidemics based on high partner change heterosexual activity. Political leadership in Australia, in partnership with community activists, clinicians and researchers, tamed the early HIV epidemic among men who have sex with men and averted an epidemic among injecting drug users. Likewise, the political leadership provided from within the Thatcher government in Great Britain ensured that an HIV epidemic among and from injecting drug users was averted by pragmatism. But political leadership in the United States of America not only deprived the citizens of that country of the benefits of evidence-based HIV prevention, but also actively exported these failed policies to other countries. Nowhere has this been clearer than in any HIV prevention policy or programme linked to injecting drug users.
At the very same time as the meeting discussed above took place in Tokyo, 22 nations attended the Programme Coordinating Board of UNAIDS in Geneva to finalise a policy paper on HIV prevention [5]. The United States had insisted during the twelve month development of the document that phrases such as 'harm reduction' and 'needle syringe programmes' must be excluded. In the June 2005 Geneva meeting, the Indian delegation noted that India and the United States of America were the world's two largest democracies and asked the delegation of the United States to respect the weight of world opinion: none of the 21 other countries supported the position of the United States. After two days of difficult discussion, the United States grudgingly allowed these (and other similarly pragmatic) phrases to be included.
Only three months earlier, at the United Nations Commission on Narcotic Drugs meeting in Vienna in March 2005, a similar debate took place. On that occasion, the United States, with the support of Japan and Russia, was able to hold out its abstinence-only position against 17 other countries who wanted the CND document to explicitly support harm reduction.
In the last few years, most of the major countries in Asia have come to realize that harm reduction policies and programmes are critical to control of HIV among and from injecting drug users. China, Vietnam, Malaysia, Indonesia, Burma and Taiwan are all now traveling down the same road. They all started as zealous supporters of a law enforcement dominated approach to drugs, and are all now moving to a more pragmatic and evidence based public health approach in which HIV control can be achieved. Methadone and needle syringe programmes are planned or already being established in these countries. Contradictions between the new harm reduction approach and the former law enforcement dominated approach are being recognized and dealt with. Thailand is now isolated as the last major Asian country to still support a scorched earth War on Drugs.
The exceptionalism of the United States, discussed since the time of Alexis de Tocqueville, has been increasing in recent years, especially since the election in 2000 of President George W Bush [6]. The United States of America is becoming increasingly isolated, not only among other developed countries, but also in the developing world.
On May 1, 2005, The Sunday Times in England published a leaked document [7] which is accepted as the official minutes of a meeting held at 10 Downing Street on 23 July 2002 to enable 'C' (Sir Richard Dearlove), then head of MI 6, to report to the British Prime Minister and his senior Cabinet colleagues and major government officials. The subject was a briefing 'C' had just received in Washington from George Tenet, then head of the CIA, regarding the forthcoming invasion of Iraq. Among the astonishing revelations in these minutes is the comment by 'C' 'but the intelligence and facts were being fixed around the policy.' The inescapable conclusion from reading these minutes is that Tenet advised 'C' that the intelligence and facts on Iraq were being adjusted in the United States to justify the decision to invade Iraq. While the revelations in these minutes have surprised and shocked many experienced foreign policy commentators, observers of the war on drugs have known for decades that 'fixing the intelligence and the facts on the policy' has been both the very basis and the central flaw of the War on Drugs.
In the lead up to the June 2005 meeting of the Programme Coordinating Board of UNAIDS, over 130 diverse individuals and organisations in the United States of America wrote (see Additional file 2) [8] on May 10, 2005 to Ambassador Randall Tobias, Coordinator of United States Government Activities To Combat HIV/AIDS Globally, 'to express our concern about recent reports that US officials have questioned the efficacy of needle exchange programs and sought to block support for needle exchange in United Nations resolutions and policy documents'. Emphasizing the importance of HIV infection among and from injecting drug users in the United States of America and globally, they noted that 'no fewer than seven federally-funded reviews and reports conducted by public health officials, researchers and US government agencies have concluded that syringe exchange programs are effective, safe and cost effective'. Recent public support for the science of needle syringe programmes was cited including endorsements from the Director of the National Institutes of Health, the Director of the National Institute on Drug Abuse and a recent World Health Organization (WHO) report which stated that the available data 'present a compelling case that needle and syringe programs substantially and cost effectively reduce the spread of HIV among injection drug users and do so without evidence of exacerbating injecting drug use at either the individual or societal level.'
In response, 35 individuals from United States' War on Drugs organizations wrote to Ambassador Randall Tobias on May 25, 2005 (see Additional file 2) [9]. This group lists only six people with medical or other degrees. These 35 individuals claimed to be a 'diverse group of citizens and organizations' who were 'better informed on prevention, intervention and treatment of addiction than any other source'. They urged Ambassador Tobias to 'continue to promote and defend the United States' position against the disease-promoting practices of needle and syringe giveaways'. Although making the remarkable claim that ' [needle syringe] programs are ineffective or, at best, weakly effective at deterring the spread of HIV', no evidence was offered to support this or any of the other propositions offered to Ambassador Tobias.
Fixing 'the intelligence and facts on the policy' has trapped the United States of America into a military quagmire in Iraq and contributed to looming economic problems resulting from the twin current account and Federal budget deficits. Fixing 'the intelligence and facts on the policy' for illicit drugs, ensured tragic health, social and economic consequences for the United States of America [10]. Extending this approach to HIV has magnified these tragic costs. But time is running out: exporting to other countries a failed and futile policy on the twin epidemics of HIV and illicit drugs will soon be a thing of the past. More and more, countries want to fix their drugs and HIV policy on intelligence and facts rather than the other way round.
Supplementary Material
Additional File 2
Ambassador Randall Tobias-Zero Tolerance file.doc A response from 35 individuals and organisations urging Ambassador Tobias to continue to promote and defend the United States' position in a document expressing UNAIDS HIV prevention policy.
Click here for file
Additional File 1
Ambassador Randall Tobias HR.doc Letter to Ambassador Randall Tobias expressing concern about recent reports that US officials have questioned the efficacy of needle exchange programs and sought to block support for needle exchange in a document expressing UNAIDS HIV prevention policy.
Click here for file
Acknowledgements
Grateful thanks to Mr. Benjamin Phillips for his help in identifying and obtaining relevant literature and editorial assistance with the manuscript.
==== Refs
National Centre for HIV Epidemiology and Clinical Research Annual Report 2004 accessed July 18th 2005
Update: Syringe Exchange Programs – United States MMWR 2002 54 673 676 Weekly July 15, 2005. accessed July 18th 2005
Health Outcomes International, the National Centre for HIV Epidemiology and Clinical Research, Drummond M: Return on Investment in Needle and Syringe Programs in Australia 2002 Australian Government Department of Health and Ageing accessed July 18th 2005
Shilts R And the Band Played On: Politics, People, and the AIDS Epidemic 1987 New York, St. Martin's Press
UNAIDS Programme Coordinating Board (PCB) accessed July 18th 2005
The Economist. A nation apart November 6, 2003. accessed July 18th 2005
Danner M The Secret Way to War 52 New York Review of Books 9 June, 2005. accessed July 18th 2005
Letter to Ambassador Randall Tobias, Office of the United States Global AIDS Coordinator May 10, 2005
Letter to Ambassador Randall Tobias, Coordinator of United States Government Activities To Combat HIV/AIDS Globally May 25, 2005
Drucker E Drug prohibition and public health: 25 years of evidence Public Health Reports 1999 114 15 19 accessed July 18th 2005 10.1093/phr/114.1.14
|
16162297
|
PMC1262743
|
CC BY
|
2021-01-04 16:36:49
|
no
|
Harm Reduct J. 2005 Sep 15; 2:14
|
utf-8
|
Harm Reduct J
| 2,005 |
10.1186/1477-7517-2-14
|
oa_comm
|
==== Front
Harm Reduct JHarm Reduction Journal1477-7517BioMed Central London 1477-7517-2-181620214510.1186/1477-7517-2-18ResearchSurvey of Australians using cannabis for medical purposes Swift Wendy [email protected] Peter [email protected] Paul [email protected] National Drug and Alcohol Research Centre, University of New South Wales, Sydney, NSW, 2052 Australia2005 4 10 2005 2 18 18 17 8 2005 4 10 2005 Copyright © 2005 Swift et al; licensee BioMed Central Ltd.2005Swift et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The New South Wales State Government recently proposed a trial of the medical use of cannabis. Australians who currently use cannabis medicinally do so illegally and without assurances of quality control. Given the dearth of local information on this issue, this study explored the experiences of medical cannabis users.
Methods
Australian adults who had used cannabis for medical purposes were recruited using media stories. A total of 147 respondents were screened by phone and anonymous questionnaires were mailed, to be returned by postage paid envelope.
Results
Data were available for 128 participants. Long term and regular medical cannabis use was frequently reported for multiple medical conditions including chronic pain (57%), depression (56%), arthritis (35%), persistent nausea (27%) and weight loss (26%). Cannabis was perceived to provide "great relief" overall (86%), and substantial relief of specific symptoms such as pain, nausea and insomnia. It was also typically perceived as superior to other medications in terms of undesirable effects, and the extent of relief provided. However, nearly one half (41%) experienced conditions or symptoms that were not helped by its use. The most prevalent concerns related to its illegality. Participants reported strong support for their use from clinicians and family. There was almost universal interest (89%) in participating in a clinical trial of medical cannabis, and strong support (79%) for investigating alternative delivery methods.
Conclusion
Australian medical cannabis users are risking legal ramifications, but consistent with users elsewhere, claim moderate to substantial benefits from its use in the management of their medical condition. In addition to strong public support, medical cannabis users show strong interest in clinical cannabis research, including the investigation of alternative delivery methods.
==== Body
Background
While cannabis has long been part of folk pharmacopeia, there is a burgeoning body of research on its therapeutic potential. This has largely drawn on scientific advances in our understanding of the pharmacology of cannabis, and its complex interactions with the central nervous system, particularly endogenous brain reward pathways [1]. In addition to basic experimental research, case reports, surveys of people using cannabis for medical conditions and prospective clinical trials of cannabis-based medicines are consolidating the evidence that cannabis may play a role in the management of some medical conditions. Authoritative reviews of this evidence indicate that cannabis has therapeutic potential for conditions such as HIV- and cancer-related wasting, nausea and vomiting resulting from chemotherapy, neurological disorders such as multiple sclerosis and chronic pain [1-4].
While current research reveals exciting therapeutic opportunities, there is an ongoing debate about the virtues of obtaining such benefits from the complex chemical cocktail contained in the whole plant or from one or more components isolated and developed into a synthetic pharmaceutical product. This debate cross-cuts important issues such as the difficulties of reliable dosing when using the natural product, whether the potential harms of smoking cannabis due to its ease of titration overshadow its therapeutic benefits, and whether different medical conditions will respond more favourably to the whole plant or to different constituents in isolation or combination. However, underlying these issues is the reality that most people who use cannabis medicinally do so by using black market supplies of an illicit drug.
As with the opiates, evaluations of the therapeutic potential of cannabis occur in the context of a vigorous political debate on the use of an illicit drug with dependence potential for medicinal purposes. This situation is clearly evident in the United States, where there is an ongoing legal challenge by the Federal Government over the States' rights to allow cannabis to be used by registered medical users. Despite Canada's recent decision to provide a controlled supply of natural cannabis to registered users, and approvals for the marketing of Sativex, a pharmaceutical cannabis extract, in some countries, currently most users would rely on home-grown cannabis, or supplies obtained from friends, families, dealers and medical compassion clubs.
To date, there has been little interest in Australia in formally investigating the therapeutic potential of cannabis or investigating the practices of current medical users. In 1999 the NSW State Government commissioned a Working Party to investigate the issue and recommend research and legislative options. Among their recommendations were: controlled clinical trials of cannabis, investigations into delivery methods other than smoking, surveys of current medical cannabis users and legislative amendments to allow compassionate use [4]. Subsequently, in 2003 the NSW Government announced it would conduct clinical trials, but despite generating significant publicity, there has been no further commitment by the NSW Government on this issue. The 2004 National Drug Strategy Household Survey found widespread public support for medical cannabis use, with 68% supporting a change in legislation to permit use for medical purposes and 74% supporting a clinical trial of medicinal cannabis use [5]. It is not known how many people use cannabis for medicinal purposes in Australia. Those who do use it engage in an illegal behaviour and risk arrest. Those that rely on black market supplies use a product of unknown source and quality.
Several surveys in the US, UK, Germany and Canada [6-12] have reported perceived improvements in a variety of medical conditions following cannabis use. However, we know very little about the experiences of Australian users, and how they compare to findings in other studies. These authors are aware of only two unpublished Australian studies conducted in northern NSW; in 1998 a survey of 202 users recruited at the Nimbin HEMP Embassy [13], and in 2003 a survey of 48 members of a medical cannabis information service [14].
This paper presents the results of a study of 128 users, which aimed to learn more about their patterns of use, experiences and concerns, and interest in participating in a medical cannabis trial.
Methods
Sample
The sample comprised 128 people who used cannabis for medical purposes. To be eligible for the study, participants had to be living in Australia and to be currently using/have previously used cannabis for medical purposes. While the study targeted residents of Australia's most populous state, NSW (pop: approximately 6.7 million), we did not exclude participants from other parts of Australia (total pop: approximately 20 million).
As it is not known how many Australians use cannabis for medical purposes it was not possible to obtain a representative sample of such users. As this was an exploratory study to see who responded to a general call for participation in the survey, we did not target groups representing people with specific medical conditions (e.g., HIV/AIDS, multiple sclerosis) or hospital departments known to treat patients who may benefit (e.g., oncology, chronic pain clinics). Participants were primarily recruited from opportunistic media stories between November 2003 and August 2004, in newspapers, on radio and television. In addition, the Medical Cannabis Information Service (MCIS) in Nimbin, NSW, offered to tell its members about the survey and the International Association for Cannabis as Medicine (IACM), in Germany, placed the questionnaire on its website.
A total of 147 enquiries were received between December 2003 and August 2004 by telephone and email and approximately 170 questionnaires distributed (some people requested multiple copies to distribute). For example, the media stories generated enquiries from several GPs who said they would inform certain patients of the study. Of the 131 questionnaires returned, 128 were used for analysis (75% of questionnaires sent out). Of the three discarded questionnaires, one respondent was a recreational cannabis user and two had never used cannabis.
Questionnaire
The survey comprised an anonymous mail-out questionnaire, adapted from one developed by the MCIS in a recent study of its members [14]. Several issues were covered, including medical conditions/symptoms experienced, patterns of medical cannabis use, symptom relief and effects of use, comparison of cannabis to other medications, source and legal concerns (e.g., arrest), other concerns over use, opinion of family, friends and medical personnel, and interest in participating in a cannabis trial. The final version incorporated comments from researchers and clinicians interested in this issue.
Procedure
The study received ethics approval from the University of New South Wales Social/Health Human Research Ethics Advisory (HREA) Panel. Interested persons were screened for eligibility over the phone and informed of the purpose of the survey; assurances of anonymity and confidentiality were provided. Questionnaires were mailed to participants, completed anonymously and returned in a stamped, self-addressed envelope. Addresses were destroyed when the questionnaire was posted.
Analyses
Data were entered into SPSS (Version 12.0.1). As this was an exploratory study with a small sample size, this paper reports descriptive statistics only. Percentages are presented for categorical data; means (for normally distributed) and medians (for skewed data) are presented for continuous data. While data are usually presented on the overall sample, gender and age differences are presented for some variables, where they are of interest.
Results
Demographics
The sample was 63% male. Participants had a median age of 45 yrs (range 24–88), with almost one third (31%) aged 50 years or over, and one in ten (9%) aged 60 years plus. While the study targeted NSW residents (who represented 58% of participants), responses came from across Australia, especially Queensland (15%) and Victoria (12%). Residents of other States and Territories each comprised less than 3% of participants.
Participants reported a wide range of medical conditions and symptoms associated in the literature with the use of medicinal cannabis (Table 1), most commonly chronic pain (53%) and arthritis (38%). Approximately one in five reported migraine (22%), weight loss (21%) and persistent nausea (20%). However, depression was the most commonly reported condition/symptom (60%). Up to 35 other conditions/symptoms were listed, most commonly post traumatic stress disorder (PTSD) (5%) and irritable bowel syndrome (4%). It is important to note that we did not ask participants to distinguish between primary symptoms/conditions for which they sought treatment (e.g., cancer) and conditions which may have been secondary to this (e.g., depression) or consequent to treatment (e.g., chronic nausea). Multiple conditions (mean = 3.7, SD = 2.1, range = 1–10), of lengthy duration, were the norm, with three quarters (84%) reporting more than one condition and two thirds (67%) at least three conditions. Congruent with this picture, cannabis was used to relieve multiple symptoms (median = 3, range = 1–12), especially chronic pain (57%), depression (56%), arthritis (35%), persistent nausea (27%) and weight loss (26%).
Table 1 Conditions/symptoms experienced, duration, and conditions/symptoms requiring cannabis relief (n = 128).
Condition (%) with condition Median duration (yrs) % used cannabis for relief of..*
Depression 60 10 56
Chronic pain 53 10 57
Arthritis 38 9 35
Migraine 22 18 17
Weight loss 21 4 26
Persistent nausea 20 6 27
Spinal cord injury 14 11 13
Spasms (spasticity) 13 8 16
Fibromyalgia 13 13 13
Wasting 13 5 11
ME (chronic fatigue) 13 16 13
Neuralgia/neuropathy 12 8 12
HIV/AIDS 9 15 8
Multiple sclerosis 7 9 7
Cancer 6 10 4
Other neurological disorder 6 5 6
PTSD 5 13 1 person
Irritable bowel syndrome 4 10 1 person
Glaucoma 3 29 2
*These figures do not necessarily equate with the % reporting a particular condition because some people reported using cannabis to relieve the particular symptoms (e.g., chronic pain, nausea) associated with a condition, rather than citing they used cannabis to relieve the condition itself (e.g., arthritis, cancer).
Patterns of medical cannabis use
Participants had first tried cannabis for medical purposes at a median age of 31 years (range = 14–77). More than one quarter (29%) had discovered its therapeutic benefits as a spin-off from recreational use; others had tried it following concerns about the side-effects of their medications (14%), or a belief their medications or treatment were ineffective (13%), or had acted on the recommendation of a medical practitioner (10%) or friend (10%).
Table 2 presents data on patterns of medical use. Most (85%) were currently using cannabis therapeutically, even if sporadically. For those who had stopped, the main reasons were: their inability to obtain a regular supply (9/19 people), its illegality (7/19), cost (7/19) and disliking the side effects or route of use (each 3/19). Of those using intermittently, many reported their use would be more regular if it were more readily availability and cheaper.
Table 2 Patterns of medical cannabis use (n = 128 unless specified)
Total (%) Male (%) Female (%)
Current use 85 86 83
Length of use
<1 year 12 9 17
1–5 yrs 27 23 35
6–10 yrs 20 26 10
11–15 yrs 9 10 8
16–20 yrs 10 10 10
>20 yrs 21 23 19
Frequency of use (n = 126)
several times a day 39 45 29
6–7 days/wk 24 19 31
1–5 days/wk 14 14 13
less than weekly 2 3 2
very seldom 2 1 2
as required 20 18 23
Method(s) of use (n = 127)
eaten as cooked recipe 49 48 50
drunk as tea 7 8 6
smoked as cigarette (joint) 65 58 77
smoked as dry pipe (chillum) 24 28 19
smoked as water pipe (bong) 54 58 46
vaporiser 8 11 2
eaten as leaf/flower matter 3 4 2
Most helpful method of use (n = 126)
eaten as cooked recipe 16 15 17
drunk as tea 2 3 2
smoked as cigarette (joint) 31 26 40
smoked as dry pipe (chillum) 10 13 4
smoked as water pipe (bong) 33 36 29
vaporiser 2 3 2
other 6 5 6
Medical use was typically long-term and regular. Use of less than one year was uncommon (12%), with more than half (61%) having used it for at least six years; one in five reported very long-term use (more than 20 years). Most used at least weekly (75%), and more than half (59%) used almost daily or daily. Approximately one in five (22%) specified they used it "as required" for their condition (e.g., when pain was severe). Women tended to report shorter term use than men (52% vs. 31% citing use of 5 years or less).
It was most common for participants' medical use to be stable (22%) or largely unchanged since they started (17%), although it was most common for the amount used to vary according to their condition (35%). About one in ten indicated some increase in dose had been required (12%), while few reported a decrease (5%). Women tended to report more variable (44% vs. 29% of men) or short term use (15% vs. 6% of men); men tended to report an increase in the amount needed (17% vs. 4% of women).
In addition to medical use, three quarters (80%) of participants had used cannabis recreationally. Recreational use was less common among older participants (75% and 97% of recreational users were aged less than 50 years and 65 years, respectively). For almost half (46%), use in the past year had been solely medicinal, but the remainder reported recent recreational use – 29% in the past week, 19% in the past month and a further 6% in the past year.
Route of use
While most people had tried multiple routes for relief, overall smoking was the route most commonly reported (91%). Approximately half the sample (49%) also smoked tobacco, and two thirds (64.1%) mixed their cannabis with tobacco.
Eating cannabis in cooked recipes was also very prevalent (49%). While vaporisers are not readily available in Australia, 8% had used them. In addition, four people had used tinctures and one used it topically in the bath or as a cream for a skin condition. Overall, smoking was also considered to be the most helpful route of use for symptom relief (74%), although concerns about this route of use were widespread. Consistent with Australian research on preferred route of use and age [15], older users (aged 50 years +) typically found joints the most helpful method of use (41% vs. 26% of younger users), while younger users preferred the use of waterpipes (43% vs. 13% of older users).
When asked to comment on the good and bad points of different methods of ingestion the most consistent response was that smoking of any form, particularly with tobacco, was detrimental to respiratory function (and health). This was of particular concern to non-smokers, some of whom did not know how to cook cannabis recipes. Despite attracting the bulk of negative comments, its popularity seemed to lie with its instant effect, its ease of titration and cost-effectiveness compared to the oral route. It seemed to "do the job". Eating was seen to be a much healthier option – it was "safer", tasty when cooked in a recipe, less obvious than smoking and could be done virtually anywhere. Some people liked its slow onset and long-lasting effects, but others claimed difficulties with titration and slow onset made it expensive and ineffective for rapid symptom relief.
Effects of cannabis use
When asked to rate the overall effects of cannabis on a Likert scale ranging from "I feel a lot worse" to "gives me great relief", cannabis was perceived to provide "great relief" (86%) or a little relief (14%). No one believed it had been detrimental to their condition or symptoms.
Positive ratings were ("great" or "good" relief) were also typical for its ability to relieve specific symptoms (Table 3). In addition, several other symptoms were noted, primarily insomnia (13% used for insomnia; of these 82% derived "great" relief).
Table 3 Symptom relief (n = 128)
Symptom relief required...* Total (%) Male (%) Female (%)
Nausea relief 48 56 44
Of these, received:
great relief 53 51 62
good relief 44 46 35
no effect 3 3 4
Pain relief 83 83 83
Of these, received:
great relief 55 49 65
good relief 45 52 35
no effect 0 0 0
Ability to cope emotionally 66 70 60
Of these, received:
great relief 45 40 54
good relief 54 58 46
no effect 1 2 0
Appetite stimulant 51 55 44
Of these, received:
great relief 52 55 48
good relief 46 46 48
no effect 2 0 5
Decrease in spasms/tremor 39 36 44
Of these, received:
great relief 43 43 43
good relief 55 54 57
no effect 2 4 0
Relief through relaxation 83 88 75
Of these, received:
great relief 72 69 78
good relief 28 31 22
no effect 0 0 0
* No-one reported their condition was made worse
Approximately three quarters of participants (71%) claimed to have experienced a return of their symptoms or condition on stopping cannabis, especially: pain (53% of those who claimed a return of symptoms), depression or anxiety (30%), insomnia (11%), spasm (10%) and nausea/vomiting or lack of appetite (9%).
Only one in ten (11%) participants reported symptoms they believed were unrelated to their medical condition upon stopping cannabis, citing symptoms congruent with cannabis withdrawal such as anxiety or mood disturbance (including paranoia), insomnia, loss of appetite, restlessness and vivid dreams.
Comparison with other medicines
Almost two thirds (62%) of respondents claimed that they decreased or discontinued their use of other medicines when they started using cannabis medicinally. This was more common in males (65% vs. 58% of females) and older participants (aged 50 years +) (70% vs. 59% among younger participants). For some people this was a substantial change, representing a shift away from chronic, high-dose medication use.
Perhaps not surprisingly, cannabis was typically perceived as superior to other medications in terms of undesirable effects, and the extent of relief provided (Table 4). Thus, cannabis was rated to produce equivalent (8%) or worse side effects (3%) by a minority of therapeutic users. It was considered to work "a bit" or "much better" than other medicines, or to be the only source of relief, by more than three quarters (82%). Two participants made the interesting comment that cannabis worked differently to other medicines, so could not be directly compared.
Table 4 Comparison of cannabis with other medications (n = 128 unless specified).
Total Male Female
Decreased or discontinued use of other medicines (n = 117*) 62% 65 58
Comparison of undesirable effects (n = 125)
Cannabis produced much worse effects than other medicines 1 0 2
Cannabis produced somewhat worse effects 2 4 0
Undesired effects about the same 8 8 9
Other meds produced somewhat worse effects than cannabis 16 14 19
Other medicines produced much worse effects than cannabis 41 40 43
I have no undesirable effects from cannabis 31 33 28
Other medicines work differently 1 1 0
Comparison of relief provided (n = 118*)
Other medicines work much better than cannabis 3 0 7
Other medicines work a bit better than cannabis 3 4 0
Other medicines work about the same as cannabis 9 8 9
Cannabis works a bit better than other medicines 13 11 15
Cannabis works much better than others medication 54 58 48
Only cannabis gives me relief from my condition 15 15 15
Other medicines work differently 2 0 4
Can't distinguish – use them together 1 1 2
Use cannabis to relieve side effects of other medicines 1 1 0
*Some people did not use other medications concurrently
Despite the very positive response to the use of cannabis, nearly one half (41%; 36% of men and 50% of women) found it did not help certain conditions/symptoms. Almost one third (29%) said cannabis was less effective for certain types of pain, or extreme pain, with a further 12% specifying migraine or headache pain. Nearly one in ten (8%) reported no effect on depression or anxiety. More than one in ten (14%) specified that while cannabis could ease their symptoms and enabled them to cope, they realised that it could not cure their underlying condition. Younger participants were more likely than older participants to claim a condition not helped by cannabis (45% vs. 32% of those aged 50 years +).
Supply issues
Participants obtained medical cannabis from multiple sources (median = 1, range = 1–6; 44% had two or more sources), especially friends or family (58%) and dealers (42%). A substantial proportion grew their own (38%) while few (6%) obtained it from a compassion club or cooperative. Among those who purchased cannabis, the median weekly outlay was $50 (range = $1–$500, n = 95).
When asked to comment on the variability of the cannabis they used, those who could obtain a consistent supply of high quality cannabis that suited their needs were in the minority. Typically, participants noticed variability along a number of lines, such as potency, effectiveness, intoxication and side-effects, which made titration difficult. While some noted the importance of factors such as the part of the plant used (e.g., leaf versus head/buds), strain (e.g., sativa versus indica), soil and climate, the overwhelming responses focussed on hydroponic versus soil-grown cannabis ("bush bud" or home grown cannabis), and home grown cannabis versus purchased cannabis.
Hydroponic cannabis was almost universally unpopular and was avoided where possible – despite its greater potency, it was also considered shorter acting, produced greater tolerance and worse side-effects than other cannabis. By comparison, soil-grown cannabis was perceived to be less unpleasantly potent, natural ("organic"), less chemically treated, and with fewer side-effects. However, it was also perceived as harder to get. Home grown cannabis was seen as the best method of obtaining a consistent, safe supply of medicinal quality. A common response was that purchased cannabis was not to be trusted, and that unscrupulous growers who were more concerned with yield and greed compromised the quality of their crop with chemicals such as growth hormone and pesticides.
Concerns
A minority (13%) had no concerns over their medical cannabis use. Concerns over potential health effects (32%) or the risk of dependence (21%) were overshadowed by those relating to its illegal status (76%), the fear of being arrested (60%) and cost (51%). Indeed, one quarter (27%) claimed to have been arrested, cautioned or convicted in relation to their medical cannabis use, with this outcome more commonly reported by men (31% vs. 19% of women) and younger users (30% vs. 16% of users aged 50 years +). Other concerns mentioned (15%) were: the stigma of using, issues around parenting, pregnancy and relationships, availability, quality and difficulties in dose adjustment.
Support from others and interest in clinical trial
Most participants had a regular doctor (90%) and about a half had a regular specialist (55%). Virtually all (90%) had informed a clinician of their therapeutic use, typically reporting a supportive response from GPs (75% of those told), specialists (74%) and nurses (81%). Family and friends were largely considered supportive of the participant's use (71%).
Not surprisingly, there was widespread support for Government provision of cannabis to patients in a variety of circumstances. At least three quarters supported the supply of cannabis to any patient who was permitted to use it by being registered under a Government scheme (82%); more specifically, those patients who: could not afford to buy it on a regular basis (82%), could only purchase it on the black market (81%), couldn't ensure a consistent supply (75%), or were worried about quality control issues (77%). More than half endorsed the supply of patients who did not know anyone capable of growing it (72%), were concerned about hydroponically grown cannabis (72%), or who needed a supply quickly (66%).
Although not all participants were NSW residents, there was almost universal interest (89%) in participating in a clinical trial, in which a controlled supply of cannabis was grown and provided to registered medical cannabis users. There was also strong, although lesser, interest in trying alternative delivery methods such as a spray or tablet (79%).
While for some people, the availability of any cannabis-derived product that worked was their prime concern, alternative delivery methods were considered attractive as they obviated the necessity to smoke, removed concern about engaging in illegal behaviour and having to access the black market, and were more portable and acceptable than smoking. The main caveats on an alternative were that it was easy to titrate, quick, efficient, reliable and natural or safe – sprays and vaporisers were mentioned specifically by some as preferable to pills in this regard. A clear theme was the desire to keep the holistic, natural properties of cannabis rather than produce a chemical/synthetic drug with numerous binding and carrying agents. Nevertheless, there was recognition that different medical conditions may require different approaches, such as different active agents (e.g., THC versus other cannabinoids), strains or methods (e.g., slow release pill versus fast-acting spray).
The main reason for not supporting alternatives appeared to be that using the whole plant in its natural state was perceived to be the best method. In addition, for some the ritual of cannabis use was perceived as part of its medicinal benefit. There was also concern at political interference and its potential for exploitation and corruption in a trial.
Discussion
This exploratory study examined the patterns of medicinal cannabis use among a sample of 128 Australian adults who responded to media stories about this issue. Firstly, we need to acknowledge its limitations. As we do not know how many Australians use cannabis medicinally or their characteristics, we relied on the recruitment of volunteers through purposive sampling. Instead of targeting a particular group we used media stories disseminated widely on the radio, television and in newspapers to attract a cross-section of people. Thus, these results may not be representative of the experiences of all medicinal users, and may be affected by selection bias by excluding those who did not have access to these media, who did not wish to or could not contact us or did not return the questionnaire. We also attracted participants whose experiences with medical cannabis were typically positive, so they have little to tell us about people who have not found cannabis helpful or pleasant therapeutically. However, they still provide important information on these people's experiences, and raise important issues regarding the use of black market supplies of the cannabis plant and the development of cannabis-based pharmaceuticals. As the questionnaire was self-completed, there was potential for misunderstanding of the questions. However, the wording was straightforward, contact details were provided in the event of misunderstanding, and the results were remarkably consistent across participants, which encourages us that the questions were understood. Despite being anonymous, several participants provided us with contact details in case further information was needed, and wrote additional comments about their experiences and attitudes. In addition, many of the findings are remarkably consistent with the findings of other local and international studies, as indicated below.
People in this study reported regular, ongoing medical use over quite long periods – with 61% using for more than five years and 20% reporting very long-term use of more than 20 years. However, as Ware and colleagues noted in their study of almost 1000 medical users [10], this was a group of chronically ill people with multiple long-standing conditions. The perceived need for alternative or additional symptom relief may reflect the fact that we recruited a sample of particularly entrenched medicinal cannabis users who were dissatisfied with conventional treatments, that medicinal cannabis use is more likely to considered an option by people who find conventional treatments and medications unsatisfactory, or that many had been exposed to its perceived medical benefits quite early due to their recreational use. Larger studies addressing a broad cross-section of users may better answer this question.
Consistent with the literature on the conditions for which cannabis has been indicated, chronic pain, arthritis, persistent nausea and weight loss were among the most common conditions for which cannabis relief was sought. However, depression was the most common condition: more than half (56%) used cannabis to relieve depression, and two thirds (66%) used it to cope emotionally, universally obtaining great or good relief. Other studies have also reported cannabis use for the relief of depression, although not at this level [8-10,14]. The relationship between depression and cannabis use is controversial, with recent literature indicating that cannabis use may be implicated in depression and suicidal thoughts and behaviours This would suggest that regular medicinal use may be contraindicated by placing people at risk of depression or self-harm. However, we do not know the type or aetiology of the depression cited by our participants. Many may have experienced depression and stress associated with their physical condition, which may have been alleviated along with any physical relief. The risk may also be greatest among heavy, younger users and those who may already be vulnerable to mental ill health due to their life circumstances [16-18]. Medical cannabis use patterns may not typically be regular enough to pose a great risk. Regardless, it is important that people considering the use of medical cannabis are aware of the risks of use [19]. A recent paper [20] has suggested that THC and cannabidiol, two major components of cannabis, may help alleviate bipolar disorder, recommending a pharmaceutical product would be a safer option than crude cannabis, in which the balance of components is variable.
Consistent with local and international research on people with a variety of medical conditions [8-12,14], most participants claimed moderate to substantial benefits from cannabis, both in terms of their overall condition and management of individual symptoms. It was typically considered more effective and less aversive than other medications in managing their condition(s), the symptoms of which commonly re-emerged upon stopping (71%). While their use was often complementary to other medications and treatment, 62% had decreased or discontinued use of other medications when they commenced medicinal cannabis use. Nevertheless, cannabis was not a panacea – it did not help all conditions, particularly certain types of pain, and there was recognition that while it substantially improved quality of life it was not a cure. This is not necessarily surprising, as overall well-being and specific symptoms have multiple causes and can be affected by several factors, and is borne out by recent controlled clinical trials, for example, on chronic pain [21].
As others have reported (e.g., [8-10] we also found that in addition to medical use, recreational use was common: most (80%) had used cannabis recreationally, with about one half (54%) of these reporting some recent use. Indeed, 29% had discovered its therapeutic potential through their recreational use. One participant raised the issue that part of the therapeutic effect for them was the ritual of use and the "high" experienced [6]. This demonstrates the difficulty of precisely identifying the therapeutic component when people are using the natural plant matter, and will continue to present a challenge for the development of cannabis pharmaceuticals. While some people may find the illegality, route of use and psychoactive effects of natural cannabis undesirable and prefer a manufactured pharmaceutical product, several in this survey claimed to prefer the holistic delivery of all the compounds present when using the natural plant. We need to know more about the effect of the different active chemicals on medical conditions and how their therapeutic potential is mediated by the context of use.
Nonetheless, this was not simply a sample of recreational users, especially as we attracted many older users who used exclusively for medical reasons (75% of those aged 50 years+). They did not fit the recreational user stereotype, were willing to take the risk of using an illicit drug, exposure to the illicit drug market and the possibility of arrest to gain symptom relief. Indeed, the most common concern over medicinal use was its illegality, fear of arrest and cost (all >50%). One quarter (27%) of participants had experienced legal ramifications due to their use. Several people commented that they had no alternative than using an illegal drug, claiming that other medicines with negative and toxic effects (e.g., opiates) were legally prescribed, and that if nothing else worked for them they had the right to access cannabis without fear or stigma. Several made pleas for medical cannabis use to be treated as a medical, rather than a legal, issue, as their health and quality of life were at stake.
Smoking was the most common method of use; in addition, many were tobacco smokers or mixed cannabis with tobacco. Given the similarities between cannabis and tobacco smoke this is of particular concern for people who are ill, especially those with compromised immune systems. Despite acknowledgement of the risks of smoking and concerns expressed over its effects, it was considered the most helpful route of use. While eating was perceived as much healthier, until satisfactory solutions are achieved on titration and dosing issues, smoking will no doubt continue to be a popular method of obtaining relief.
Cannabis dependence was a concern for one in five participants (21%). This study provided indirect evidence that participants were unlikely to experience withdrawal symptoms on ceasing medical use, but this was only a crude measure. While the risk of dependence is probably low when used medicinally, this risk needs to be weighed up with the other concerns of the patient – for example, it may be low on the list of concerns for those with terminal illness [19].
Finally, participants reported that family and friends were likely to know about and support their medical cannabis use. These data also indicate that the medical profession is encountering, and frequently supporting, patients who use cannabis for symptom relief. Given their central role in the management of illness, it is important that clinicians are educated about the effects of cannabis, in order to assist patients in making informed decisions about their treatment. There was also clearly great interest among participants in a clinical trial and scope to investigate methods of delivery that avoid the health concerns associated with smoking cannabis, keeping in mind that some participants were reluctant to use a pharmaceutical product. In addition to distrust of unscrupulous participants in the black market, some were also distrustful of Government's motives and role in therapeutic research. It is therefore vital that any clinical trials are conducted in a rigorous, independent manner.
Conclusion
Overall, these findings are consistent with those of other surveys, in revealing the perceived effectiveness of cannabis for the relief of symptoms associated with several medical conditions. While a small study, it has several implications. Firstly, people are risking the use of an illicit drug for its perceived therapeutic effects, and in some cases being arrested. Secondly, they are informing their clinicians about their medical use and frequently receiving support, highlighting the importance of ensuring clinicians are informed about cannabis. Finally, in addition to strong public support, medical cannabis users show strong interest in clinical cannabis research, including the investigation of alternative delivery methods.
Competing interests
The author(s) declare they have no competing interests.
Authors' contributions
WS conceived the study, designed the methodology, adapted the questionnaire, cleaned and analysed the data and wrote the paper.
PG assisted in questionnaire adaptation, managed data collection, entered the data, assisted with preliminary data analyses and commented on the manuscript.
PD assisted in questionnaire adaptation, recruited participants and commented on the manuscript.
All authors read and approved the final manuscript.
Acknowledgements
Thanks to all the participants for sharing their experiences and to: Andrew Kavisilas for permission to adapt his questionnaire and ongoing support; and Graham Irvine, Franjo Grotenhermen, Laurie Mather, Wayne Hall and Louisa Degenhardt for comments on the questionnaire.
==== Refs
Grotenhermen F Russo E Cannabis and cannabinoids: Pharmacology, toxicology and therapeutic potential 2002 NY, Haworth Integrative Healing Press
Joy JE Watson SJ Benson JAJ Marijuana and medicine:Assessing the science base 1999 Washington, DC, National Academy Press
House of Lords Select Committee on Science and Technology Cannabis: The scientific and medical evidence 1998 London, The Stationery Office
New South Wales Parliament. Working Party on the Use of cannabis for Medical Purposes Report of the Working Party on the use of cannabis for medical purposes 2000 Sydney, New South Wales Parliament. Working Party on the Use of cannabis for Medical Purposes 42
Australian Institute of Health and Welfare 2004 National Drug Strategy Household Survey: First results 2005 Canberra, Australian Institute of Health and Welfare
Coomber R Oliver M Morris C Using cannabis therapeutically in the UK: A qualitative analysis Journal of Drug Issues 2003 33 325
Gieringer D Grotenhermen F and Russo E Medical use of cannabis: Experience in California Cannabis and cannabinoids: Pharmacology, toxicology and therapeutic potential 2002 NY, Haworth Integrative Press 143 151
Grotenhermen F Schnelle M Survey on the medical use of cannabis and THC in Germany Journal of Cannabis Therapeutics 2003 3 17 40 10.1300/J175v03n02_03
Ogborne AC Smart RG Weber T Birchmore-Timney C Who is using cannabis as a medicine and why: An exploratory study Journal of Psychoactive Drugs 2000 32 435 443 11210205
Ware MA Adams H Guy GW The medicinal use of cannabis in the UK: results of a nationwide survey. International Journal of Clinical Practice 2005 59 291 295 15857325 10.1111/j.1742-1241.2004.00271.x
Ware MA Doyle CR Woods R Lynch ME Clark AJ Cannabis use for chronic non-cancer pain: results of a prospective study. Pain 2003 102 211 216 12620613 10.1016/s0304-3959(02)00400-1
Ware MA Rueda S Singer J Kilby D Cannabis use by persons living with HIV/AIDS: patterns and prevalence of use Journal of Cannabis Therapeutics 2003 3 3 15 10.1300/J175v03n02_02
Helliwell D Medicinal Cannabis survey.
Kavasilas A Medical uses of cannabis: Information for medical practitioners 2004 Sydney, Inn Press
Hall W Swift W The THC content of cannabis in Australia: Evidence and implications Australian and New Zealand Journal of Public Health 2000 24 503 508 11109687
Rey JM Sawyer MG Raphael B Patton GC Lynskey M Mental health of teenagers who use cannabis: results of an Australian survey. British Journal of Psychiatry 2002 180 216 221 11872513 10.1192/bjp.180.3.216
Fergusson DM Horwood LJ Swain-Campbell NR Cannabis dependence and psychotic symptoms in young people. Psychological Medicine 2003 33 15 21 12537032 10.1017/S0033291702006402
Patton GC Coffey C Carlin JB Degenhardt L Lynskey M Hall W Cannabis use and mental health in young people: cohort study. British Medical Journal 2002 325 1195 1198 12446533
Swift W Hall W Grotenhermen F and Russo E Cannabis and dependence Cannabis and cannabinoids: Pharmacology, toxicology, and therapeutic potential 2002 NY, Haworth Integrative Healing Press
Ashton CH Moore PB Gallagher P Young AH Cannabinoids in bipolar affective disorder: a review and discussion of their therapeutic potential Journal of Psychopharmacology 2005 19 293 300 15888515 10.1177/0269881105051541
Notcutt W Price M Miller R Newport S Phillips C Simmons S Sansom S Initial experiences with medicinal extracts of cannabis for chronic pain: Results from 34 "N of 1" studies. Anaesthesia 2004 59 440 452 15096238 10.1111/j.1365-2044.2004.03674.x
|
16202145
|
PMC1262744
|
CC BY
|
2021-01-04 16:36:49
|
no
|
Harm Reduct J. 2005 Oct 4; 2:18
|
utf-8
|
Harm Reduct J
| 2,005 |
10.1186/1477-7517-2-18
|
oa_comm
|
==== Front
Health Qual Life OutcomesHealth and Quality of Life Outcomes1477-7525BioMed Central London 1477-7525-3-571615330810.1186/1477-7525-3-57ResearchApplication of a disease-specific mapping function to estimate utility gains with effective treatment of schizophrenia Lenert Leslie A [email protected] Marcia FT [email protected] Christine [email protected] Veterans Administration San Diego Health Care System, San Diego, California, USA2 University of California, San Diego, California, USA3 Janssen Medical Affairs, L.L.C., Titusville, NJ, USA4 Health Services Research and Development Service, Department of Veteran Affairs, Washington, DC, USA2005 11 9 2005 3 57 57 15 6 2005 11 9 2005 Copyright © 2005 Lenert et al; licensee BioMed Central Ltd.2005Lenert et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Most tools for estimating utilities use clinical trial data from general health status models, such as the 36-Item Short-Form Health Survey (SF-36). A disease-specific model may be more appropriate. The objective of this study was to apply a disease-specific utility mapping function for schizophrenia to data from a large, 1-year, open-label study of long-acting risperidone and to compare its performance with an SF-36-based utility mapping function.
Methods
Patients with schizophrenia or schizoaffective disorder by DSM-IV criteria received 25, 50, or 75 mg long-acting risperidone every 2 weeks for 12 months. The Positive and Negative Syndrome Scale (PANSS) and SF-36 were used to assess efficacy and health-related quality of life. Movement disorder severity was measured using the Extrapyramidal Symptom Rating Scale (ESRS); data concerning other common adverse effects (orthostatic hypotension, weight gain) were collected. Transforms were applied to estimate utilities.
Results
A total of 474 patients completed the study. Long-acting risperidone treatment was associated with a utility gain of 0.051 using the disease-specific function. The estimated gain using an SF-36-based mapping function was smaller: 0.0285. Estimates of gains were only weakly correlated (r = 0.2). Because of differences in scaling and variance, the requisite sample size for a randomized trial to confirm observed effects is much smaller for the disease-specific mapping function (156 versus 672 total subjects).
Conclusion
Application of a disease-specific mapping function was feasible. Differences in scaling and precision suggest the clinically based mapping function has greater power than the SF-36-based measure to detect differences in utility.
==== Body
Background
Estimation of cost-effectiveness in clinical trial settings requires measurement of changes in utility. This is particularly difficult in diseases that impact cognitive functioning, such as schizophrenia or Alzheimer's disease, because these impairments may preclude direct elicitation of utilities in trial participants. Even in cognitively intact persons, direct elicitation often is logistically difficult in clinical trial settings and therefore rarely done. To overcome these issues, researchers have developed health index models to assign a utility to each individual. Health index models, including the Health Utilities Index, EuroQol (EQ-5D), and the Quality of Well-Being Scale (QWB), present comprehensive models of quality of life [1-4]. Measurement of attributes within these models allows assignment of utility scores based on population models of values summarized by the index; however, if these measures were not performed during the study, the direct application of a health index model is not possible.
Many trials include measurements of health status performed with short-form measures such as the 36-Item Short-Form Health Survey (SF-36) [5,6]. An alternative to health index models is to use these data to estimate utility values. Efforts in this vein began with work by Fryback and colleagues on mapping between the SF-36 and the QWB [7] and have been extended by many others [8,9]. A second approach, described by Brazier and colleagues, has been to develop a health index based on the content of a short-form measure and to measure utilities for the models with larger numbers of states as defined by the short-form measure [10,11]. A refinement of this method, which involves using k-means clustering to find a small number of states that represent patterns of disease effects on health status, has been described by Lenert and colleagues [12]. The primary advantage of this approach is that it allows direct comparison of implications of differences between patient and general population values, as is recommended by the cost-effectiveness analysis guideline panel [13]. In comparison studies, this approach appeared to be more responsive than other mapping functions for short-form measures in depression [14].
Use of a health index model or a short-form measure of health status does not address the issue of disease-specific effects on quality of life, however. To capture disease-specific effects, investigators typically measure disease activity with validated scales specific to a disorder. Other times, modeling of the symptoms of the disease itself may be required. In schizophrenia, a commonly used disease activity measurement scale is the Positive and Negative Syndrome Scale (PANSS) [15]. This scale measures, via interview, the total burden of symptoms and impact of disease. Change in the PANSS may or may not be reflected in short-form measures such as the SF-36, although they often are responsive in schizophrenia and other mental illnesses. In this paper, we describe the application of a disease-specific utility mapping function based on the PANSS, incorporating an additional assessment of the impact of adverse effects of medication into the model. The methods used to create the model have been described elsewhere [16,17]. Briefly, a set of disease states for schizophrenia were developed from the PANSS based on data from a large, 1-year study that prospectively compared oral risperidone with conventional antipsychotic agents among patients with schizophrenia who were treated under usual practice conditions [18]. Data were analyzed by cluster analysis for five factor domains; cluster analysis results were compared with an expert-developed conceptual framework of disease states. Using a combination of the empirical data and the conceptual framework, eight disease states with varying levels of positive, negative, and cognitive impairment were established. Health states were described in a holistic fashion that included interactions between effects of symptoms of the disease and other aspects of quality of life. Utilities were measured in the general public from a convenience sample of a large Internet survey panel [19]. Participants viewed digital videos of actors depicting the eight health states and five common antipsychotic side effects (akathisia, pseudo-parkinsonism, orthostatic hypotension, dyskinesia, and clinically important weight gain) and rated the states using a visual analog scale (VAS) followed by a standard gamble (SG). The mean utility rating for each state and the reduction in utility with each adverse effect were estimated by re-weighting responses so that calculated mean values matched US population demographic profiles in age, ethnicity, and gender [17].
We report application of this mapping function to a 50-week, multicenter, international, open-label study of long-acting risperidone in patients with schizophrenia and schizoaffective disorder. Detailed safety and efficacy results of this study have been published and presented elsewhere [20,21]. Changes in utility associated with long-term use of long-acting risperidone were estimated using the mapping function in patients who completed 50 weeks of therapy. We then compared results to a mapping function for the SF-36.
Methods
To assess the responsiveness of the utility mapping function, we applied it to data from an open-label, international (Europe and Canada), 50-week trial evaluating the long-term safety and tolerability of long-acting risperidone in 725 patients with schizophrenia or schizoaffective disorder considered to be stable at study entry [20,21]. The final protocol for, and any amendments to, the original study were reviewed and approved by independent Ethics Committees or by appropriately constituted institutional review boards (IRBs) according to specifications outlined in the US Code of Federal Regulations (CFR). This trial was conducted in accordance with current International Conference on Harmonisation (ICH)-Good Clinical Practice guidelines and the Declaration of Helsinki and its subsequent revisions. Patients were aged 18 to 85 years with a diagnosis of schizophrenia or schizoaffective disorder according to DSM-IV criteria [22] and were judged to be clinically stable (stable symptoms and antipsychotic dose for at least 1 month).
Trial medication
During a 2-week run-in period, antipsychotics other than risperidone were discontinued, and patients not currently being treated with risperidone received flexible doses of 1 to 6 mg/daily of oral risperidone. Assessments performed prior to this run-in period, however, were considered as the baseline for this analysis. By protocol, pharmacokinetic considerations, and investigator judgment, patients were assigned to flexible-dose treatment with 25, 50, or 75 mg long-acting risperidone given by intramuscular gluteal injection every 2 weeks. The investigator could adjust the dose of long-acting risperidone when necessary. Medications other than long-acting risperidone that could be initiated or continued during the trial included anticholinergic agents, antidepressants, mood stabilizers, propranolol for akathisia, and benzodiazepines for agitation and insomnia.
Assessments
PANSS total scores [15] and health-related quality-of-life assessments, measured by the SF-36 [5,6], were collected at weeks 1, 12, 24, 36, 50, and at endpoint. Adverse effects of treatment were assessed at the same time points, using the Extrapyramidal Symptom Rating Scale (ESRS) [23] to determine the Clinical Global Impression (CGI) of severity of parkinsonism, dystonia, and dyskinesia, and adverse-effect reporting to document weight gain and hypotension. Values for systolic blood pressure, pulse, and weight were obtained at baseline and endpoint.
Analysis plan
Disease states were assigned at each observation based on the mean value of each patient's PANSS items using the model developed by Mohr and colleagues [16]. For patients completing at least 50 weeks of treatment, the mean values of VAS and SG utility were calculated from the PANSS, and VAS and SG utility were calculated from the SF-36. VAS and SG utilities calculated from the PANSS were adjusted for adverse effects using a multiplicative model. An individual with a score of ≥4 on any of the ESRS subscales for parkinsonism, dystonia, or dyskinesia was ascribed as having that adverse effect. An individual was ascribed as having orthostatic hypotension for the entire duration of the study if upon exit from the trial, the patient exhibited a ≥20-mm Hg drop in standing blood pressure. Patients with a gain of ≥10 kg (≥22.4 lb) during the study were assessed a utility tariff for weight gain. These measurements were performed only at 24 weeks, 50 weeks, and endpoint. Missing data were estimated using a last-value-carried-forward approach in patients completing the study.
Estimated utility values at each point in time were compared using the Wilcoxon signed rank test. Overall gains in utility over the course of the study were calculated by subtracting baseline from endpoint values and compared with those estimated from an SF-36-based mapping function developed by Nichol and colleagues [8]. These calculations were used to compare both the magnitude of the estimated utility gains and the correlation of the gains between the two mapping functions. To compare the responsiveness of the measures, we took the standard deviation of SF-36-based and clinically based utility measures at baseline and endpoint and the effect size seen in this study for each measure and estimated the sample size required for a clinical trial to confirm the effect seen in this observational study.
Results
A total of 725 symptomatically stable patients with schizophrenia (n = 615) or schizoaffective disorder (n = 110), received long-acting risperidone treatment. Four hundred seventy-four patients (65.3%) completed the trial. Demographic characteristics of patients who completed or discontinued the study are displayed in Table 1. The only significant difference in baseline characteristics between those who completed the study and those who discontinued was the mean age.
Table 1 Demographics and Baseline Disease Characteristics for Patients Who Completed or Discontinued the Study
Parameter Patients Who Completed the Study (n = 474) Patients Who Discontinued (n = 251) P value Between Groups*
Age, y (mean ± SE) 43.7 ± 0.7 39.3 ± 0.9 <0.0001
Sex 0.730
Female, n (%) 162 (34.2) 89 (35.5)
Male, n (%) 312 (65.8) 162 (64.5)
Race 0.502
Caucasian, n (%) 440 (92.8) 231 (92.0)
Black, n (%) 8 (1.7) 8 (3.2)
Asian, n (%) 6 (1.3) 5 (2.0)
Hispanic, n (%) 4 (0.8) 2 (0.8)
Other, n (%) 16 (3.4) 5 (2.0)
Diagnosis 0.650
Schizophrenia, n (%) 400 (84.4) 215 (85.7)
Schizoaffective disorder, n (%) 74 (15.6) 36 (14.3)
SE indicates standard error. *Chi-square test.
A graphic depiction of the eight health states used in the PANSS-based mapping function is provided in Figure 1. Each health state represents a set of symptoms ranging from mild to very severe, with patients having mild disease (health state 1) displaying low symptoms, and patients in the very severe disease state (health state 8) displaying high symptoms, with the exception of cognitive impairment, which could be either high or low. Two separate groups are considered to have moderate symptoms (health states 2, 3), while 4 health states (states 4–7) describe patients with severe symptoms [16]. The distribution of these health states at the beginning and endpoint of the trial are shown in Figure 2. Patients who completed treatment with long-acting risperidone experienced substantial symptomatic improvement over the 1-year study. Importantly, the percentage of patients in health state 1 (representing full remission of symptoms) increased significantly, from 25% to 42% over the course of the study (P < 0.001, McNemar's test). The impact of this shift was significant. Considering symptoms of schizophrenia alone, mean SG-weighted utilities increased significantly, from 0.729 at baseline to 0.775 at endpoint (P < 0.001, Wilcoxon signed rank test), with a net gain of 0.046. VAS-weighted ratings yielded similar results, with a gain in utility equaling 0.058 from baseline (0.538) to endpoint (0.596, P < 0.001, Wilcoxon signed rank test).
Figure 1 Symptom description for the eight health states used in the PANSS-based mapping function. PANSS indicates Positive and Negative Syndrome Scale; Neg, negative symptoms; Pos, positive symptoms; Cog, cognitive impairment; MOD, moderate symptom impairment. Adapted from Mohr PE, Cheng CM, Claxton K, et al. [16]. Reproduced with permission.
Figure 2 Distribution of health states at baseline and at endpoint of the 1-year study. Numbers of patients evaluated were 471 at baseline and 474 at endpoint. This figure illustrates both the floor effects of the measurement model as well as its descriptive validity: the percentage of patients in health state 2 shifted to a higher level of health in state 1 at the study endpoint. *P < 0.001 vs baseline, McNemar's test.
The incidence of common antipsychotic-associated adverse effects over the course of the study (parkinsonism, akathisia, dyskinesia, orthostatic hypotension, and weight gain) was assessed. Movement disorder side effects decreased over time, reflected in lower frequencies of moderate or severe parkinsonism (from 25.6% to 15.4%), akathisia (from 9.4% to 4.2%), and dyskinesia (from 13.6% to 9.5%) at endpoint. The occurrence of orthostatic hypotension overall was low; only 4 cases were reported during the study. Weight gain was the only adverse effect, with increased frequency over time. During the study, 51 patients gained ≥20 pounds, thus meeting the criteria for the utility tariff.
Adverse effects, as may be expected, impacted utility gains. Because adverse effects overall decreased with treatment over the course of the study, further gains in utility were realized (Table 2). Gains were 0.051 for SG-weighted comparisons and 0.064 for VAS-weighted comparisons after adjusting for adverse effects. Changes in utility from baseline to endpoint were statistically significant for both comparisons (P < 0.001, Wilcoxon signed rank test).
Table 2 Utility Gains Adjusted for Adverse Effects
VAS SEM SG SEM
Baseline 0.519 0.00725 0.712 0.00578
12 weeks 0.570* 0.00676 0.751† 0.00560
24 weeks 0.583* 0.00668 0.762† 0.00542
36 weeks 0.594* 0.00658 0.768† 0.00551
50 weeks 0.591* 0.00668 0.766† 0.00554
Endpoint 0.583* 0.00647 0.763† 0.00552
VAS indicates visual analog scale; SEM, standard error of the mean; SG, standard gamble.
*P < 0.001 versus baseline visual-analog-scale measurement (Wilcoxon signed rank test).
†P < 0.001 versus baseline standard-gamble measurement (Wilcoxon signed rank test).
Utility changes were estimated by a second method, which used the approach devised by Nichol and colleagues of mapping SF-36 domain scores to Health Utility Index (HUI) Mark II scores [8]. By this method, we found the average baseline utility to be 0.762. As with the disease-specific PANSS by adverse-effect method, utility attributable to long-acting risperidone treatment increased at endpoint but by a smaller degree, 0.0285 units (95% confidence interval: 0.039–0.017). The SF-36 mapping function was significantly but not strongly correlated with the PANSS by adverse-effect mapping function (r = 0.20 Pearson correlation coefficient; Figure 3).
Figure 3 Correlation between gains in utility, estimated using PANSS mapping function adjusted for averse effects (PMF+) and the SF-36 mapping function.
To compare the responsiveness of both measures, we estimated the sample size that would be required for a randomized clinical trial to have 80% power to detect the changes in utility found in this observational study, at an alpha of 0.05. The standard deviation of the clinical mapping function was 0.127 at baseline and 0.125 at endpoint. The change in observed utility was 0.051, or about 0.4 standard deviations (a moderate effect, according to Cohen [24]). This translates to a requirement for about 156 total subjects to achieve the specified power. The change in utility seen with SF-36 function was smaller (0.0285), and the standard deviation was slightly larger (0.136 at baseline and 0.132 at endpoint). This translates to an effect size of 0.211, or about half the effect size of that seen with the clinical mapping function. A clinical trial designed to detect the observed change with the SF-36 mapping function would need 672 total subjects, or about four times the number that would be required if the study used the clinical mapping function. By way of comparison, change score for relevant PANSS items was 6.9 with a standard deviation 9.92. This translates to an effect size of .70. Only 19 subjects would be required to detect a positive change in the overall score.
Discussion
Generation of utility weights for cost-effectiveness analysis is often a difficult task. This analysis applied a mapping function for the PANSS, with preference weights from a diverse sample of the US population, to a clinical observational study. Results demonstrate both the feasibility and the responsiveness of the function as a tool in cost effectiveness analysis. Estimates of gains in utility based on the disease-specific mapping function ranged from 0.046 to 0.064, depending on the scaling method and whether adverse effects of medication were included in the model. The effect was greater than that calculated using an SF-36-based mapping function, and the disease-based measure had greater precision and power to detect differences observed with treatment; however, its power was still not close to the change score for the PANSS items used in the mapping function.
These data confirm that utility calculations from disease specific and generic instruments may not be directly comparable. The relatively low correlation (r = 0.2) is probably due to the instruments covering different content areas. It could be argued that the optimal mapping system might incorporate both health status effects and disease effects in a utility model. To address the issue of avoidance of double counting of gains, one would need to apply methods to address the correlation that does exist between symptoms and their effects on health related quality of life. This might be done at the model formulation level through use of principal components analysis and cluster analysis to define states using both PANSS and SF-36 data. Methods described by Sugar and co-authors [25] might be suitable for this task.
Nonetheless, the estimates provided by the clinical function are better suited to use in a cost-effectiveness analysis than the ones derived from the SF-36 mapping function in this domain. The PANSS records an interviewer's perceptions of disease effects on the patient. The SF-36 is a self administered instrument. If an individual lacks the insight to appreciate health related quality of life impacts (lack of insight into disease effects is common in schizophrenia), then mapping functions based on self-report data might lack construct validity.
In mental illness disease effects and health related quality of life are highly convolved and it would be difficult to separate health related quality of life from disease experience. The clinical function was based on health state descriptions that included impacts of the disease on health related quality of life [17]. These descriptions were designed to be sufficiently comprehensive of health related quality of life to warrant direct usage in a cost-utility analysis. If descriptions had been limited to disease effects, further adjustments to utility estimates might be necessary prior to use in a cost-effectiveness analysis [26].
A few studies provide comparisons of utility gains with treatment in schizophrenia. Rosenheck and colleagues constructed a mapping function for schizophrenia with a quality-adjusted-life-year (QALY) like weight, based on subjectively defined "best" and "worst" possible health states [27]. They estimated that treatment of refractory patients with clozapine increased the quality-of-life measure by 0.049 units during a 1-year study. Pyne and coworkers estimated the utility gain with clinical improvement using the QWB scale and Brazier's mapping function for the SF-36 [28]. They found that "clinically significant" improvement in schizophrenia was associated with a 0.048 gain in utility using the QWB scale, and a 0.043 gain using the VAS-scaled version of Brazier's SF-36 mapping function.
This study had several limitations. First, the data were from an open-label study, which began with a 2-week oral risperidone run-in period. Estimates of gain in utility depended on the degree of symptom control that was achieved during this oral-dosing period; if symptoms were poorly controlled during this period, benefits of long-acting risperidone treatment for this population of clinically stable patients could have been overestimated. Second, this analysis included only patients who completed the trial. However, baseline demographics and disease characteristics of patients who completed the trial versus those who discontinued were not significantly different, with the exception of mean patient age (Table 1). While these design limitations limit the generalizability of findings of utility gains for treatment with long-acting risperidone, they do not impact assessments of the mapping function. Another important limitation of the clinical mapping function is the limited set of adverse effects of antipsychotic treatment accounted for in this model. While not all adverse effects were included in the mapping function, the features included have been described as the key benefits or liabilities of atypical agents versus conventional antipsychotics [29,30]. Thus, the most important and relevant medication side effects for contemporary pharmacoeconomic analyses have been included; however, the model may need to be expanded as new drugs are developed.
The mapping function applied in this study has technical advantages and disadvantages. The health states are based on a combination of clinical data and expert judgment. We believe that this is an optimal mix because it is a data-driven approach that compensates for under-representation of certain types of patients in clinical trials [16]. A second advantage is the software program used to elicit utilities. The software used multimedia video clips to describe the health effects of schizophrenia and adverse effects. This most likely improves the face validity of measurements because the health effects of schizophrenia can be difficult to comprehend to those without direct experience. A second advantage of the software program is its use of advanced methods for error correction in utility elicitations that have been proven to yield more accurate population estimates of utility values [31]. However, the computerized approach also brings limitations. Computer surveys are difficult to administer to "representative" samples. To limit data collection costs for the model, data were measured in members of an Internet survey panel [19]. Although participants were a diverse group in terms of geography, age, and ethnicity, the sample may not be representative of the US population because they were all Internet users (and members of a research panel) and because of drop-out due to technical issues with survey software.
Conclusion
In summary, this paper describes the application of a new disease-specific utility mapping function, based on the PANSS and adverse events, to estimate gains in utility in a clinical study. This function is easy to apply and appears to have greater precision than a SF-36-based mapping function. One of the greatest advantages of the disease-specific mapping function is that it uses data generally available in clinical trials for schizophrenia (PANSS), and thus it could have wide applicability.
Authors' contributions
LAL designed the study, developed the analysis plan, contributed to statistical analyses, and drafted the manuscript. MR participated in the design of the study, contributed to the analysis plan, and helped draft the manuscript. CE performed statistical analyses and helped draft the manuscript. All authors have read and approved the final manuscript.
Acknowledgements
This study was supported by Janssen Medical Affairs, L.L.C. The authors are thankful for the clinical advice provided by Robert Lasser, MD, and critical review of the study results by Julie Locklear, PharmD.
==== Refs
Furlong WJ Feeny DH Torrance GW Barr RD The Health Utilities Index (HUI) system for assessing health-related quality of life in clinical studies Ann Med 2001 33 375 384 11491197
Torrance GW Feeny DH Furlong WJ Barr RD Zhang Y Wang Q Multiattribute utility function for a comprehensive health status classification system. Health Utilities Index Mark 2 Med Care 1996 34 702 722 8676608 10.1097/00005650-199607000-00004
Kind P Dolan P Gudex C Williams A Variations in population health status: results from a United Kingdom national questionnaire survey BMJ 1998 316 736 741 9529408
Kaplan RM Ganiats TG Sieber WJ Anderson JP The Quality of Well-Being Scale: critical similarities and differences with SF-36 Int J Qual Health Care 1998 10 509 520 9928590 10.1093/intqhc/10.6.509
Ware J JrKosinski M Keller SD A 12-Item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity Med Care 1996 34 220 233 8628042 10.1097/00005650-199603000-00003
Ware JE JrSherbourne CD The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection Med Care 1992 30 473 483 1593914
Fryback DG Lawrence WF Martin PA Klein R Klein BE Predicting Quality of Well-being scores from the SF-36: results from the Beaver Dam Health Outcomes Study Med Decis Making 1997 17 1 9 8994146
Nichol MB Sengupta N Globe DR Evaluating quality-adjusted life years: estimation of the health utility index (HUI2) from the SF-36 Med Decis Making 2001 21 105 112 11310943 10.1177/02729890122062352
Sengupta N Nichol MB Wu J Globe D Mapping the SF-12 to the HUI3 and VAS in a managed care population Med Care 2004 42 927 937 15319619 10.1097/01.mlr.0000135812.52570.42
Brazier JE Roberts J The estimation of a preference-based measure of health from the SF-12 Med Care 2004 42 851 859 15319610 10.1097/01.mlr.0000135827.18610.0d
Brazier J Roberts J Deverill M The estimation of a preference-based measure of health from the SF-36 J Health Econ 2002 21 271 292 11939242 10.1016/S0167-6296(01)00130-8
Lenert LA Sherbourne CD Sugar C Wells KB Estimation of utilities for the effects of depression from the SF-12 Med Care 2000 38 763 770 10901359 10.1097/00005650-200007000-00008
Siegel JE Torrance GW Russell LB Luce BR Weinstein MC Gold MR Guidelines for pharmacoeconomic studies. Recommendations from the panel on cost effectiveness in health and medicine. Panel on cost Effectiveness in Health and Medicine Pharmacoeconomics 1997 11 159 168 10172935
Donald Sherbourne C Unutzer J Schoenbaum M Duan N Lenert LA Sturm R Wells KB Can utility-weighted health-related quality-of-life estimates capture health effects of quality improvement for depression? Med Care 2001 39 1246 1259 11606878 10.1097/00005650-200111000-00011
Kay SR Opler LA Lindenmayer JP The Positive and Negative Syndrome Scale (PANSS): rationale and standardisation Br J Psychiatry 1989 59 67
Mohr PE Cheng CM Claxton K Conley RR Feldman JJ Hargreaves WA Lehman AF Lenert LA Mahmoud R Marder SR Neumann PJ The heterogeneity of schizophrenia in disease states Schizophr Res 2004 71 83 95 15374576 10.1016/j.schres.2003.11.008
Lenert LA Sturley AP Rapaport MH Chavez S Mohr PE Rupnow M Public preferences for health states with schizophrenia and a mapping function to estimate utilities from positive and negative symptom scale scores Schizophr Res 2004 71 155 165 15374583 10.1016/j.schres.2003.10.010
Mahmoud RA Engelhart LM Janagap CC Oster G Ollendorf D Risperidone versus conventional antipsychotics for schizophrenia and schizoaffective disorder: symptoms, quality of life and resource use under customary clinical care Clin Drug Invest 2004 24 275 286
Lenert LA Sturley AE Kohane IS Use of the Internet to study the utility values of the public Proceedings of the 2002 AMIA Annual Symposium: 9–13 November 2002; San Antonio 2002 Bethesda: AMIA 440 444
Fleischhacker WW Eerdekens M Karcher K Remington G Llorca PM Chrzanowski W Martin S Gefvert O Treatment of schizophrenia with long-acting injectable risperidone: a 12-month open-label trial of the first long-acting second-generation antipsychotic J Clin Psychiatry 2003 64 1250 1257 14658976
Lasser R Bossie CA Gharabawi G Eerdekens M Nasrallah HA Efficacy and safety of long-acting risperidone in stable patients with schizoaffective disorder J Affect Disord 2004 83 263 275 15555724 10.1016/j.jad.2004.05.008
American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders 1994 4 Washington, DC: American Psychiatric Association
Chouinard G Ross-Chouinard A Annable L Jones BD Extrapyramidal symptom rating scale Can J Neurolog Sci 1980 7 233
Cohen J Statistical Power Analysis for the Behavioral Sciences 1988 2 Hillsdale, New Jersey: Lawrence Erlbaum Associates
Sugar CA James GM Lenert LA Rosenheck RA Discrete state analysis for interpretation of data from clinical trials Med Care 2004 42 183 196 14734956 10.1097/01.mlr.0000108748.13206.ba
Stolk EA Busschbach JJ Validity and feasibility of the use of condition-specific outcome measures in economic evaluation Qual Life Res 2003 12 363 371 12797709 10.1023/A:1023453405252
Rosenheck R Cramer J Xu W Grabowski J Douyon R Thomas J Henderson W Charney D Multiple outcome assessment in a study of the cost-effectiveness of clozapine in the treatment of refractory schizophrenia. Department of Veterans Affairs Cooperative Study Group on Clozapine in Refractory Schizophrenia Health Serv Res 1998 33 1237 1261 9865219
Pyne JM Sullivan G Kaplan R Williams DK Comparing the sensitivity of generic effectiveness measures with symptom improvement in persons with schizophrenia Med Care 2003 41 208 217 12555049 10.1097/00005650-200302000-00003
Rosenheck R Cramer J Xu W Thomas J Henderson W Frisman L Fye C Charney D A comparison of clozapine and haloperidol in hospitalized patients with refractory schizophrenia. Department of Veterans Affairs Cooperative Study Group on Clozapine in Refractory Schizophrenia N Engl J Med 1997 337 809 815 9295240 10.1056/NEJM199709183371202
Rosenheck R Perlick D Bingham S Liu-Mares W Collins J Warren S Leslie D Allan E Campbell EC Caroff S Corwin J Davis L Douyon R Dunn L Evans D Frecska E Grabowski J Graeber D Herz L Kwon K Lawson W Mena F Sheikh J Smelson D Smith-Gamble V Department of Veterans Affairs Cooperative Study Group on the Cost-Effectiveness of Olanzapine Effectiveness and cost of olanzapine and haloperidol in the treatment of schizophrenia: a randomized controlled trial JAMA 2003 290 2693 2702 14645311 10.1001/jama.290.20.2693
Lenert LA Sturley A Rupnow M Toward improved methods for measurement of utility: automated repair of errors in elicitations Med Decis Making 2003 23 67 75 12583456 10.1177/0272989X02239649
|
16153308
|
PMC1262745
|
CC BY
|
2021-01-04 16:38:14
|
no
|
Health Qual Life Outcomes. 2005 Sep 11; 3:57
|
utf-8
|
Health Qual Life Outcomes
| 2,005 |
10.1186/1477-7525-3-57
|
oa_comm
|
==== Front
Health Qual Life OutcomesHealth and Quality of Life Outcomes1477-7525BioMed Central London 1477-7525-3-581616805010.1186/1477-7525-3-58ResearchModification of the asthma quality of life questionnaire (standardised) for patients 12 years and older Juniper Elizabeth F [email protected] Klas [email protected]örk Ann-Christin [email protected]åhl Elisabeth [email protected] Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Canada2 AstraZeneca R&D, Lund, Sweden2005 16 9 2005 3 58 58 8 7 2005 16 9 2005 Copyright © 2005 Juniper et al; licensee BioMed Central Ltd.2005Juniper et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The age limit for some adult asthma clinical trials has recently been lowered to 12 years. In this study we have made minor modifications to the standardised version of the adult Asthma Quality of Life Questionnaire (AQLQ(S)) to make it valid for patients 12 years and older (AQLQ12+).
Methods
We have used two clinical trial databases, in which the AQLQ12+ was used, to compare the measurement properties of the questionnaire in patients 12–17 years and patients 18 years and older. A total of 2433 patients (12–75 years), with current asthma and with data that could be evaluated both at randomisation and end of treatment, were included.
Results
The analysis showed that internal consistency, responsiveness and correlations with other clinical indices were very similar in patients 12–17 years and patients 18 years and older.
Conclusion
The measurement properties of the AQLQ12+ are similar in adolescents and adults and therefore the instrument is valid for use in adult studies which include children 12 years and older.
==== Body
Background
The Asthma Quality of Life Questionnaire (AQLQ) was developed to measure the functional impairments experienced by adults 17 years and older [1]. It has 32 items in four domains (symptoms, activity limitations, emotional function and environmental stimuli). In the original AQLQ, five of the activity questions are patient-specific but this proved time-consuming and troublesome in large clinical trials. To address this problem, we developed the standardised version (AQLQ(S)). In the AQLQ(S) all the activity questions are generic and its measurement properties that are almost the same as those of the original AQLQ [3]. The Paediatric Asthma Quality of Life Questionnaire was developed to measure the problems that children 7–17 years experience [2]. It has 23 items in three domains (symptoms, activity limitations and emotional function). Recently, adolescents 12 years and older have been included in adult asthma clinical trials. To avoid using two separate health-related quality of life questionnaires in these studies, we have modified the AQLQ(S) [3] to make it suitable for both adolescents and adults. The AQLQ(S) was selected in preference to the original AQLQ because it is the version of the questionnaire most commonly used in clinical trials.
The aim of this adaptation was to ensure that the problems that are most troublesome to adolescents were included whilst making as few modifications to the original as possible. We have used two clinical trial databases to compare the measurement properties of the AQLQ(S) for 12 years and older (AQLQ12+) in patients 12–17 years and patients ≥ 18 years.
Methods
Modification of the AQLQ(S)
Both the AQLQ [1] and the Paediatric Asthma Quality of Life Questionnaire (PAQLQ) [2] were developed by asking adults and children respectively about the problems and limitations that are most important to them in their daily lives as a result of their asthma. Items that were most frequently experienced and most troublesome for the two groups and which are included in the two questionnaires [1,2] are shown in Tables 1 and 2. After reviewing the two questionnaires, only one word needed to be added to the AQLQ(S) to make it suitable for adolescents 12 years and older. As can be seen in Tables 1 and 2, symptoms and activity limitations are very similar in adults and children and the only change necessary was to alter the activity question about 'work-related limitations' to ask about 'work-/school-related limitations'. Although children identified sleep as a troublesome activity, this is already included in the symptom domain of the AQLQ. There is no environmental stimuli domain in the PAQLQ because children tend to express their problems with the environment in terms of activity limitation. For instance, an adult will say 'I am bothered by cigarette smoke', a young child will say 'I can't go to my friend's house because her mum smokes'. We considered that as adolescents (12–17 years) are moving towards adulthood, they would be old enough to express directly, rather than indirectly, their problems with environmental stimuli. Although children experience similar fears and frustrations to adults, they also 'feel different and left out'. Since none of the emotional function questions in the AQLQ could be modified to take this into account and since adding a separate question would have altered the weighting of the domain and overall score, 'feel different and left out' has not been included in the AQLQ12+.
Table 1 Functional impairments most important to adults (17–70 years) (1)
Symptoms Emotions Activities Environment
Short of breath Afraid of not having medications available Exercise/sports Cigarettes
Chest tightness Hurrying Dust
Wheeze Afraid of getting out of breath Social activities Air pollution
Cough Concerned about the need to use medications Pets Cold air
Tired Frustrated Work/housework Pollen
Table 2 Functional impairments most important to children (7–17 years) (2)
Symptoms Activities Emotions
Short of breath Sports and games Feel different and left out
Chest tightness Activities with friends Frustrated
Cough Playing with pets Angry
Wheeze School activities Sad
Tired Sleeping Frightened/anxious
Studies and patients
The analysis was conducted using databases from two clinical trials. Full details of one trial have been published elsewhere [4]. The second trial has been published as abstracts [5,6]. The first trial was a 12-month, randomised, double-blind, parallel group study comparing two active interventions. Of the 1890 patients randomised, 1770 completed the AQLQ12+ at baseline and either at the end of 12 months or on withdrawal. The second study was a 12-week randomised trial comparing three active interventions. Of the 680 patients randomised, 655 completed the AQLQ12+ at randomisation and either at 12 weeks or withdrawal. In both studies, patients were required to have inadequately controlled asthma with no evidence of any other respiratory disease and to be between the ages of 12 and 80 years.
Outcomes
Asthma Quality of Life Questionnaire for 12 years and older (AQLQ12+)
Patients are asked to recall their experiences during the previous 2 weeks and to score each of the 32 questions on a 7-point scale from 7 = no impairment to 1 = severe impairment. The overall score is calculated as the mean response to all questions. The four domain scores (symptoms, activity limitations, emotional function and environmental stimuli) are the means of the responses to the questions in each of the domains.
Symptom and Medication Diary
Each morning and evening patients scored how much they were bothered by their asthma symptoms on a 4-point scale (0 = no symptoms and 3 = unable to do normal daily activities (or sleep) because of asthma) and recorded the amount of rescue medication taken. Each morning they recorded whether they had been woken during the night by asthma symptoms. They also measured pre-bronchodilator PEF each morning and evening, recording the best of three blows. For this analysis, we have estimated the mean diary scores for the 2 weeks that were co-incident with the AQLQ12+ two week recall period.
Spirometry
Pre-bronchodilator FEV1% predicted normal was recorded at all clinic visits.
Statistical analysis
Since the validity of the AQLQ(S) has already been established in adults and because only one word was added to the AQLQ(S) for the modification, AQLQ12+ scores for adults (18–80 years) have been considered the gold standard for this analysis (criterion validity). Data collected at baseline were used to determine differences between age groups (unpaired t-test) and internal consistency (Cronbach's alpha). Change in scores between baseline and end of treatment, adjusted for treatment effect and baseline values by a linear ANOVA model, were used to determine responsiveness. Cross-sectional and longitudinal construct validity were evaluated by examining Pearson correlation coefficients between the AQLQ12+ and both diary symptoms and airway calibre.
Results
2423 patients were included in the analysis. There were 2207 over 18 years and 216 between 12 and 17 years (Table 3). In the older patients there were slightly more women than men and in the younger patients slightly more men than women. FEV1% predicted was slightly higher in the younger patients.
Table 3 Demographics and baseline values
Study 1 Study 2
≥ 18 years 12–17 years ≥ 18 years 12–17 years
Number of patients 1652 116 555 100
Mean age (range) 44.8 (18–80) 14.3 (12–17) 44.6 (18–79) 13.9 (12–17)
Gender M/F (%) 41.1/58.9 57.8/42.2 33.5/66.5 53.0/47.0
FEV1% pred. (range) 75.4 (32–136) 83.9 (47–125) 73.3 (41–107) 77.8 (54–114)
At baseline in both studies, there was no evidence of any difference in AQLQ12+ scores both for overall AQLQ12+ scores and for the symptom and emotional function domain scores (Table 4). However, in study 1, 12–17 year old patients were less troubled by activity limitations and environmental stimuli than older patients but the differences were small (< 0.35) and cannot be considered of clinical importance [5]. These differences were not seen in study 2. After adjusting for treatment and baseline values, changes in AQLQ12+ scores during the treatment period were similar in the two age groups for both studies (Table 4). As further evidence of the validity of the AQLQ12+ in adolescents, internal consistency was similar in the two age groups (Table 4) and correlations between each domain of the AQLQ12+ and other measures of asthma clinical status were also very consistent in the two age groups (Tables 5 and 6).
Table 4 Standardised version of the Asthma Quality of Life Questionnaire for 12 years and older (AQLQ12+)
AQLQ12+ Mean Score at baseline Mean change score during treatment
(adjusted means) Internal Consistency at baseline
Cronbach's alpha
≥ 18 yr 12–17 yr p value ≥ 18 yr 12–17 yr p value ≥ 18 yr 12–17 yr
Study 1
Overall 4.95 5.14 0.063 0.58 0.47 0.17 0.96 0.95
Symptoms 4.87 4.96 0.40 0.65 0.57 0.41 0.93 0.92
Activity limitation 5.09 5.41 0.002 0.52 0.39 0.12 0.91 0.86
Emotional function 5.06 5.11 0.68 0.57 0.46 0.26 0.87 0.77
Environmental stimuli 4.64 4.94 0.028 0.54 0.4 0.16 0.82 0.77
Study 2
Overall 4.69 4.72 0.82 0.66 0.59 0.52 0.97 0.97
Symptoms 4.68 4.64 0.73 0.73 0.67 0.60 0.94 0.95
Activity limitation 4.81 4.96 0.25 0.57 0.51 0.57 0.92 0.91
Emotional function 4.74 4.68 0.70 0.76 0.62 0.27 0.86 0.91
Environmental stimuli 4.24 4.43 0.24 0.57 0.56 0.96 0.84 0.88
Table 5 Cross-sectional construct validation (Baseline) (Pearson correlation coefficients)
Study 1
AQLQ12+ Age FEV1% pred PEF Symptoms Night waking Rescue bd
Overall ≥ 18 yr 0.15 0.29 -0.50 -0.50 -0.35
12–17 yr -0.11 0.13 -0.37 -0.38 -0.21
Symptoms ≥ 18 yr 0.14 0.22 -0.53 -0.54 -0.39
12–17 yr -0.11 0.15 -0.34 -0.40 -0.23
Activities ≥ 18 yr 0.14 0.32 -0.47 -0.45 -0.29
12–17 yr -0.13 0.15 -0.37 -0.35 -0.16
Emotions ≥ 18 yr 0.15 0.25 -0.36 -0.35 -0.28
12–17 yr -0.10 0.05 -0.28 -0.27 -0.23
Environment ≥ 18 yr 0.07 0.24 -0.34 -0.36 -0.22
12–17 yr -0.01 0.06 -0.25 -0.25 -0.12
Study 2
AQLQ12+ Age FEV1% pred PEF Symptoms Night waking Rescue bd
Overall ≥ 18 yr 0.06 0.23 -0.49 -0.47 -0.35
12–17 yr 0.03 0.37 -0.46 -0.39 -0.28
Symptoms ≥ 18 yr 0.06 0.17 -0.56 -0.54 -0.42
12–17 yr 0.04 0.32 -0.49 -0.42 -0.32
Activities ≥ 18 yr 0.06 0.28 -0.40 0.41 -0.28
12–17 yr 0.05 0.35 -0.42 -0.41 -0.21
Emotions ≥ 18 yr 0.06 0.18 -0.36 -0.36 -0.25
12–17 yr -0.02 0.36 -0.32 -0.31 -0.25
Environment ≥ 18 yr 0.03 0.21 -0.34 -0.34 -0.23
12–17 yr -0.01 0.40 -0.43 -0.36 -0.25
Table 6 Longitudinal construct validity (Baseline – End of study) (Pearson correlation coefficients)
Study 1
AQLQ12+ Age FEV1% pred PEF Asthma symptoms Night-time awakening Rescue medication use
Overall ≥ 18 yr 0.3 0.33 -0.43 -0.42 -0.39
12–17 yr 0.06 0.29 -0.33 -0.2 -0.24
Symptoms ≥ 18 yr 0.32 0.34 -0.46 -0.46 -0.43
12–17 yr 0.03 0.3 -0.33 -0.26 -0.25
Activity limitation ≥ 18 yr 0.27 0.28 -0.39 -0.37 -0.33
12–17 yr 0.08 0.26 -0.33 -0.18 -0.21
Emotional function ≥ 18 yr 0.25 0.29 -0.3 -0.31 -0.33
12–17 yr -0.06 0.12 -0.22 -0.11 -0.2
Environmental stimuli ≥ 18 yr 0.16 0.19 -0.25 -0.2 -0.2
12–17 yr 0.19 0.3 -0.21 0 -0.16
Study 2
AQLQ12+ Age FEV1% pred PEF Asthma symptoms Night-time awakening Rescue medication use
Overall ≥ 18 yr 0.21 0.41 -0.51 -0.41 -0.37
12–17 yr 0.23 0.24 -0.45 -0.3 -0.2
Symptoms ≥ 18 yr 0.23 0.4 -0.56 -0.47 -0.43
12–17 yr 0.26 0.29 -0.47 -0.31 -0.24
Activity limitation ≥ 18 yr 0.17 0.36 -0.43 -0.33 -0.3
12–17 yr 0.23 0.15 -0.44 -0.28 -0.1
Emotional function ≥ 18 yr 0.14 0.33 -0.4 -0.28 -0.27
12–17 yr 0.1 0.17 -0.21 -0.22 -0.23
Environmental stimuli ≥ 18 yr 0.17 0.36 -0.32 -0.28 -0.2
12–17 yr 0.16 0.23 -0.48 -0.22 -0.15
Discussion
The results of this analysis have shown that measurement properties of the AQLQ12+ in both adolescents and adults are very similar and that the AQLQ12+ can therefore be used in adult clinical trials that include adolescents.
Since only one word change was needed to make the AQLQ(S) to be suitable for adolescents and because both the original AQLQ and the AQLQ(S) have undergone extensive validation in adults [7-11], we have considered the AQLQ12+ in adults (≥18 years) to be the gold standard with which to compare the measurement properties of the AQLQ12+ in adolescents. In both studies at baseline, there was no evidence of any differences in the overall or domain scores except for the activity limitation and environmental stimuli domains in study 1, where small but clinically unimportant differences were observed (the minimal important difference for the AQLQ is 0.5 on the 7-point scale [12]). Changes in scores during treatment and internal consistency were very similar in both age groups in both studies. These data strongly support the validity of the AQLQ(S)12+ in adolescents. There was a very slight tendency for correlations with other clinical indices to be slightly lower in adolescents but this is most likely attributable to a slight difference in the relationship between quality of life and clinical asthma in the two age groups. Even if this is not the reason, the tendency is so small that it not sufficient to suggest lack of validity of the AQLQ12+ in adolescents.
The results of this analysis should not be interpreted to mean that the AQLQ12+ is the most appropriate asthma-specific quality of life questionnaire for adolescents. To evaluate the impact of asthma on individual adolescents in the clinic or to estimate the effect of interventions on adolescents alone, it would be wise to continue to use an instrument that has been specifically developed for this age group. The PAQLQ [2], for instance, includes all the problems that children between 7 – 17 years have identified as important and uses the words that they are most likely to use.
Conclusion
The results of this analysis suggest that the AQLQ12+ is valid for measuring asthma-specific quality of life in adolescents 12–17 years. The similarity of the measurement properties of the AQLQ12+ in patients 12–17 years and over 18 years provides evidence that data from the two groups can be combined for analysis of adult clinical trials and surveys that included patients 12 years and older.
List of abbreviations
AQLQ Asthma Quality of Life Questionnaire
AQLQ(S) Standardised version of the Asthma Quality of Life Questionnaire
AQLQ12+ Standardised version of the Asthma Quality of Life Questionnaire for 12 years and older
FEV1 Forced Expiratory Volume in 1 second
PAQLQ Paediatric Asthma Quality of Life Questionnaire
Authors' contributions
EFJ: Design of the analysis, interpretation of data, primary author of manuscript.
KS: Statistical analysis, development of study question, drafting of manuscript.
ACM: Development of study question, drafting of manuscript, study co-ordinator.
ES: Development of study question, drafting of manuscript.
All four authors have played a major part throughout the entire study process from the development of the study question to the revision of the final manuscript. Each of the four authors has made a significant contribution at each phase of the study.
All four authors have reviewed and approved the final version of this manuscript.
Acknowledgements
This study was supported by Financial Support from AstraZeneca, R&D Lund, Sweden.
==== Refs
Juniper EF Guyatt GH Epstein RS Ferrie PJ Jaeschke R Hiller TK Evaluation of impairment of health-related quality of life in asthma: development of a questionnaire for use in clinical trials Thorax 1992 47 76 83 1549827
Juniper EF Guyatt GH Feeny DH Ferrie PJ Griffith LE Townsend M Measuring quality of life in children with asthma Qual Life Res 1996 5 35 46 8901365 10.1007/BF00435967
Juniper EF Buist AS Cox FM Ferrie PJ King DR Validation of a standardized version of the Asthma Quality of Life Questionnaire Chest 1999 115 1265 1270 10334138 10.1378/chest.115.5.1265
Scicchitano R Aalbers R Ukena D Manjara A Fouquert L Centanni S Boulet LP Naya IP Hultquist C Efficacy and safety of budesonide/formoterol single inhaler therapy versus a higher dose of budesonide in moderate to severe asthma Curr Med Res Opin 2004 20 1403 1418 15383189 10.1185/030079904X2051
Morice AH Osmanliev D Arheden L Beckman O Therapeutic equivalence of a novel budesonide/formoterol pMDI versus budesonide/formoterol Turbuhaler(R) in adolescents and adults with asthma J Allergy Clin Immunol 2005 115 S2 Abs 8 10.1016/j.jaci.2004.12.025
Morice AH Kukova Z Arheden L Beckman O The novel budesonide/formoterol pMDI is therapeutically equivalent to budesonide/formoterol Turbuhaler(R) in children with asthma J Allergy Clin Immunol 2005 115 S209 Abs 833 10.1016/j.jaci.2004.12.846
Juniper EF Guyatt GH Ferrie PJ Griffith LE Measuring quality of life in asthma Am Rev Respir Dis 1993 147 832 838 8466117
Sanjuas C Alonso J Sanchis J Casan P Broquetas JM Ferrie PJ Juniper EF Anto JM The quality of life questionnaire with asthma patients; the Spanish version of the Asthma Quality of Life Questionnaire Arch Bronconeumol 1995 31 219 226 7788083
Leidy NK Coughlin C Psychometric performance of the Asthma Quality of Life Questionnaire in a US sample Qual Life Res 1998 7 127 134 9523494 10.1023/A:1008853325724
Rutten-van Molken MPMH Clusters F Van Doorslaer EKA Jansen CCM Heurman L Maesen FPV Smeets JJ Bommer AM Raaijmakers JAM Comparison of performance of four instruments in evaluating the effects of salmeterol on asthma quality of life Eur Respir J 1995 8 888 898 7589374
Juniper EF Norman GR Cox FM Roberts JN Comparison of the standard gamble, rating scale, AQLQ and SF-36 for measuring quality of life in asthma Eur Respir J 2001 18 38 44 11510803 10.1183/09031936.01.00088301
Juniper EF Guyatt GH Willan A Griffith LE Determining a minimal important change in a disease-specific quality of life instrument J Clin Epidemiol 1994 47 81 87 8283197 10.1016/0895-4356(94)90036-1
|
16168050
|
PMC1262746
|
CC BY
|
2021-01-04 16:38:13
|
no
|
Health Qual Life Outcomes. 2005 Sep 16; 3:58
|
utf-8
|
Health Qual Life Outcomes
| 2,005 |
10.1186/1477-7525-3-58
|
oa_comm
|
==== Front
Health Res Policy SystHealth Research Policy and Systems1478-4505BioMed Central London 1478-4505-3-61620215310.1186/1478-4505-3-6CommentaryResearch ethics committees: agents of research policy? Hemminki Elina [email protected] Health and Social Services, National Research and Development Centre for Welfare and Health STAKES, P.O. BOX 220, 00531 Helsinki, Finland2005 4 10 2005 3 6 6 31 1 2005 4 10 2005 Copyright © 2005 Hemminki; licensee BioMed Central Ltd.2005Hemminki; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The purpose of this commentary is to describe the unintended effects ethics committees may have on research and to analyse the regulatory and administrative problems of clinical trials.
Discussion
The Finnish law makes an arbitrary distinction between medical research and other health research, and the European Union's directive for good clinical trials further differentiates drug trials. The starting point of current rules is that clinical trials are lesser in the interest of patients and society than routine health care. However, commercial interests are not considered unethical. The contrasting procedures in research and normal health care may tempt physicians to continue introducing innovations into practice by relying on unsystematic and uncontrolled observations. Tedious and bureaucratic rules may lead to the disappearance of trials initiated by researchers. Trying to accommodate the special legislative requirements for new drug trials into more complex interventions may result in poor designs with unreliable results and increased costs. Meanwhile, current legal requirements may undermine the morale of ethics committee members.
Conclusion
The aims and the quality of the work of ethics committees should be evaluated, and a reformulation of the EU directive on good clinical trials is needed. Ethical judgement should consider the specific circumstance of each trial, and ethics committees should not foster poor research for legal reasons.
ethics committeesclinical trialsethicslaw
==== Body
Introduction
Various ethical rules and, lately, an increasing amount of legislation have been introduced to protect research participants. Ethical codes target the researchers, but ethical committees have come into existence to aid the researcher and to ensure that rules are followed. Ethical rules and codes have emerged within medical research, and still today in most countries, the legal requirements for research protocols to be checked by ethics committees are confined to medical research.
Clinical trials are used variously to obtain unbiased estimates of the value of health technologies, for administrative purposes (e.g. for registering medicines), to get opinion leaders to commit, and for marketing. Currently, most clinical trials concern new drugs or new indications for drugs. Dorman et al.'s [1] analysis of trials in acute stroke from 1955–95 showed that only 12% tested non-drug interventions. Active and successful research in biomedicine along with powerful financing – most commonly by the drug firms – have made drug trials a central element of clinical research.
The main aim of ethics committees is to protect patients involved in research. However, it has been asked whether the committees do more harm than good, with researchers expressing concern at the negative impact on research [see for e.g. [2,3]]. Experts have identified serious problems in the processes [4,5,3], while a great variability in the criteria used by the committees is reported both within and between countries [6-10,3], with most of the empirical studies looking at ethics committees come from the U.K. The problems listed include too much work for voluntary committee members, slowness (crucial for studies with time-limited budgets), varying criteria used by different committees, a lack of training of the members of ethics committees, repeat handling of multicenter trials, and excessive costs. Less criticism has been made of omissions – what is allowed to happen that should not be [11], or the waste of limited clinical research capacity on trivial research [12].
The purpose of this article is to describe the unintended effects ethics committees may have on research, using Finland as an example and focusing on clinical trials. I will analyse some of the regulatory and administrative problems associated with clinical trials and suggest solutions.
Finland is a Northern European welfare state with a population of about 5 million people. Trials of new drugs are common in Finland (tenfold higher than might be expected on the basis of the size of the Finnish population) [13], with 287 new drug trials reported to the drug regulatory authority in 2001 [14].
This article is based on observations drawn from reports of clinical trials, both published and unpublished; authoritative advice for conducting trials; my own experience of submitting research to ethics committees, judging research protocols in funding organizations and sitting in Finnish ethics committees. Much of my experience is based on material that is, or has been, confidential, and so I will amalgamate the information for purposes of anonymity.
Discussion
The regulation of clinical trials and ethics committees
In Finland, ethics committees were first established voluntarily in the 1970s, whilst later their tasks and composition were regulated through a 1999 Act on medical research [15,13]. Only medical research ("research, which affects human integrity and is led by physicians") is expected to be submitted to ethics committees for prior evaluation. The European Union good clinical trials directive [16] was integrated into the Finnish national law in spring 2004, differentiating drug trials as a subgroup that is subject to more detailed regulation. In addition, the directive added many administrative and surveillance tasks in regard to drug trials, some of which are only vaguely related to ethics, such as insurance, compensation of injuries, and external surveillance of adverse drug events.
The Finnish law requires that informed consent is obligatory in all drug trials, and the procedures are regulated in detail; in non-drug clinical trials, informed consent may be waived in exceptional circumstances. Informed consent is defined such that the people who are "the targets" of the trial will be included in the study only after the trial has been described to them honestly and in detail, and they have then voluntarily given their permission, free of any pressures. The consent "must be written, dated and signed", and in the case of incapable persons, this should be made by his or her legal representative. If the person cannot write, oral consent may be accepted in the presence of at least one witness [16]. The Finnish Act (Sublaw 986/1999) gives a detailed list of the formal requirements of informed consent and additional details on how to ask for informed consent and which kinds of documents are needed are given by drug control authorities and the central ethics committee. One requirement is that a copy of the signed consent is to be given to the patient or other research participant, suggesting a contract arrangement.
Informed consent on data acquisition and its use is much vaguer. In general, the legal and ethical issues of data use in health research are a muddle due to law changes over time, historic data sets, commercial interest, and special questions of genetic data or other biological samples. In multinational studies, the issue is complicated by the effects of varying data protection laws and their interpretation in different countries. Many researchers have decided to avoid the problems by asking for informed consent from the patient to also use his/her data outside the current research frame. This consent is likely to be uninformed, because an average patient does not know, for example, what information can be obtained from a blood sample or what data national registers contain, or what does anonymity of genetic samples imply. In clinical trials, informed consent for data storage or transfer often ends up being a quasi-action from an ethical point of view.
Current ethical rules in Finland do not classify commercial interests as unethical, and they do not need to be revealed to trial participants. Patients may be asked to join trials with designs favouring the studied therapy or they are asked to join trials intended to accustom physicians to prescribing a certain therapy through their trial participation. A large number of trials remain unpublished (within pharmaceuticals, this is more than half), and those favouring the therapy being studied are more likely to be published [17-22]. A common motive for patients to take part in research is an altruistic wish to help medicine develop for the common good, and were the patients to know that the trial was created for commercial reasons and/or the results would remain unavailable, they may not have participated.
Double standards on informed consent
Current rules for Finnish ethics committees – for example, as expressed in the EU directive on good clinical trials – start from the idea that clinical trials are, by definition, lesser in the interests of the patient and society than routine health care. The ethical codes do not explain why an intervention that is already used in patient care automatically requires an informed consent within a research setting [23-25]. In ordinary health care, the same intervention may be prescribed by less experienced and knowledgeable practitioners. In the words of the paediatrician Richard Smithells: "I need permission to give a drug to half of my patients, but not to give it to them all" [cited by [23]].
Even though the idea of a physician being a consultant to help the patient make an informed choice has been put forward, it is likely that this consumerism in health care will only expand into issues where lay-persons have the main responsibility for health decisions, such as in the field of prevention and some chronic diseases. In medical research, current ethics have adopted the consumer model, largely putting the responsibility on the patient regardless of the nature of the issue being studied.
The contrast between the procedures required in research and normal health care is striking. Additional research requirements may tempt physicians to continue the old method of introducing innovations into practice by relying on unsystematic and uncontrolled observations. In the case of new drugs, the inflexible requirement of informed consent in emergency situations with critically ill patients may be unethical and/or unfeasible. It may result in the future with emergency medication being based on uncontrolled experimentation.
In practice, asking for informed consent is often something of a performance, in which patients and physicians are acting as if a truly informed consent was asked for. In many situations, patients cannot or prefer not to make an informed choice [26]. Moreover, even if such an attempt at truly informed consent is made, it is not often successful. Diseases and ill health are concentrated among the aged. Most patient information leaflets and consent forms are long with difficult language [27,28].
Impact on ethics committees
In many countries, ethics committees have been transformed from bodies providing advice into administrative bodies observing that rules are followed [for the U.K., see [29]]. If in conflict, the law precedes ethics. The legal requirements used in judging the ethicality of clinical trials may undermine the morale of the members of ethics committees. Giving positive statements on research protocols fulfilling legal requirements, yet wasting resources and bringing no value to health may create cynicism and decrease true interest in their work. Such situations include accepting marketing research disguised as scientific research, or accepting patient information leaflets which contain all the necessary information, but which are unlikely to be understood by the target group. Giving a negative statement on important ethical research that does not fulfil legal requirements may also undermine morale.
Impact on research orientation
Tedious and bureaucratic rules may result in more and more incentives being necessary to persuade physicians and their employers to carry out trials, making trials more expensive. Trials initiated by researchers may disappear and only those trials having a rich sponsor will survive. We are already in a situation where most drug trials are paid for by drug companies [1,30-33]. In the U.K., non-commercial sources have also extensively supported clinical trials, including non-drug trials, but in recent years there has been a clear decline [34].
"Ethical" codes and legislation may lead to trials becoming tedious, expensive and factory-like, alienating interested minds and health service providers. In the worst scenario, research resources are wasted, answers are received to unimportant questions, and scientists turn to other types of health research.
The current Finnish legislation and many international codes have been made to accommodate the special requirements of new drug trials and they do not fit well into established therapies or into more complex interventions, such as prevention and ways of organizing services. Trying to accommodate the legal requirements may result in poor designs with unreliable results and increased costs, and interest in studying complex interventions may diminish. For example, if the requirement of informed consent is interpreted rigidly, cluster randomization would become unfeasible.
Conclusion
Comparative trials that answer important health or health service questions and which are not biased by commercial interests are needed. They should be applicable to real life situations, and use resources prudently to allow many questions to be studied.
The research question and field circumstances should determine how clinical trials should be done. When the Helsinki Declaration was formulated, Bradford Hill [35], an innovator behind clinical trials, claimed that there is no one way of doing clinical trials ethically, and giving detailed advice as if there were will harm both research and ethics. He argued that ethical judgement has to consider the specific circumstances of each trial. General advice for trial design and ethics are useful in giving inexperienced researchers help in their work, but as Foster [29] has argued, to decide what is ethically appropriate requires a thoughtful balancing between different moral approaches and cannot simply be substituted by regulation and rules.
It seems that ethics committees have sometimes become an extra burden instead of an aid to bettering clinical trials. The external review by ethics committees should be advisory and they should not be censoring and preventing research, but advising and helping researchers to carry out responsible research. Ethics committees should judge the ethical components, free from rigid detailed rules, guided by general principles, enriched in international debate. Ethics include the fair use of health care resources and the potential value of the study. The burden of judging the benefits and risks should not be put solely onto individual patients via informed consent.
The concept and practical application of informed consent should be rethought for trials with interventions which can be used without informed consent in everyday practice. Normal health care and research on existing practices should have similar ethical rules [24,25]. Opting out and "non-compliance" are the rights of a person, both in research and normal care. But people cannot (individually) beforehand decline from being asked to enter mass screening, or choose hospital wards randomized to varying (established) treatment policies. Their role in deciding on emergency care is limited, too. Whether or how individuals are informed about the trial should depend on the intervention. For screening, Irwig et al. [36] have proposed a survey of the target population's interest in participating after being fully informed before the offer of screening. For trials comparing different treatment policies (pragmatic trials), providing information to and gaining permission from the communities in which the research is being carried out is an option. For collective permission to be useful, it may, however, need public education on what the trials are and why they are valuable.
Participation in trials with non-commercial interests should be seen as a professional responsibility [37], and clinical trials with existing therapies or service provision should be considered a part of health services.
The role of ethics committees should be expanded to cover commercial interests. Ethics committees should guarantee that all potential participants, both physicians and patients, are aware of the financer, and what, if any, are the commercial aims, as well as what compensation is paid, and whether results will be publicly available.
The rules and legislation governing the work of ethics committees as well as the quality of their work should be evaluated. Observations suggest that ethics committees are doing tasks which do not suit them, and which prevent them from concentrating on the real issues. Furthermore, there is a worry that the new EU legislation may worsen the opportunities to do trials which are in the patients' interest. An urgent task in Europe is to reformulate the EU directive on good clinical trials and to discuss ethics from a wider perspective.
Competing interests
The author(s) declare that they have no competing interests.
Acknowledgements
I thank Piret Veerus for her valuable comments.
==== Refs
Dorman PJ Counsell C Sandercock P Reports of Randomized Trials in Acute Stroke, 1955 to 1995. What Proportions Were Commercially Sponsored? Stroke 1999 30 1995 1998 10512897
Doll R What are the effects of the fifth revision of the Declaration of Helsinki: research will be inpeded BMJ 2001 323 1421 1422
Glasziou P Chalmers I Ethics review roulette: what can we learn? That ethics review has costs and one size doesn't fit all BMJ 2004 328 121 122 14726317 10.1136/bmj.328.7432.121
Saunders J Research ethics committees – time for change? Clin Med 2002 2 534 538 12528967
Jamrozik K The case for a new system for oversight of research on human subjects Med Ethics 2000 26 334 339 10.1136/jme.26.5.334
Hotopf M Wessely S Noah N Are ethical committees reliable? J R Soc Med 1995 88 31 33 7884767
Redshaw ME Harris A Baum JD Research ethics committee audit: differences between committees J Med Ethics 1996 22 78 82 8731532
Jamrozik K Kolybaba M Are ethics committees retarding the improvement of health care services in Australia? Med J Aust 1999 170 26 28 10026670
Maskell NA Jones EL Davies RJO Variations in experience in obtaining local ethical approval for participation in a multi-centre study Q J Med 2003 96 305 307
Hearnshaw H Comparison of requirements of research ethics committees in 11 European countries for a non-invasive interventional study BMJ 2004 328 140 141 14726341 10.1136/bmj.328.7432.140
Savulescu J Chalmers I Blunt J Are research ethics committees behaving unethically? Some suggestions for improving performance and accountability BMJ 1996 313 1390 1393 8956711
Chalmers I Current Controlled Trials: an opportunity to help improve the quality of clinical research Curr Control Trials Cardiovasc Med 2000 1 3 8 11714397 10.1186/CVM-1-1-003
Keinonen T Clinical drug trials in Finland Quality and characteristics 2003 Kuopio University Publications A, Pharmaceutical Sciences 66. Kuopio: University of Kuopio
National Agency for Medicines and KELA Finnish Statistics on Medicines 2001 2002 Helsinki: Edita Prima Oy
Halila R The role of national ethics commissions in Finland Bioethics 2003 17 357 368 14569982 10.1111/1467-8519.00350
Directive 2001/20/EC of the European Parliament and of the Council of 4 April 2001 on the approximation of the laws, regulations and administrative provisions of the Member States relating to the implementation of good clinical practice in the conduct of clinical trials on medicinal products for human use Official Journal of the European Communities No L121, 2001 May 1 44/34
Bardy A Bias in reporting clinical trials Br J Clin Pharmacol 1998 46 147 150 9723823 10.1046/j.1365-2125.1998.00759.x
Melander H Ahlqvist-Rastad J Meijer G Beermann B Evidence b(i)ased medicine – selective reporting from studies sponsored by pharmaceutical industry: review of studies in new drug applications BMJ 2003 326 1171 1173 12775615 10.1136/bmj.326.7400.1171
Garland EJ Facing the evidence: antidepressant treatment in children and adolescents Can Med Assoc J 2004 170 489 491 14970097
Svensson S Mansfield PR Escitalopram: superior to citalopram or a chiral chimera? Psychotherapy & Psychosomatics 2004 73 10 16 14665791 10.1159/000074435
Whittington C Kendall T Cottrel D Cotgrove A Boddington E Selective serotonin reuptake inhibitors in childhood depression: systematic review published versus unpublished data Lancet 2004 363 1341 1345 15110490 10.1016/S0140-6736(04)16043-1
Djulbegovic B Lacevic M Cantor A Fields KK Bennett CL Adams JR The uncertainty principle and industry-sponsored research Lancet 2000 356 635 638 10968436 10.1016/S0140-6736(00)02605-2
Chalmers I Lindley RI Doyal L, Tobias JS Double standards on informed consent to treatment Informed Consent in Medical Research 2000 BMJ Books 266 75
Chalmers I Trying to do more good than harm in policy and practice: The role of rigorous transparent, up-to-date evaluations The Annals of the American Academy of Political and Social Science 2003 589 22 40 10.1177/0002716203254762
Chalmers I The James Lind Initiative J R Soc Med 2003 96 575 576 14645604 10.1258/jrsm.96.12.575
Corrigan O Abraham J, Lawton SH The limitations of current ethical regulations Regulation of the pharmaceutical industry 2003 Palgrave, Mcmillan 195 211
Paasche-Orlow MK Taylor HA Brancati FL Readability standards for informed-consent forms as compared with actual readability N Engl J Med 2003 348 721 726 12594317 10.1056/NEJMsa021212
Flory J Emanuel E Interventions to Improve Research Participants' Understanding in Informed Consent for Research JAMA 2004 292 1593 1601 15467062 10.1001/jama.292.13.1593
Foster C Abraham J, Lawton SH Regulation for Ethical Purposes: Medical Research on Humans Regulation of the pharmaceutical industry 2003 Palgrave, Mcmillan 181 194
Yaphe J Edman R Knishkowy B Herman J The association between funding by commercial interests and study outcome in randomized controlled drug trials Family Practice 2001 18 565 568 11739337 10.1093/fampra/18.6.565
Bekelman JE Li Y Gross CP Scope and Impact of Financial Conflicts of Interest in Biomedical Research. A Systematic Review JAMA 2003 289 454 464 12533125 10.1001/jama.289.4.454
Bhandari M Busse JW Jackowski D Montori VM Schünemann H Sprague S Association between industry funding and statistically significant pro-industry findings in medical and surgical randomized trials CMAJ 2004 170 477 480 14970094
Buchkowsky SS Jewesson PJ Industry sponsorship and authorship of clinical trials over 20 years Ann Pharmacother 2004 38 579 585 14982982 10.1345/aph.1D267
Chalmers I Rounding C Lock K Descriptive survey of non-commercial randomised controlled trials in the United Kingdom, 1980–2002 BMJ 2003 327 1017 1020 14593034 10.1136/bmj.327.7422.1017
Bradford HillA Medical ethics and controlled trials Brit Med J 1963 5337 1043 1049 13954493
Irwig L Glasziou P Informed consent for screening by community sampling Effective Clinical Practice 2000 3 47 50 10788038
Chalmers I Well informed uncertainties about the effects of treatments. How should clinicians and patients respond? BMJ 2004 328 475 476 14988164 10.1136/bmj.328.7438.475
|
16202153
|
PMC1262747
|
CC BY
|
2021-01-04 16:37:15
|
no
|
Health Res Policy Syst. 2005 Oct 4; 3:6
|
utf-8
|
Health Res Policy Syst
| 2,005 |
10.1186/1478-4505-3-6
|
oa_comm
|
==== Front
Int J Behav Nutr Phys ActThe International Journal of Behavioral Nutrition and Physical Activity1479-5868BioMed Central London 1479-5868-2-131616228510.1186/1479-5868-2-13ResearchThe Physical Activity Resource Assessment (PARA) instrument: Evaluating features, amenities and incivilities of physical activity resources in urban neighborhoods Lee Rebecca E [email protected] Katie M [email protected] Jacqueline Y [email protected] Gail [email protected] Hugh H [email protected] Health and Human Performance, University of Houston, Garrison Gymnasium 104E, 4800 Calhoun Rd, Houston, TX 77204, USA2 Department of Psychology, University of Missouri-Kansas City, 4825 Troost, Suite 123, Kansas City, MO 64110, USA3 Psychology, University of Kansas, 315 Fraser Hall, 1415 Jayhawk Blvd., Lawrence, KS 66045-7556, USA4 Department of Psychology, Castleton State College,62 Alumni Dr., Castleton, VT 05735, USA5 Department of Geography, American River College, 4700 College Oak Dr., Sacramento, CA 95841, USA2005 14 9 2005 2 13 13 20 12 2004 14 9 2005 Copyright © 2005 Lee et al; licensee BioMed Central Ltd.2005Lee et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Neighborhood environment factors may influence physical activity (PA). The purpose of this study was to develop and test a brief instrument to systematically document and describe the type, features, amenities, quality and incivilities of a variety of PA resources.
Method
The one-page Physical Activity Resource Assessment (PARA) instrument was developed to assess all publicly available PA resources in thirteen urban lower income, high ethnic minority concentration neighborhoods that surrounded public housing developments (HDs) and four higher income, low ethnic minority concentration comparison neighborhoods. Neighborhoods had similar population density and connectivity. Trained field coders rated 97 PA resources (including parks, churches, schools, sports facilities, fitness centers, community centers, and trails) on location, type, cost, features, amenities, quality and incivilities. Assessments typically took about 10 minutes to complete.
Results
HD neighborhoods had a mean of 4.9 PA resources (n = 73) with considerable variability in the type of resources available for each neighborhood. Comparison neighborhoods had a mean of 6 resources (n = 24). Most resources were accessible at no cost (82%). Resources in both types of neighborhoods typically had about 2 to 3 PA features and amenities, and the quality was usually mediocre to good in both types of neighborhoods. Incivilities at PA resources in HD neighborhoods were significantly more common than in comparison neighborhoods.
Conclusion
Although PA resources were similar in number, features and amenities, the overall appearance of the resources in HD neighborhoods was much worse as indicated by substantially worse incivilities ratings in HD neighborhoods. The more comprehensive assessment, including features, amenities and incivilities, provided by the PARA may be important to distinguish between PA resources in lower and higher deprivation areas.
==== Body
Background
Regular physical activity has been associated with prevention of overweight and obesity, [1] as well as numerous other diseases [2]. Epidemiological studies have linked area of residence to physical activity [3-5]. Given consistently low rates of physical activity in the US, [2] it is important to develop an understanding of neighborhood factors that influence physical activity.
Residence in areas with high levels of material deprivation, an indicator of low area socioeconomic status (e.g., low median household income, and low educational attainment) has been associated with low levels of physical activity [3,5-7] as well as poor overall health [5,8] that may indicate low levels of physical activity. Despite the consistency of these findings, it is difficult to pinpoint the physical characteristics of material deprivation that contribute to physical inactivity. Deprived neighborhoods often have a lower proportion of owned homes compared to non-deprived neighborhoods, resulting in a low tax base for municipal improvements, have higher crime rates and reduced collective efficacy, [5,9,10] and have fewer goods and services available [7,11]. Although studies in two countries (UK, USA) have found that physical activity resources [12,13] vary by the socioeconomic status of the neighborhood, with lower SES areas having fewer physical activity resources, another study in Australia found greater access to physical activity resources in more deprived areas [14]. Although lack of access may be a driving factor to lower rates of physical activity in some deprived areas, there are likely additional qualitative elements of the physical activity resources that have not been well described or documented. For example, one study reported that attractiveness of public open space was associated with higher rates of walking [15]. However, there remain many elements of physical activity resources that have yet to be described or investigated in detail.
Although research in this area has been limited by access to geocoded databases to provide environmental information, the widespread adoption of Internet-based "user friendly" formats has resulted in a proliferation of research. Wider availability of these kinds of data have produced several studies investigating urban development constructs including walkability, a measure of how convenient and pleasant a neighborhood is for walking, [16,17] along with other street-scale elements, such as protected pedestrian pathways and availability of goods and services [18].
Despite these innovations, there remains no widely accepted system or protocol for describing or evaluating physical activity resources. The potential for endless permutations in type, equipment, size, shape, and condition of resources has hampered development of a common protocol. Further, until recently, little interest had been devoted to investigating the role that the physical environment plays in physical activity [19]. Recognition of physical activity as a universal good health recommendation, associated with the prevention of many chronic diseases, and success of policy and environmental approaches to manage health behaviors in other domains (e.g., tobacco control) has led researchers and funding agencies alike to pursue environmental approaches [19].
Available evidence suggests that there is a link between neighborhood factors and physical activity. Before this link can be clearly defined, it is important to develop strategies to assess neighborhood factors that may influence physical activity. The goals of this study were threefold. First, we aimed to develop an assessment protocol to describe physical activity resources based on existing literature and extensive pilot testing. Second, we assessed the type, quantity, features, amenities, and quality of all publicly accessible physical activity resources in urban neighborhoods available to public housing residents. Last, we compared the resources in the public housing neighborhoods to less deprived, comparison neighborhoods matched on age of housing and urban design (as measured by street network connectivity) and population density.
Method
Neighborhood Selection and Characteristics
Seventeen neighborhoods were selected for this study. Thirteen neighborhoods were defined as an 800 meter radius circumscribed around a public housing development managed by the state housing authority offices in Kansas City, Kansas and Missouri. Defining the neighborhood as the area within the boundaries of the circle has several advantages [20]. First, it captures all area to which a resident may be exposed on a daily basis during both foot and automobile travels. Second, the straight line distance allows for capture of distance traveled on footpaths and other "short cut" routes that may not be captured by using a street network strategy. Third, it may reduce the effect of spatial correlation that arises from using census boundaries where points near the boundary of the census area are influenced by factors in adjacent census areas, as housing developments were selected to be at least 1600 meters apart. Two housing developments violated this selection; however, they were separated by a major interstate that effectively eliminated daily exposure to the neighboring area.
Although Kansas City spans two states, it is a seamless city visually and practically. The public housing developments in this area serve a predominantly African-American population. Residents meet the 2003 US Department of Health and Human Service's Poverty guidelines, an annual household income of $18,400 or less per year for a family of four [21]. Aggregated neighborhood characteristics are presented in Table 1.
Table 1 Aggregated urban neighborhood characteristics.
Housing Development Neighborhoods Median Household Income Population Density Percent Ethnic Minority Street Connectivity
H01 29,299 5,548 62.1 118
H02 32,205 1,877 50.7 50
H03 16,902 3,925 71.1 108
H04 15,832 2,957 93.0 138
H05 11,930 1,770 98.1 94
H06 18,053 3,210 72.9 81
H07 18,644 2,960 70.3 54
H08 18,719 3,818 71.1 77
H09 22,519 4,087 60.6 93
H10 30,452 2,125 88.1 93
H11 34,303 2,779 71.4 84
H12 22,625 1,974 66.8 72
H13 25,844 3,218 52.6 98
Mean 22,871 3,096 71.4 89.2
(SD) (7,005) (1,073) (14.3) (24.2)
Comparison Neighborhoods
C01 38,099 3,664 14.6 93
C02 48,383 2,889 9.1 99
C03 43,006 3,403 18.2 107
C04 39,970 2,167 10.7 105
Mean 42,364 3,031 13.2 101
(SD) (4,493) (660) (4.1) (6.3)
All housing development neighborhoods were located in urban areas that were predominantly lower income, with higher proportions of ethnic minorities. Four comparison neighborhoods were selected in areas that were similar in age and urban design, but higher in income with lower proportions of ethnic minorities. In comparison neighborhoods, an 800 meter radius was circumscribed around the centroid of an apartment complex or set of buildings with multiple family residences that were similar in size and appearance to public housing developments.
Measures
Neighborhood level variables
United States Census data from the year 2000 were used to compute the aggregate median household income, population density and percentage of ethnic minorities for each neighborhood. All variables were drawn in aggregated form at the census block group level [22]. Neighborhoods often included parts of several block groups; thus, all values were calculated as weighted sums based on the overlap of housing development neighborhood buffer boundaries and block group boundaries [23,24]. For example, if a neighborhood buffer boundary encircled 30% of one block group with a median household income of $20,000, plus 20% of a second block group with a median household income of $25,000, plus 50% of a third block group with a median household income of $30,000, the weighted sum median income value for that neighborhood would be calculated as ($20,000 * 0.3) + ($25,000 * 0.2) + ($30,000 * 0.5) = $26,000. Population density was the number of people per square kilometer in each neighborhood. Proportion (reported as a percentage) of ethnic minorities was calculated as the sum of people identifying themselves as Black or African American alone; American Indian and Alaska Native alone; Asian alone; Native Hawaiian and Other Pacific Islander alone; some other race alone; or Hispanic, divided by the total population in that neighborhood. Street connectivity was calculated by counting the number of three or more street intersections in each neighborhood [18]. A three or more street intersection is any intersection where at least three streets are joined. These intersections may form a "T" or a "+" or a star shape when viewed from above.
Physical activity resources
The census of physical activity resources available to the general public was identified using a three step strategy [13]. First, Internet and telephone book searches were performed to generate an initial list of all physical activity resources in each neighborhood. Searches were done using an exhaustive list of terms identified previously [13] and built on by pilot testing for this study. Terms included physical activity, gym, fitness, dance studios, school, park, church, health clubs, bikeways, martial arts, sports, et cetera. All resources were mapped and verified initially by phone to confirm the presence and availability of the resource. Next, trained assessors conducted windshield surveys to confirm locations of resources and find any resources that had not been identified by existing databases.
Trained field coders assessed each physical activity resource on overall characteristics, the number, type and quality of features and amenities it possessed, and overall incivilities using the Physical Activity Resource Assessment (PARA) instrument (available from the principal investigator at ). Each physical activity resource was classified as a fitness club, a trail, a gym or sports facility, a park, a school, a church, or a community center. Resources that had multiple uses were coded based on the primary function of the resources. All resources were rated on hours of use, cost for use, and size (i.e., small, < ½ city block; medium, ½ city block < 1 city block; large, 1 city block or larger).
Data collectors counted and coded 25 unique possible elements of each physical activity resource that included 13 features used specifically for physical activity (e.g., basketball courts, soccer fields, playgrounds) and 12 amenities (e.g., benches, lighting, sidewalks). Each feature or amenity was also rated for quality by a three category quantitative system, which was developed based on extensive pilot testing of physical activity resources not in study neighborhoods. Ratings were listed as 3 "good," 2 "mediocre," and 1 "poor," with specific operational definitions developed by the research team for each item in each category. Definitions were constructed based on objective standards of quality. For example, an outdoor soccer field's rating of good was defined as "Field has uniform grass coverage and is well-mowed, no trash or debris on field; nets, if furnished are intact;" mediocre was defined as "Grass coverage may be sparse in a few places, grass may be too high, some trash or debris on field;" and poor was defined as "Grass coverage may be poor in 50% or > of the field, rough surface, hazards and/or trash on the field." Good sidewalks were defined as "Sidewalk is smooth, clear of debris," while mediocre sidewalks had "some debris, cracks or uneven surfaces, but [were] otherwise usable," and poor sidewalks had "major damage and need repair, almost unusable."
Each physical activity resource was also rated on overall incivilities. Incivilities included 9 elements that would reduce the pleasure associated with using that physical activity resource. These included auditory annoyances, broken glass, dog refuse, unattended dogs, evidence of alcohol and substance use, graffiti, litter, not enough grass or overgrown grass, sex paraphernalia, and vandalism [25]. Incivilities is a term that was originally coined by criminologists Wilson and Kelling [26] and has been investigated in sociological and anthropological contexts to describe the quality and social order of a neighborhood [27,28]. The presence of incivilities has been associated with less physical activity [25] and other poor health outcomes [10]. Incivilities were coded on a four category rating system of 4 "not present," 3 "a little," 2 "a medium amount," and 1 "a lot." Objective operational definitions were created and pilot tested for each item. For example, unattended dogs were defined as, good "1 dog unattended," mediocre "2–4 dogs unattended; may be associated noise," and poor "5 or > dogs unattended, definitely unsafe, may be associated noise." Litter was defined as, good "A few items (< 5) are on the ground," mediocre "Several items (5–10) are on the ground," and poor "Many items are on the ground (11+)."
Procedures
The physical activity resource assessment instrument was developed over a nine month period. The instrument was pilot tested and revised numerous times to achieve the final form. Reliability tests of a 10% overlap showed good reliability (rs > .77).
After neighborhoods were selected, the physical activity resource census was developed using the above described method. Trained field assessors (three doctoral candidates in psychology) used the instrument to systematically describe each physical activity resource. All assessments were conducted during daylight hours in the spring, summer and fall seasons when the ground was free from snow or ice. Assessors were accompanied by a second student for safety reasons in the housing development neighborhoods, and procedures included safety protocols in case of imminent perceived danger. Field assessments typically took about 10 minutes to complete; however, in a few cases of larger resources (e.g., a large park) the instrument could take up to 30 minutes to complete. Data were entered and proofed by trained graduate assistants. All analyses were conducted using SPSS [29].
Results
Neighborhood characteristics
Aggregate neighborhood characteristics are described in Table 1. Housing development neighborhoods had a median household income range of $11,930–$34,303, (M = $22,871, SD = $7,004), a population density range of 881–2,761, (M = 1,541, SD = 534), a non-white population range of 50.7%–98.1% (M = 71.4%, SD = 14.3%), and a street intersections range of 50–138 (M = 89.2, SD = 24.2). Comparison neighborhoods had a median household income range of $38,099–$48,383, (M = $42,364, SD = $4,493), a population density range of 1,079–1,824, (M = 1,508, SD = 328), an ethnic minority population range of 9.1%–18.2% (M = 13.2%, SD = 4.1%), and a street intersections range of 93–107 (M = 101, SD = 6.3).
Housing development neighborhoods had a range of 0 to 8 physical activity resources (M = 4.85, SD = 2.82), including fitness clubs, parks, sport facilities, community centers, churches, and schools, with considerable variability in the type of resources available for each neighborhood. As shown in Figure 1, one in three PA resources in HD neighborhoods were parks (n = 22, 35%), possibly reflecting preferences of early city developers. One fourth of PA resources were public school yards (n = 16, 25%) illustrating an important, and underrecognized, role that public schools play in communities. Most (n = 8, 62%) communities also had access to a community center. Comparison neighborhoods had a range of 2 to 9 PA resources (M = 6, SD = 3.56), including fitness clubs, parks, sport facilities, trails, community centers, churches, and schools. As shown in Figure 1, 38% (n = 9) of the resources in comparison neighborhoods were churches, but only one neighborhood (25%) had access to a community center. Most resources were freely accessible at no cost (82%), and appeared evenly distributed throughout neighborhoods.
Figure 1 Physical Activity Resources by Neighborhood
Table 2 presents the physical activity features and resource amenities and their quality by neighborhood. Thirteen possible physical activity features within each resource were assessed for availability and quality. HD neighborhoods had slightly more physical activity features within each resource (M = 2.71, SD = 1.65) than did resources in comparison neighborhoods (M = 2.17, SD = 1.63). Although not shown in the tables, it is interesting to note that in HD neighborhoods, fitness clubs and community centers had the most physical activity features available, on average, (M = 4, SD = 1.91); however, in comparison neighborhoods, parks had the most physical activity features available (M = 4.67, SD = 1.53). Sport facilities had the lowest average number of physical activity features within each resource available for both the HD neighborhoods (M = 1.25, SD = .50), and the comparison neighborhoods (which had none of the features on the assessment form). In HD neighborhoods, quality ratings for physical activity features within resources ranged from 1 to 3 (M = 2.45, SD = .81), and from 1 to 3 (M= 2.41, SD = .96) at comparison neighborhoods.
Table 2 Mean count and quality ratings for amenities and physical activity features by neighborhood type.
HD ID Count of Resources Mean Count of Features Mean Quality of Features Mean Count of Amenities Mean Quality of Amenities
H01 4 1.25 1.25 1.75 2.06
H02 0 0 0 0 0
H03 8 2.25 2.64 4.37 2.60
H04 5 4.20 2.25 4.40 2.54
H05 1 2.00 3.00 4.00 2.25
H06 5 2.60 2.45 2.80 2.68
H07 6 2.33 2.32 4.83 2.47
H08 8 2.63 2.50 3.50 2.27
H09 1 2.00 3.00 5.00 2.20
H10 5 4.00 2.89 4.40 2.31
H11 4 3.25 2.80 3.25 2.28
H12 8 3.00 2.60 3.88 2.36
H13 8 2.25 2.31 3.63 2.38
Total 63
Mean 2.71 2.45 3.79 2.40
(SD) (1.65) (.81) (2.16) (.68)
C01 2 3.00 2.75 2.00 2.00
C02 4 2.00 2.17 4.50 2.69
C03 9 1.22 1.89 1.00 1.72
C04 9 3.00 2.96 4.44 2.81
Total 24
Mean 2.17 2.41 2.96 2.32
(SD) (1.63) (.96) (2.42) (.97)
Twelve possible amenities were assessed for availability and quality at each resource. HD neighborhoods had more amenities per resource, on average (M = 3.79, SD = 2.16) than did comparison neighborhoods (M = 2.96. SD = 2.42). Although not presented in the table, on average, community centers had the most amenities available for both the HD neighborhoods (M = 5.5, SD = 1.98) and the comparison neighborhoods (N = 8, SD = 0). In HD neighborhoods, churches had the fewest amenities available (M = 1.50, SD = 1.31), although in comparison neighborhoods, fitness clubs had the fewest amenities available (M = 1.67, SD = 1.53). For amenities, quality ratings ranged from 1 to 3 (M = 2.40, SD = .68) in HD neighborhoods, and from 2 to 3 (M = 2.32, SD = .97) in comparison neighborhoods. However, quality ratings within each neighborhood varied widely.
Eighty percent of resources in all HD neighborhoods had incivilities (M = 1.81 per resource, SD = 1.72). In contrast, incivilities were found at only 11% of the PA resources in only half of the comparison neighborhoods (M = .29, SD = .75). This relationship was significant (t = 12.60, p < .001) as illustrated in Figure 2. Litter was the most frequently reported incivility for HD resources (65%, N = 41), with 20 resources having litter ratings of a medium amount to a lot. Broken glass was found at 25% of the resources (N = 16), with 14 ratings of a medium amount to a lot. Twelve resources (19%) had evidence of alcohol use, with 7 ratings of a medium amount to a lot. An auditory annoyance was reported at 16% (N = 10) of HD resources, with 8 of the resources having auditory annoyance ratings of a medium amount to a lot, mostly for traffic noise. Graffiti or tagging was found at 14% (N = 9) of the resources, with 6 having a medium amount to a lot. Nine of the resources (14%) lacked grass, while 10 (16%) of the resources had overgrown grass. Dog refuse (N = 1), unattended dogs (N = 2), evidence of substance use (N = 1), and vandalism (N = 1) were seen at less than two percent of the resources, and sex paraphernalia was not present at any. Comparison neighborhoods had few incivilities. Litter was present at 17% (N = 4) of the resources, with 2 having a medium amount to a lot. Graffiti, overgrown grass, and vandalism were all found at one (4%) of the resources, with ratings of a little bit to a medium amount.
Figure 2 Percent of PA Resources with Incivilities by Neighborhood
Discussion
This study (1) developed the PARA instrument that provides a brief, reliable and effective strategy to objectively assess neighborhood factors that may influence physical activity by describing the type, quantity, features, amenities and incivilities of physical activity resources, and (2) used the PARA to evaluate the physical activity resources in thirteen lower income, high ethnic concentration neighborhoods that surrounded public housing developments in comparison to physical activity resources in higher income, low ethnic minority concentration neighborhoods.
Lower income, higher ethnic concentration, housing development neighborhoods varied widely in the number and type of resources that were available for physical activity. All but one of the neighborhoods had access to parks, and most had accessible public school yards and community centers. The high number of parks found in housing development neighborhoods likely reflects city planners and landscape architects from over a century ago, such as Frederick Law Olmstead [30] who advocated for park development for the civilizing influence of "neighborly recreation" and its benefits to human health. As was posited then, more recent data suggest that people who live near attractive, public open spaces may be almost twice more likely to walk at moderately active levels than were those who do not have access to public open spaces [31]. Higher income, low ethnic concentration, comparison neighborhoods also varied widely in the number and type of physical activity resources available. The neighborhoods were selected to be similar in age and urban design; thus, the housing development neighborhoods and comparison neighborhoods might have similar amounts of parks, schoolyards and other public structures resulting from similar urban planning strategies. Although efforts were made to select appropriate comparison neighborhoods for the purposes of the current study, these neighborhoods do not represent the entire universe of possible neighborhoods.
Although the net number of physical activity resources did not vary by neighborhood income and ethnic concentration, the overall environment of physical activity resources was strikingly different in the neighborhoods, and suggests that evaluating merely the presence or absence of physical activity resources may be an overly simplistic way to investigate access to resources. In this sample, incivilities were consistently present and conspicuously bad and offensive at physical activity resources in lower income, higher ethnic concentration neighborhoods. In comparison, we found no incivilities at the resources in half of the comparison neighborhoods, and very few incivilities at the resources of the other half. It would seem to make sense that the presence of litter and debris, as well as lack of apparent maintenance might be important detractors from using a physical activity resource for physical activity, because people who walk more frequently typically rate their environment more positively [31]. In fact, a high proportion of incivilities might suggest lack of attention to an area, and might even encourage less desirable behavior (e.g., drug trafficking, prostitution) clearly not promoting favorable conditions for recreational physical activity [25].
Perhaps most distressing about these findings is the suggestion that the enduring relationship between social inequalities and poor health in an ostensibly wealth country, the US, is complex and appalling. The results suggest that merely building a park in a deprived area may be insufficient for insuring its intended use and maintenance. It is critical to provide ongoing support for maintenance and civic improvements. There is a deep need for policy makers and political leaders to work with communities to improve the quality of publicly available physical activity resources to improve the quality of life for all.
Conclusion
Previous work suggests that convenient neighborhood resources are critical for meeting physical activity recommendations [31,32]. An important contribution of this study is the inclusion of quality and incivility ratings in our instrument. To date, investigations that have examined physical activity resources have been limited to investigating the type and accessibility of resources [13,25]. This study is among the first to comprehensively evaluate a broad range of physical activity resources on a number of dimensions. At the same time, the PARA instrument was developed to be easily and rapidly administered, promoting broad scale dissemination and relatively, easily translatable outcomes for policy implementation. Future research is needed that tests the associations among these physical activity resource dimensions and individual physical activity levels to guide individual and environmental intervention development.
Authors' contributions
REL conceived of the study, and directed all aspects of the study including development, assessment and analyses, and lead the writing of the manuscript. KMB, JRS, GR all helped to develop measures, conduct assessment and analyses and assisted with the writing of the manuscript. HH provided geographic support including neighborhood maps, location of resources and assistance with variable construction, as well as assisting with the writing of the manuscript. All authors read and approved the final manuscript.
Acknowledgements
The authors would like to gratefully thank Dr. Paul A. Estabrooks for his thoughtful comments and feedback on this manuscript. This work made possible by a grant from the American Heart Association Heartland Affiliate to Dr. Rebecca E. Lee.
==== Refs
U.S. Department of Health and Human Services The Surgeon General's Call to Action to Prevent and Decrease Overweight and Obesity 2001 Rockville, MD , U.S. Department of Health and Human Services, Public Health Service, Office of the Surgeon General
USDHHS Physical activity and health: A report of the Surgeon General 1996 Atlanta , Centers for Disease Control and Prevention
Yen IH Kaplan GA Poverty area residence and changes in physical activity level: Evidence from the Alameda County Study Am J Public Health 1998 88 1709 1712 9807543
Ross CE Walking exercising and smoking: Does neighborhood matter? Soc Sci Med 2000 51 265 274 10832573 10.1016/S0277-9536(99)00451-7
Cubbin C Hadden WC Winkleby MA Neighborhood context and cardiovascular disease risk factors: the contribution of material deprivation Ethn Dis 2001 11 687 700 11763293
Ecob R Macintyre S Small area variations in health related behaviours; do these depend on the behaviour itself, its measurement, or on personal characteristics? Health Place 2000 6 261 274 11027952 10.1016/S1353-8292(00)00008-3
Ellaway A Macintyre S Does where you live predict health related behaviours?: A case study in Glasgow Health Bull (Edinb) 1996 54 443 446 8990608
Diez Roux AV Investigating neighborhood and area effects on health Am J Public Health 2001 91 1783 1789 11684601
Sampson RJ Raudenbush SW Earls F Neighborhoods and violent crime: A multilevel study of collective efficacy Science 1997 277 918 924 9252316 10.1126/science.277.5328.918
Ross CE Mirowsky J Neighborhood disadvantage disorder and health J Health Soc Behav 2001 42 258 276 11668773
Morland K Wing S Diez Roux A Poole C Neighborhood characteristics associated with the location of food stores and food service places Am J Prev Med 2002 22 23 29 11777675 10.1016/S0749-3797(01)00403-2
MacIntyre S Ellaway A Neighbourhood cohesion and health in socially contrasting neighbourhoods: implications for the social exclusion and public health agendas Health Bulletin 2000 58 450 4460 12813776
Estabrooks PA Lee RE Gyurcsik NC Resources for physical activity participation: Does availability and accessibility differ by neighborhood socioeconomic status? Ann Behav Med 2003 25 100 104 12704011 10.1207/S15324796ABM2502_05
Giles-Corti B Donovan RJ Socioeconomic status differences in recreational physical activity levels and real and perceived access to a supportive physical environment Prev Med 2002 35 601 611 12460528 10.1006/pmed.2002.1115
Giles-Corti B Broomhall MH Knuiman M Collins C Douglas K Ng K Lange A Donovan RJ Increasing walking: how important is distance to, attractiveness, and size of public open space? Am J Prev Med 2005 28 169 176 15694525 10.1016/j.amepre.2004.10.018
Atkinson JL Sallis JF Saelens BE Cain KL Black JB The association of neighborhood design and recreational environments with physical activity Am J Health Promot 2005 19 304 309 15768925
Saelens BE Sallis JF Black JB Chen D Neighborhood-based differences in physical activity: an environment scale evaluation Am J Public Health 2003 93 1552 1558 12948979
Handy SL Boarnet MG Ewing R Killingsworth RE How the built environment affects physical activity: views from urban planning Am J Prev Med 2002 23 64 73 12133739 10.1016/S0749-3797(02)00475-0
Sallis J Kraft K Linton LS How the environment shapes physical activity A transdisciplinary research agenda American Journal of Preventive Medicine 2002 22 208 11897466 10.1016/S0749-3797(01)00435-4
Lee RE Reese-Smith J Regan G Booth K Howard H Applying GIS Technology to Assess the Obesogenic Structure of Neighborhoods Surrounding Public Housing Developments Medicine and Science in Sports and Exercise 2003 35S
USDHHS OS Federal Register 2003 68 6456 6458
Census of Poulation and Housing, 2000: Summary Files 1 and 3
Geographic Data Technology, Inc.
ESRI
Brownson RC Baker EA Housemann RA Brennan LK Bacak SJ Environmental and policy determinants of physical activity in the United States Am J Public Health 2001 91 1995 2003 11726382
Wilson JQ Kelling GL Broken Window Atlantic Monthly 1982 211 29 38
Scahfer K Atzwanger K Wallner B Grammer K Human evolutionary aspects and urban dwelling features Coll Antropol 1999 23 369 378 10646213
Airey L "Nae as nice a scheme as it used to be": lay accounts of neighbourhood incivilities and well-being Health Place 2003 9 129 137 12753796 10.1016/S1353-8292(03)00013-3
SPSS SPSS 12.0 for Windows
Olmstead FL LeGates RT, Stout F The Uses of Parks The City Reader 1870 Second edition New York, NY , Routledge
Carnegie MA Bauman A Marshall AL Mohsin M Westley-Wise V Booth ML Perceptions of the physical environment stage of change for physical activity and walking among Australian adults Res Q Exerc Sport 2002 73 146 155 12092889
Powell KE Martin LM Chowdhury PP Places to walk: Convenience and regular physical activity. American Journal of Public Health 2003 93 1519 1521 12948973
|
16162285
|
PMC1262748
|
CC BY
|
2021-01-04 16:38:04
|
no
|
Int J Behav Nutr Phys Act. 2005 Sep 14; 2:13
|
utf-8
|
Int J Behav Nutr Phys Act
| 2,005 |
10.1186/1479-5868-2-13
|
oa_comm
|
==== Front
Int J Behav Nutr Phys ActThe International Journal of Behavioral Nutrition and Physical Activity1479-5868BioMed Central London 1479-5868-2-151621612410.1186/1479-5868-2-15ResearchChanges in accessibility and preferences predict children's future fruit and vegetable intake Bere Elling [email protected] Knut-Inge [email protected] Department of Nutrition, Faculty of Medicine, University of Oslo, Norway2 Department of Nutrition, Box 1046 Blindern, 0316 Oslo, Norway2005 10 10 2005 2 15 15 3 12 2004 10 10 2005 Copyright © 2005 Bere and Klepp; licensee BioMed Central Ltd.2005Bere and Klepp; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Most children eat fewer fruits and vegetables than recommended. To be able to design effective interventions, understanding the aetiology of the behaviour is important. Accessibility and preferences have shown to be strong correlates of fruit and vegetable intake in several cross-sectional studies. The aim of this study was to identify predictors of future fruit and vegetable intake and to explore longitudinal patterns of interactions between accessibility and preferences.
Methods
Data presented are based on baseline (September 2001) and follow-up (May/June 2002) surveys of 20 control schools in the Norwegian intervention study Fruits and Vegetables Make the Marks. A total of 816 pupils (77%) completed both baseline and follow-up questionnaires. The average age of the sample at baseline was 11.8 years. The research instrument assessing potential predictor variables was guided by Social Cognitive Theory, and included Accessibility at home, Accessibility at school, Modelling, Intention, Preferences, Self-Efficacy and Awareness of the 5-a-day recommendations. Multiple regression analyses were performed.
Results
All independent variables (measured at baseline) were significantly correlated to future fruit and vegetable intake (measured at follow-up). When reported fruit and vegetable intake at baseline (past intake) was included in this model, the effect of the other independent variables diminished. Together with past intake, the observed change in the independent variables from baseline to follow-up explained 43% of the variance in the reported intake at follow-up. Past intake remained the strongest predictor, but changes in accessibility at home and at school, as well as changes in preferences for fruits and vegetables, also explained significant amounts of the variance in fruit and vegetable intake at follow-up. In addition, baseline accessibility was found to moderate the relationship between change in preferences and change in intake.
Conclusion
Change in accessibility and preferences appear to be important predictors of future fruit and vegetable intake among school children. Interventions should focus on strategies to modify these factors.
==== Body
Background
Most children eat fewer fruits and vegetables than recommended. To be able to design effective interventions, in order to increase fruit and vegetable consumption, it is critical to understand the aetiology of the behaviour. An intervention should aim at changing the strongest determinants of the behaviour, in order to be successful [1].
Behavioural theories, like Social Cognitive Theory (SCT) [2], provide frameworks for understanding health behaviour and can guide the selection of potential determinants [3]. SCT is extensively used when children's fruit and vegetable intake is the behavioural outcome, and it served as the theoretical framework for four out of five multi-component intervention studies recently reviewed [4].
Several factors have been suggested as determinants of children's fruit and vegetable intake [5]. Among these factors, accessibility and preferences have been most strongly correlated to intake in several studies [6-8]. These studies have, however, been conducted using cross-sectional designs. This is a limitation as cross-sectional relationships could be due to a third antecedent, cannot state causality, and the relationships could be functionally different in longitudinal studies [1]. Longitudinal studies are therefore highly requested in order to prospectively investigate such relationships [1,6,8].
Adolescent fruit and vegetable intake declines with increasing age, but has shown to be stable with respect to the relative intake between individuals [9]. Lien and colleagues [10] reported that the only significant variable in a longitudinal study investigating the variance in fruit and vegetable intake (at age 21) was past intake (at age 15). Eight percent of the variance in fruit and vegetable intake at age 21 was explained by fruit and vegetable intake at age 15 for boys, and 26% (low SES) and 20% (high SES) for girls [10]. Changes in determinants must, however, explain the variance in future fruit and vegetable intake, in addition to the variance explained by past intake, as the change from past to future intake has to be explained. A change in intake is indeed the ultimate goal for an intervention.
The aim of the present study was to identify predictors of future fruit and vegetable intake, to assess whether these factors predicted future fruit and vegetable intake when controlling for past intake, and to assess whether changes in these factors over time were related to future intake and to the change in intake over time. In addition, a secondary aim was to explore longitudinal patterns of interactions between accessibility and preferences.
Methods
Sample and procedure
Data presented are based on the baseline (September 2001) and follow-up (May/June 2002) surveys of the 20 control schools in the intervention project Fruits and Vegetables Make the Marks (FVMM). These schools were randomly selected from two Norwegian counties, Hedmark and Telemark, and all 6 th and 7 th graders in each school were invited to participate. All schools were public schools, as are most schools in Norway. Informed consent was sought from the children and their parents prior to the study. Ethical approval and research clearance was obtained from The National Committees for Research Ethics in Norway and from The Norwegian Social Science Data Services.
A survey questionnaire was completed by the pupils in the classroom in the presence of a trained project worker. One school-lesson (45 minutes) was used to complete the questionnaire. Out of 1065 eligible pupils, 896 completed the baseline questionnaire. A total of 816 (77%) also completed the follow-up questionnaire: 406 boys and 410 girls (444 6 th graders and 372 7 th graders). The average age of the sample at baseline was 11.8 years.
Instrument
A questionnaire to measure the children's fruit and vegetable intake and potential predictors of intake was developed as part of the FVMM project. Repeated pre-testing, a test-retest study [11,12] and a validation study [11] of the questionnaire were conducted prior to the baseline survey.
The questionnaire items assessing potential predictors were guided by SCT. SCT postulates that behaviour (here fruit and vegetable intake) is a result of environmental and personal factors, but it also states that all three sets of factors affect each other in constant reciprocal relationships [2]. The following factors were measured:
Fruit and vegetable Intake was measured by four frequency questions: 'How often do you eat vegetables for dinner?,' 'How often do you eat other vegetables (e.g., carrot for school lunch)?' 'How often do you eat apple, orange, pear or banana?' and 'How often do you eat other fruits or berries?' All four questions had ten response alternatives ranging from 'Never' = 0 to 'Several times a day' = 10. These items were added together, and the test retest (14 days in-between) correlation of this scale in a sample of 114 6 th grade pupils was 0.75 [11]. The correlation between this scale and a validation method (7-day food diaries) was 0.32 in a separate validation study of 85 6 th grade pupils, which is similar to the results found in other studies of this age group [11].
A total of seven potentially mediating factors were measured; three environmental and four personal. All questionnaire items included in these seven scales are provided in Table 1, including response alternatives, scale ranges and psychometric properties (test retest correlation and internal consistency reliability [6,11,12]. The environmental factors were Accessibility at home, Accessibility at school and Modelling. The personal factors were Intention (to eat 5-a-day), Preferences, Self-Efficacy (to eat 5-a-day) and Awareness (of the 5-a-day recommendation). All scales except Awareness and Accessibility at school included one to five statements with response alternatives ranging from 'I fully disagree' to 'I fully agree.' Awareness (of the 5-a-day recommendation) was measured by one question: 'How many servings of fruit and vegetables should a person your age eat every day?' This question had seven response alternatives ranging from 'None' to 'More than 5 a day.' Accessibility at school was a dichotomous variable assessing whether the pupils subscribed to the Norwegian School Fruit Programme or not. As few Norwegian elementary schools have canteens, the only accessible fruits and vegetables in most Norwegian schools are through this programme. This programme offers pupils a piece of fruit or a carrot every day at the cost of NOK 2.50 per day (€ 0.30) [13]. All Norwegian elementary schools are offered the chance to participate in this programme. If the school participates, fruit and vegetables are available to the pupils, but it is not accessible to them unless they subscribe to the programme. As the Norwegian School Fruit Programme started in October 2001 in Hedmark and Telemark, no pupils subscribed at baseline (September 2001). Thus, the baseline score for all pupils was zero.
Table 1 Questionnaire items, response alternatives, and reliability coefficients (test-retest correlation and internal consistency reliability) of fruit and vegetable intake and the SCT variables assessed in the FVMM study.
Scale Response Possible scale range Test-retest correlation (14 days in between)** Internal consistency reliability***
Intake (times/week) 0/40 0.75 NA
How often do you eat:
1. Vegetables for dinner?
2. Vegetables on bread?*
3. Other vegetables (e.g. carrot for school lunch)?
4. Apple, orange, pear or banana?
5. Other fruits or berries? Ten alternatives: Never (0), less than once a week (0.5), once a week (1) to every day (7), several times a day (10). Question 3 did not count in the scale.*
Accessibility at home -10/10 0.66 0.49
1. At home we usually always have fruits and vegetables in the refrigerator
2. At home I am allowed to eat fruits and vegetables whenever I want
3. Mother or father do sometimes cut up fruits or vegetables for me as a snack
4. At home we usually have vegetables at dinner every day
5. At home we usually have fruits available in a (fruit-) bowl Five alternatives: I fully disagree (-2), I disagree (-1), I do not agree nor disagree (0), I agree (1), I fully agree (2)
Accessibility at school 0/1 No data NA
1. Do you subscribe to the School Fruit Programme? Yes (1), no (0).
Modelling -8/8 0.70 0.46
1. My mother eats lots of fruits and vegetables
2. My father eats lots of fruits and vegetables
3. Many of my friends and siblings eat lots of fruits and vegetables
4. My home-economy teacher eats lots of fruits and vegetables Five alternatives: I fully disagree (-2) to I fully agree (2)
Intention (to eat 5-a-day) -2/2 0.51 NA
1. I intend to eat at least 5 servings of fruits and vegetables every day Five alternatives: I fully disagree (-2) to I fully agree (2)
Preferences -8/8 0.74 0.68
1. Fruits and vegetables make my meals taste better
2. I really like raw vegetables
3. Fruits are among the best [foods] I know
4. Fruits and vegetables are very suitable as snacks Five alternatives: I fully disagree (-2) to I fully agree (2)
Self-Efficacy (to eat 5-a-day) -6/6 0.61 0.44
1. For me, it would be easy to eat fruits or vegetables to every meal, every day, if I decided to do so
2. For, me it would be easy to eat fruits or vegetables on Saturday evenings, even if everybody else are eating snacks
3. For me, it would be easy to eat more than 5 servings of fruits and vegetables every day Five alternatives: I fully disagree (-2) to I fully agree (2)
Awareness (of 5-a-day) 0/6 No data NA
1. How many servings of fruit and vegetables should a person at your age eat every day? Seven alternatives: None (0), 1-a-day (1) to 5-a-day(5), more than 5-a-day (6)
* Some Norwegians have vegetables on their sandwiches, but usually in small amounts. Therefore, this question was not included in the intake scale. The question was included in the questionnaire so that the participants should not include their vegetables on bread in the 'other vegetables' question.
** Intake: Spearman's r From Andersen and colleagues [11], all other scales: Pearson's r from Bere and Klepp [12].
*** Cronbach's alpha: From Bere and colleagues [6].
NA = not applicable
Statistics
Missing values on any item were substituted with the mean value for the remaining group on the respective item, if more than 50% of the scale items were answered. A total of 214 pupils had one or more missing values substituted. Multiple regression assumptions regarding normality, linearity and homoscedasticity were found to be acceptable, and therefore parametric statistics were used. Multiple regressions were performed to determine the explained variance of the children's fruit and vegetable intake and of the change in intake. Pearson's correlation coefficients (r) and standardized regression coefficients (beta) are given for each independent variable. In addition, the unique amount of variance in intake explained by an independent variable is given by the square of the semi-partial correlation (sri2) [14,15]. The square of the multiple correlation (= explained variance) is given by the multiple correlation (R2) and the adjusted multiple correlation (adj. R2).
The effect of potential interactions between baseline values of Preferences (dicotomised) and the change in Accessibility at home (positive or negative (including 0)), and baseline values of Accessibility at home (dicotomised) and change in Preferences (positive (including 0) or negative) and change in fruit and vegetable intake was assessed by including the respective cross-product terms into linear regression models. These models did also include the dicotomised change in Accessibility or the dicotomised change in Preferences respectively.
Paired sample t-tests were used in the attrition analyses. All analyses were conducted using SPSS version 12.
Results
Mean values of intake and the SCT constructs at baseline and follow-up, as well as change scores are presented in Table 2.
Table 2 Baseline, follow-up and change mean scores and standard deviations (SD) of the variables assessed.
Baseline Follow-up Change
Scale n Mean SD n Mean SD n Mean SD
Intake (times/week) 804 14.1 7.1 810 13.2 7.1 799 -0.9 6.6
Accessibility at home 815 3.8 3.6 813 4.2 3.5 812 0.5 3.6
Accessibility at school 816 0 NA 816 0.1 NA 816 0.1 NA
Modelling 796 2.0 2.7 791 2.0 2.7 774 -0.1 2.9
Intention (to eat 5-a-day) 809 0.2 1.3 812 0.1 1.3 805 -0.1 1.4
Preferences 810 2.7 3.8 813 2.1 3.9 807 -0.6 3.3
Self-Efficacy (to eat 5-a-day) 813 0.1 2.6 814 0.2 2.7 811 0.2 2.7
Awareness (of 5-a-day) 805 3.5 1.5 792 3.4 1.6 782 -0.1 1.8
NA = not applicable
Correlates of intake cross-sectionally
Cross-sectionally, Accessibility at home and Preferences were most strongly correlated to intake (r = 0.43 and 0.45, respectively) (Table 3). At baseline, the independent variables explained 29% (adj. R2) of the variance in intake at baseline (Table 4, analysis I).
Table 3 Correlation (Pearson's r) between the SCT variables (baseline and change scores) and fruit and vegetable intake (baseline, follow-up and change scores).
Fruit and vegetable intake
Baseline Follow-up Change
Baseline:
Intake (= past intake) 1** 0.57** NA
Accessibility at home 0.43** 0.31** NA
Modelling 0.24** 0.18** NA
Intention (to eat 5-a-day) 0.33** 0.27** NA
Preferences 0.45** 0.35** NA
Self-Efficacy (to eat 5-a-day) 0.35** 0.30** NA
Awareness 0.22** 0.19** NA
Change in:
Accessibility at home NA 0.13** 0.27**
Accessibility at school NA NA NA
Modelling NA 0.10** 0.11**
Intention (to eat 5-a-day) NA 0.07* 0.14**
Preferences NA 0.17** 0.28**
Self-Efficacy (to eat 5-a-day) NA 0.12** 0.16**
Awareness NA 0.11** 0.10**
NA = not applicable
* p < 0.05
** p < 0.01
Table 4 Multiple regressions of fruit and vegetable intake (baseline, follow-up and change) by the SCT variables (baseline and change) including the standardized regression coefficients (beta) and the semi-partial correlation (sri2).
Fruit and vegetable intake
Baseline Analysis I (n = 766) Analysis IIa (n = 770) Follow-up Analysis IIb (n = 762) Analysis IIc (n = 722) Change Analysis III (n = 722)
beta sri2 beta sri2 beta sri2 beta sri2 beta sri2
Baseline
Accessibility at home 0.24** 0.04 0.15** 0.02 0.04 0.00
Modelling 0.06 0.00 0.03 0.00 0.01 0.00
Intention (to eat 5-a-day) 0.03 0.00 0.04 0.00 0.03 0.00
Preferences 0.26** 0.04 0.18** 0.02 0.06 0.00
Self-Efficacy (to eat 5-a-day) 0.08* 0.00 0.11** 0.01 0.07 0.00
Awareness 0.09** 0.01 0.09* 0.01 0.05 0.00
Past intake 0.47** 0.15 0.59** 0.34
Change in
Accessibility at home 0.14** 0.02 0.21** 0.04
Accessibility at school 0.17** 0.03 0.16** 0.03
Modelling 0.00 0.00 -0.02 0.00
Intention (to eat 5-a-day) -0.03 0.00 -0.03 0.00
Preferences 0.17** 0.02 0.21** 0.03
Self-Efficacy (to eat 5-a-day) 0.05 0.00 0.05 0.00
Awareness 0.08** 0.01 0.07* 0.00
R2: 0.30 0.18 0.33 0.44 0.16
Adj. R2: 0.29 0.17 0.33 0.43 0.15
Sum sri2: 0.10 0.05 0.16 0.42 0.11
* p < 0.05
** p < 0.01
Prediction of future intake
All independent variables (measured at baseline) were significantly correlated to future intake (measured at follow-up) (Table 3). These variables explained 17% of the variance in the pupils' fruit and vegetable intake at follow-up, with Modelling and Intention as the only non-significant variables (Table 4, analysis IIa). Overall, 5% (sum sri2) of the variance was explained by unique contribution to the explanation, while the remaining 12% was shared variance by two or more concepts. Accessibility at home and Preferences contributed most of the unique variance explained (explaining 2% each).
When reported fruit and vegetable intake at baseline (past intake) was included in the model, none of the other baseline variables remained significant (Table 4, analysis IIb). This model explained 33% of future fruit and vegetable intake. A model with past intake as the only independent variable explained 32% of the variance in future intake (data not shown).
In addition to past intake, the change in the independent variables explained 43% of the variance of follow-up intake, almost all by unique contribution by; past intake (35%), change in Accessibility at home (2%), change in Accessibility at school (3%), change in Preferences (2%) and change in Awareness (1%) (Table 4, analysis IIc).
Correlates of change in intake
The changes in the independent variables were all significantly correlated to the change in intake (Table 3), and they explained 15% of the variance in the change in intake between baseline and follow-up, with Accessibility at home, Accessibility at school, Preferences and Awareness being significant (Table 4, analysis III). Overall, 11% of the variance was explained by unique contribution to the explanation. Accessibility at home, Accessibility at school and Preferences contributed most of the unique variance explained (4%, 3% and 3%, respectively).
Interaction analyses
The cross-product of baseline Preferences and change in Accessibility at home was not significant when introduced in the model (p = 0.29). The cross-product of baseline Accessibility and change in Preferences was significant (p = 0.03), and therefore the relationship between change in Preferences and change in intake are presented in Table 5, stratified by baseline Accessibility at home. Table 5 shows that the difference in change in intake between those with positive and negative changes in Preferences was much greater among those with a high baseline Accessibility at home than those with a low accessibility (4.4 vs. 2.3 times/week).
Table 5 Changes in fruit and vegetable intake (times/week) related to changes in Preferences, stratified by baseline Accessibility at home
Baseline Accessibility at home Change in Preferences Change in intake Confidence intervals
LOW (n = 424) Negative (n = 188) -1.7 (-2.6, -0.8)
Positive (n = 230) 0.6 (-0.2, 1.4)
p-value < 0.01
HIGH (n = 391) Negative (n = 216) -3.4 (-4.2, -2.5)
Positive (n = 173) 1.0 (0.0, 1.9)
p-value < 0.01
Attrition analyses
No significant differences were seen between the cohort participants (n = 816) and the baseline-only participants (n = 80) for any of the variables assessed in this study. The pupils with scores on all scales assessed (n = 722, same sample as analyses IIc and III) had higher Preferences than those without follow-up data or without scores on all scales (n = 174, p = 0.05). Of the pupils with scores on all scales, those without missing data (n = 508) did not show different scores on any of the scales compared to pupils with one or more missing values replaced (N = 214).
Discussion
These results from the FVMM project show that changes in Accessibility (at home and at school) and Preferences were correlated to changes in intake, and that these changes explained some of the variance of follow-up fruit and vegetable intake, when controlling for past intake. This suggests that these factors play a role as potential mediators in future intervention studies.
Prospectively, the change in SCT factors explained 15% of the change in intake, and together with past intake, 43% of the variance in future intake. We are not aware of any other studies assessing the prospective nature of fruit and vegetable predictors in children, and this is more explained variance than what has been reported for adults [16]. The present study contributes to the literature by showing that longitudinal relationships exist between accessibility, preferences and fruit and vegetable intake. Longitudinal relationships are necessary, but however, not a sufficient premise for causality.
While it is a prerequisite that fruits and vegetables are available and accessible, it is not necessarily sufficient to ensure high intake. A recent review of environmental interventions to promote fruit and vegetable consumption among youth in school settings reported only three stand-alone environmental interventions [4]. Only one of them was a stand-alone availability study, assessing the effect of a Danish pilot project of a school fruit and vegetable subscription programme [17]. The subscription programme increased 6–10 year old children's intake of fruit among both subscribing and non-subscribing pupils at the intervention schools, with about 0.4 pieces/school day, compared to control schools. The two other stand-alone environmental studies reported significant effects of lowering prices of fruit and vegetable on pupil purchases [18], and of a multi-environmental strategy [19]. In addition, a number of multi-component fruit and vegetable and multi-behavioural (including fruit and vegetables) interventions were included in this review [4]. These studies did not separately evaluate the availability/accessibility component, and unfortunately, the effect of that component cannot be stated. More recently, we have evaluated the effect of free participation in the Norwegian School Fruit Programme [13]. Seventh-graders at nine elementary schools were given a piece of fruit or a carrot every school day for a school year for free, and the pupils' mean intake of fruit and vegetables at school increased by about 0.9 portions compared to control pupils [13]. Offering free fruit at school can be seen as increasing the accessibility of fruit and vegetables at school, and this increased accessibility clearly led to increased intake. Increasing accessibility is theoretically simple; just offer children fruit and vegetables – at school or at home.
Food preferences have been suggested as determinants for food intake [20], including fruit and vegetable intake [6,8]. Previous research suggests that children's dislike of foods can be transformed into liking of foods with repeated tasting or 'exposure' to those foods [21,22]. It has also been reported that children's food preferences are often guided by taste alone, while food choices of adults also tend to be influenced by nutritional beliefs and attitudes toward weight and dieting [20]. However, a few studies have reported that children's preferences for and consumption of disliked vegetables were enhanced when children had opportunities to observe peers selecting and eating those foods, and that adults can also be effective in increasing fruit and vegetable intake by encouraging children to try new foods [23]. We are, however, not aware of any intervention studies that have increased children's or adolescents' fruit and vegetable intake through increased preferences.
In the present study we also found an interaction between baseline accessibility at home and the relationship between change in preferences and change in intake, indicating that baseline accessibility mediate this relationship. For those with high baseline accessibility, changes in preferences were related to significantly larger changes in intake than for those with low baseline accessibility, indicating again that high access to fruits and vegetables are extremely important for a sufficient fruit and vegetable intake. This result is in line with previously cross-sectionally reported interactions between accessibility and preferences. Neumark-Stainer et al. [8] found that, in a group of adolescents (mean age 14.9 years), preferences was more related to intake for those with higher levels of accessibility. For those with the lowest accessibility, preferences were not related to intake. Similarly, Cullen and colleagues [7] found in a group of 4–6 graders that among those with low preferences, both availability and accessibility were significant in explaining the variance in fruit and vegetable intake. For those with high preferences, only availability was significant. This again indicates that those with lower preferences need a higher access to fruit and vegetables in order to eat sufficient amounts of fruit and vegetables.
In addition to changes in Accessibility and Preferences, change in Awareness of the 5-a-day recommendations contributed significantly to the explanation of the variance in future fruit and vegetable intake. A change in Awareness also contributed significantly to the explanation of variance in the change in intake. Recently, Reynolds and colleagues [24] showed that a similar scale was a significant mediator in the High 5 Alabama intervention study (an increase in Awareness explained 9.8% of the increase observed in fruit and vegetable intake). It has also been reported from several countries that several people are not aware of national fruit and vegetable recommendations [25-28]. Thus, our results are encouraging, and relevant information about existing 5-a-day recommendations should be included in future intervention studies.
The strength of this study is that it includes a prospective cohort of a rather large random sample of schools. There are, however, also some limitations with the present study. The study was geographically confined to two of Norway's 19 counties. As Norway is a rather homogeneous country, we believe the results are likely to be generalizable to the other counties. A second limitation is the validity of the intake measure as this scale showed a rather low correlation with the validation method [11]. However, the correlation was not lower than found in other studies of same age pupils, and the scale showed good test-retest reliability. A third limitation is that the follow-up period was only 8–9 months. In such a short time span, large changes in fruit and vegetable intake can not be expected. A small change in intake will be a limiting factor for observing relations between change in intake and its determinants. However, due to an age-related decline in fruit and vegetable intake previously observed in Norway [9] and elsewhere in Europe [29], as well as seasonal variations in Norway, the average change in intake was -0.9 times/week (Table 2). Finally, when using observational data, prospective relationships can, as for cross-sectional studies, be due to a third antecedent. Thus we can still not state causality.
Conclusion
Changes in Accessibility and Preferences and Awareness were significantly correlated to changes in reported fruit and vegetable intake, and as hypothesised, these changes also explained added variance in future fruit and vegetable intake when adjusting for past intake. Baseline accessibility was a moderator of the relationship between change in preferences and change in intake. These results point to the potential role of these factors, especially accessibility, as mediators in future fruit and vegetable interventions.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
EB collected and analysed the data and drafted the manuscript. KIK conceived the study, participated in its design and coordination, and provided critical revision of the paper. Both authors have read and approved the final manuscript.
Acknowledgements
This study was funded by the Norwegian Research Council.
==== Refs
Baranowski T Lin LS Wetter DW Resnicow K Hearn MD Theory as mediating variables: why aren't community interventions working as desired? Ann Epidemiol 1997 7 S89 S95
Bandura A Social foundations for thought and action: A social cognitive theory 1986 Inglewood Cliffs, NJ: Prentice Hall
MacKinnon DP Analysis of mediating variables in prevention and intervention research NIDA Res Monogr 1994 139 127 153 8742555
French SA Staples G Environmental interventions to promote vegetable and fruit consumption among youth in school settings Prev Med 2003 37 393 610 10.1016/j.ypmed.2003.09.007
Baranowski T Cullen KW Baranowski J Psychosocial correlates of dietary intake: Advancing dietary intervention Annu Rev Nutr 1999 19 17 40 10448515 10.1146/annurev.nutr.19.1.17
Bere E Klepp K-I Correlates of fruit and vegetable intake among young adolescents – parental and self-reports Public Health Nutr 2004 7 991 999 15548337 10.1079/PHN2004619
Cullen KW Baranowski T Owens E Marsh T Rittenberry L de Moor C Availability, accessibility, and preferences for fruit, 100% fruit juice, and vegetables influence children's dietary behavior Health Educ Behav 2003 30 615 626 14582601 10.1177/1090198103257254
Neumark-Sztainer D Wall M Perry C Story M Correlates of fruit and vegetable intake among adolescents. Findings from Project EAT Prev Med 2003 37 198 208 12914825 10.1016/S0091-7435(03)00114-2
Lien N Lytle LA Klepp K-I Stability in consumption of fruit, vegetables, and sugary foods in a cohort from age 14 to age 21 Prev Med 2001 33 217 226 11522162 10.1006/pmed.2001.0874
Lien N Jacobs DR JrKlepp K-I Exploring predictors of eating behaviour among adolescents by gender and socio-economic status Public Health Nutr 2002 5 671 681 12372162 10.1079/PHN2002334
Andersen LF Bere E Kolbjørnsen N Klepp K-I Validity and reproducibility of self-reported intake of fruit and vegetable among 6 th graders Eur J Clin Nutr 2004 58 771 777 15116080 10.1038/sj.ejcn.1601875
Bere E Klepp K-I Reliability of parental and self-reported determinants of fruit and vegetable intake among 6 th graders Public Health Nutr 2004 7 353 356 15003144 10.1079/PHN2003529
Bere E Veierød MB Klepp K-I The Norwegian School Fruit Programme: evaluating paid vs. no-cost subscriptions Prev Med 2005 41 463 470 15917042 10.1016/j.ypmed.2004.11.024
Hankins M French D Horne R Statistical guidelines for studies of the theory of reasoned action and the theory of planned behaviour Psychol Health 2000 15 151 161
Tabachnick BG Fidell LS Tabachnick BG, Fiedell LS Multiple Regression Using Multivariate Statistics 2001 4 Needham Heights, MA: Allyn & Bacon 110 130
Kvaavik E Lien N Tell GS Klepp K-I Psychosocial predictors of eating habits among adults in their mid-30s: The Oslo Youth Study follow-up 1991–1999 IJBNPA 16076386
Eriksen K Haraldsdottir J Pederson R Flyger HV Effect of a fruit and vegetable subscription in Danish schools Public Health Nutr 2002 6 57 63 12581466 10.1079/PHN2002356
French SA Story M Jeffery RW Snyder P Eisenberg M Sidebottom A Murray D Pricing strategy to promote fruit and vegetable purchase in high school cafeterias J Am Diet Assoc 1997 97 1008 1010 9284880 10.1016/S0002-8223(97)00242-3
Perry CL Bishop DB Taylor GL Davis M Story M Gray C A randomized school trial of environmental strategies to encourage fruit and vegetable consumption among children Health Educ Behav 2004 31 65 76 14768658 10.1177/1090198103255530
Drewnowski A Taste preferences and food intake Annu Rev Nutr 1997 17 237 253 9240927 10.1146/annurev.nutr.17.1.237
Birch LL Fisher JO Development of eating behaviors among children and adolescents Pediatrics 1998 101 539 549 12224660
Wardle J Cooke LJ Gibson EL Sapochnik M Sheiham A Lawson M Increasing children's acceptance of vegetables; a randomized trial of parent-led exposure Appetite 2003 40 155 162 12781165 10.1016/S0195-6663(02)00135-6
Birch LL Development of food preferences Annu Rev Nutr 1999 19 41 62 10448516 10.1146/annurev.nutr.19.1.41
Reynolds KD Bishop DB Chou C-P Xie B Nebeling L Perry CL Contrasting mediating variables in two 5-a-day nutrition intervention programs Prev Med 2004 39 882 893 15475020 10.1016/j.ypmed.2004.03.022
Krebs-Smith SM Cook A Subar AF Cleveland L Friday J US adults' fruit and vegetable intakes, 1989 to 1991: a revised baseline for the Healthy People 2000 objective Am J Public Health 1995 85 1623 1629 7503335
Lechner L Brug J DeVries H Misconceptions of fruit and vegetable consumption: Differences between objective and subjective estimation of intake J Nutr Educ 1997 29 313 320
Parmenter K Waller J Wardle J Demographic variation in nutrition knowledge in England Health Educ Res 2000 15 163 174 10751375 10.1093/her/15.2.163
Bere E Increasing school-children's intake of fruit and vegetables: Fruits and Vegetables Make the Marks Phd thesis 2004 University of Oslo: Department of Nutrition
World Health Organization Health behaviour on school-aged children: international report from the 2001–2002 survey 2004 World Health Organization: Copenhagen
|
16216124
|
PMC1262749
|
CC BY
|
2021-01-04 16:38:04
|
no
|
Int J Behav Nutr Phys Act. 2005 Oct 10; 2:15
|
utf-8
|
Int J Behav Nutr Phys Act
| 2,005 |
10.1186/1479-5868-2-15
|
oa_comm
|
==== Front
Int Semin Surg OncolInternational seminars in surgical oncology : ISSO1477-7800BioMed Central London 1477-7800-2-201620216010.1186/1477-7800-2-20ResearchHepatocellular Carcinoma in The Gambia and the role of Hepatitis B and Hepatitis C Mboto Clement Ibi [email protected] Angela [email protected] Mark [email protected] Andrew Paul [email protected] Royal Victoria Hospital, Banjul, The Gambia2 School of Life Science, Kingston University, Surrey KT1 2EE, UK3 Faculty of Health and Social Care Sciences, Kingston University and St George's University of London, Surrey KT1 2EE, UK2005 4 10 2005 2 20 20 3 7 2005 4 10 2005 Copyright © 2005 Mboto et al; licensee BioMed Central Ltd.2005Mboto et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Objectives
Hepatocellular Carcinoma is the commonest form of cancer in The Gambia, and although Hepatitis B and Hepatitis C are known risk factors, accurate baseline data on Hepatitis B and Hepatitis C distribution in the region are limited. Similarly data including information on the involvement of the viruses in HCC remains unknown. The current study was undertaken to estimate the risk of HCC in relation to HCV and HBV in The Gambia.
Methods
Thirteen patients with histological proven history of HCC and 39 healthy controls were enrolled in the study. Each subject blood was screened individually for anti-HCV using ORTHO HCV 3.0 ELISA test system (Ortho-Clinical Diagnostics, Inc, U.S.A) and for HBsAg using QUADRATECH CHECK 4-HBs one step generation hepatitis B surface antigen test kit (VEDALAB, France) following the manufacturers instructions.
Results
HBsAg and anti-HCV was detected in 38.5 %(5/13) and 7.7% (1/39) of the persons with a history of HCC respectively. HBsAg but not anti-HCV was detected in 12.8% (5/39 of the case control subjects. HBsAg and HCV rates among the HCC patients were higher in men than women. Rates were highest in patients 48 years and above (37.5%; 3/8). No patient was found with anti-HCV and anti-HBV.
Conclusion
These results indicate that the involvement of HBV and HCV in HCC in the country is in a ratio of 5:1 and that these two viruses might be independently involved in the pathogenesis of the disease. The study revealed a statistically significant association (p = 0.04) between HBsAg and HCC patients.
The results also indicate that up to 50% of HCC cases in the country may be due to non viral factors and calls for further studies in this regard. These findings call for provision of diagnostic facilities for these viruses in hospitals and for their routine screening in blood banks while intervention programmes should be put in place.
Hepatitis Bhepatitis Chepatocellular carcinomaThe Gambia
==== Body
Introduction
Hepatocellular carcinoma (HCC) is a common cancer worldwide, which occurs substantially as a complication of liver cirrhosis [1]. Chronic infection with Hepatitis B virus (HBV) and Hepatitis C virus (HCV) has been associated with the disease [2,3] with higher incidences reported in countries where Hepatitis B and or Hepatitis C are endemic [4,5] HCV has a lower global prevalence than HBV and is more often associated with HCC in economically developed regions [6,7]. Globally, HCC is increasingly becoming a major health concern with estimates of 500,000 new cases reported annually [8]. Some studies have shown a direct correlation between the geographical distribution of HBV and HCV and HCC prevalence [7]. In Japan and Italy where higher HCV prevalence is very high, HCC has been shown to be more prevalent [8]. In the United States of America, it is estimated that the disease burden from HCV is likely to rise considerably over the next 10–20 years increasing demands on liver transplantation [9]. In West Africa, some studies have shown that both HBV and HCV infections are common but the role of HCV in acute infection is still not clear [10-13].
In the Gambia hepatocellular carcinoma (HCC) has been defined as the country's commonest form of cancer [14], and Hepatitis B virus (HBV) infection is endemic [15,16]. However, like other developing countries in the West African region, accurate data including information about incidence and prevalence of both HBV and HCV or the involvement of the viruses in HCC is lacking or limited [13]. The problem is further compounded by the non-existence of facilities for HCV or HBV diagnosis in established hospitals in the country making it difficult to have a baseline data on HBV or HCV distribution in the country. The few available studies however, have shown Hepatitis B prevalence in the Gambia to be quite high [15,16]. The Gambia National blood transfusion services (GAMBLOOD), which were recently put in place, are yet to commence screening of blood for HCV or HBV. The present study was carried out to compare the involvement of HBV and HCV in HCC in the country.
Subjects, Materials and Methods
This study is part of an on-going study on HIV and HCV coinfection and was approved by the Department of State for Health. The study population consisted of a total of 13 HCC patients seen consecutively at the Royal Victoria Teaching Hospital (RVTH), Banjul between the months of July to December 2002. The patients were aged 32 years to 76 years and were made up of 11 men and 2 women. Patients were enrolled for the study following informed consent. Each patient was matched by three persons on the basis of age and sex. The primary choice of control group persons were blood donors, however due to the lack of female blood donors, the two female HCC patients were matched with two women attending antenatal clinic in their first trimester of pregnancy, and four other female patients with history of malaria. In all a total of thirty-nine healthy controls made up of 33 blood donors and 6 women were enrolled for the study. Both the HCC patients and control subjects were unaware of their HCV or HBV status prior to the commencement of the study.
Blood samples were collected from each participant and linked by name and code number. Samples were separated within 8 hours of collection, screened individually for anti-HCV using ORTHO HCV 3.0 ELISA test system (Ortho-Clinical Diagnostics, Inc, U.S.A) a third generation enzyme linked immunosorbent assay (ELISA). Persons reactive to ORTHO HCV 3.0 ELISA test were considered anti-HCV positive [17]. Hepatitis B surface antigen (HBsAg) test was carried using QUADRATECH CHECK 4-HBs one-step generation hepatitis B surface antigen test kit (VEDALAB, France) following the manufacturers instructions.
HBsAg and anti-HCV prevalence rates were calculated to reflect the relative frequency of each disease while Odds ratio (OR) and ninety five percent confidence interval (95% CI) was calculated using the Fisher Exact Test to estimate the strength of the association between each infection and possible risk factor[18].
Results
The mean age of the HCC patients was 46 years and 43 years for the men and women respectively. The mean age of the control subjects was 45.7 years for the men and 43 years for the women. The mean age of the HCC patients with HbsAg was 47.8 years, while the only patient with anti-HCV was aged 54 years.
Hepatitis B surface antigen (HbsAg) was present in 38.5 %(5/13) of the HCC patients (p = 0.04; 95% CI: 1.03–8.73) and in 12.8% (5/39) of the control subjects (p = 0.046; 95% CI: 0.11–0.97). Anti-HCV antibodies were detected in 7.7 % (1/13) of the HCC patients. No anti-HCV antibody was detected among any of the control subjects. Similarly no HBsAg or nor anti-HCV was detected in more than half (53%; 7/13) of the HCC patients.
The male HCC patients had an anti-HCV prevalence of 9.1 %(1/11) and a hepatitis B surface antigen (HBsAg) prevalence of 36.4 %(5/11). The two female HCC patients who participated in this study were both anti-HCV and HBsAg negative. No patient was found with anti-HCV and anti-HBV.
The hepatitis B surface antigen (HBsAg) prevalence for the control group was 12.12% (4/33) for the males (95% CI: 0.097–5.43; OR 0.69). The female control subjects had an HBsAg prevalence of 16.7 % (1/6) (CI: 0.18–10.27; OR: 1.45). All the HBsAg positive control subjects were blood donors give an HBsAg prevalence of 15.2 % (5/33). None of the control subjects demonstrated antibody to HCV.
Hepatitis B surface antigen rates were highest in patients 48 years and above (37.5%, 3/8) (CI: 0.23–3.79; OR: 0.9). Control subjects aged 48 years and above had a lower HbsAg prevalence (11.1%, 2/18) than those less than 48 years (15.8 %, 3/19)}; however this was not statistically significant (p > 0.05). There was a marginal statistically significant association between HBsAg and HCC patients (p = 0.04; 95%CI: 1.03–8.73, OR: 4.25).
A summary of the results of the hepatitis B surface antigen and anti-HCV tests for the HCC patients are shown in figure 1 below.
Figure 1 Distribution of HBsAg and anti-HCV among HCC patients according to age range.
Discussion
Hepatocellular carcinoma is generally associated with increasing age and significantly higher HBV and HCV prevalence have been reported among persons in their 40's and 50's respectively [4,19]. Some studies conducted in the West African region have found a comparatively higher HBsAg positivity in those 41 years and above.
In this study, patients had no prior knowledge of their HCV status because HCV testing was not performed routinely, and this test was not available before this study. A summary of the results showed that the mean age of the HCC patients with HbsAg as 48.2 years, while the only patient with anti-HCV was aged 54 years. This finding is in line with similar reports [4,19]
HCC is more associated with males than females [8]. The finding of a comparatively higher prevalence of HbsAg and HCV among the male subjects in this study reflects the results of other studies [8]. However, in this study the difference was not statistically significant (p > 0.05), although it may be due to the small number of women participants. A similar reason may be advanced for the finding of a comparatively higher prevalence of HBsAg among the female case control subjects than the males
This study reveals an HBsAg prevalence of 15.2 % (5/33) among the apparently healthy Gambian population. Even though the sample size in the study was small, the finding is of major public health significance. The finding also supports a report that suggests that the virus is endemic in the country [20]. An aggressive HBV immunization exercise carried out in the country is believed to have drastically reduced the incidence of HBV [14]. An earlier work almost a decade ago estimated an HBV prevalence of 15–20% among the Gambian population [20]. The prevalence found in this study may therefore be suggestive of stable existence of the virus in the region.
The finding of lower anti-HCV prevalence among the HCC patients and none among the control subjects reflects the lower prevalence of HCV and the possible low involvement in HCC in the country. These findings are in line with a recent report and also provide support for a similar work conducted in Senegal a country that shares an extensive border, language and cultural similarities with the Gambia [14,20]. Globally 52.3% of HCC is attributed to HBV while HCV account for about 25% [8].
The observation of higher rates of hepatitis B surface antigen (HBsAg), and anti-HCV in patients with HCC than in control subjects suggests the involvement of these viruses in HCC [21]. Similarly, the finding of anti-HCV and HBsAg independently among HCC patients suggests that these two viruses might contribute independently to the pathogenesis of HCC. This finding supports the assertion of the independent roles of HBV and HCV in the pathogenesis of HCC [5]. Some studies have reported the synergistic role of HBV and HCV in HCC [2]. In this study no patient was found with HBV and HCV infection simultaneously. This may however be due to the number of participants enrolled in the study.
The observed involvement of HBV and HCV in HCC patients in a ratio of 5:1 in this study is higher than that reported previously [14]. This difference could have resulted from the comparatively small study population. The World Health Organization estimated an HCV prevalence of 2.4% for West Africa region [22]. However, a close epidemiological association between the HCC and HCV was not found in Senegal [20]. Their findings suggest that the main viral cause of HCC in the Senegal remains HBV.
In this study the possible attributable fraction of HCC due to HBV or HCV is 46%, thus suggestive of the involvement of other factors. In some countries HCC has been associated with chronic exposure to toxins originating from Aspergillus infected grains and peanuts [8]. Other associated risk factors includes cigarette smoking, prolonged abuse of alcohol in addition to some hereditary factors [5,8,23]. These factors were not evaluated in this study nor are their contributory role as causative agents of HCC in the Gambia known, however grains are the country most staple food while cigarette smoking and groundnut consumption are very common habits in the Gambia. There is therefore need for studies to evaluate the possible involvement of non-viral factors in HCC in the country.
Conclusion
These results suggest that HBV is endemic in the country and is present in apparently healthy persons. It also reveals that both HBV and HCV are actively involved in HCC in the region, in a ratio of 5:1 and that these two viruses might be independently involved in the pathogenesis of the disease. The results indicate that more than 50% of HCC cases in the country may be due to non-viral factors and calls for further studies to address this.
The study revealed a marginally statistically association (p = 0.04) between HBsAg and HCC patients (95%CI: 1.03–8.73, OR: 4.25). A similar level of association was found between HBsAg and with the case control subjects. No such association was found for HCV. These findings make it necessary for provision of diagnostic facilities for these viruses in hospitals and blood banks while intervention programmes should be put in place.
CIM designed the study and carried out laboratory work
AD analysed data and revised manuscript
MF critically revised manuscript
AJ conceived and organized the study, and revised the manuscript
Acknowledgements
CI is particularly grateful to Dr Sam Omar of the department of State for health for approval of this work, the chief Medical Director of Royal Victoria Teaching Hospital Banjul, Dr (Mrs.) Esangbedo and the head of the pathology department Mr. Jaye for permitting the use of laboratory facilities, Mr. Sylvester Onovo for the HCC patients, Messer Emmanuel Ogenekun, Batchilly, Faal, Jobarteh who gave generously their time and Dr Mark Egbe for some reagents and material support.
==== Refs
Mets T Smitz J Ngendahayo P Sabbe L Bigilimana I Ngirabatware B Hepatitis C virus infection in African patients with liver cirrhosis or primary hepatocellular carcinoma Scand J Gastroenterol 1993 28 331 334 8387695
Chuang WL Chang WY Lu SN Su WP Lin ZY Chen SC Hsieh MY Wang LY You SL Chen CJ The role of hepatitis B and C viruses in hepatocellular carcinoma in a hepatitis B endemic area. A case-control study Cancer 1992 69 2052 2054 1311978
Hassan MM Zaghloul AS El-Serag HB Soliman O Patt YZ Chappell CL Beasley RP Hwang LY The role of hepatitis C in hepatocellular carcinoma: a case control study among Egyptian patients J Clin Gastroenterol 2001 33 123 6 11468438 10.1097/00004836-200108000-00006
Shiratori Y Shiina S Imamura M Kato N Kanai F Okudaira T Characteristic difference of hepatocellular carcinoma between hepatitis B- and C- viral infection in Japan Hepatology 1995 22 1027 1033 7557847 10.1016/0270-9139(95)90605-3
Sun CA Wu DM Lin CC Lu SN You SL Wang LY Wu MH Chen CJ Incidence and cofactors of hepatitis C virus-related hepatocellular carcinoma: a prospective study of 12,008 men in Taiwan Am J Epidemiol 2003 157 674 682 12697571 10.1093/aje/kwg041
Monto A Wright TL. The epidemiology and prevention of hepatocellular carcinoma Semin Oncol 2001 28 441 449 11685737 10.1053/sonc.2001.26947
El Saadany S Tepper M Mao Y Semenciw R Giulivi A An epidemiologic study of hepatocellular carcinoma in Canada Can J Public Health 2002 93 443 446 12448868
Montalto G Cervello M Giannitrapani L Dantona F Terranova A Castagnetta LA Epidemiology, risk factors, and natural history of hepatocellular carcinoma Ann N Y Acad Sci 2002 963 13 20 12095924
Sulkowski MS Mast EE Seeff LB Thomas DL Hepatitis C virus infection as an opportunistic disease in persons infected with human immunodeficiency virus Clin Infect Dis 2000 30 77 84 10.1086/313842
Martinson FE Weigle KA Mushahwar IK Weber DJ Royce R Lemon SM Seroepidemiological survey of hepatitis B and C virus infections in Ghanaian children J Med Virol 1996 48 278 283 8801290 10.1002/(SICI)1096-9071(199603)48:3<278::AID-JMV11>3.0.CO;2-9
Hodges M Sanders E Aitken C Seroprevalence of hepatitis markers; HAV, HBV, HCV and HEV amongst primary school children in Freetown, Sierra Leone West Afr J Med 1998 17 36 37 9643158
Ampofo W Nii-Trebi N Ansah J Abe K Naito H Aidoo S Nuvor V Brandful J Yamamoto N Ofori-Adjei D Ishikawa K Prevalence of blood-borne infectious diseases in blood donors in Ghana J Clin Microbiol 2002 40 3523 3525 12202610 10.1128/JCM.40.9.3523-3525.2002
Ola SO Otegbayo JA Odaibo GN Olaleye OD Olubuyide OL Serum hepatitis C virus and hepatitis B surface antigenaemia in Nigerian patients with acute icteric hepatitis West Afr J Med 2002 21 215 217 12744571
Kirk GD Lesi OA Mendy M Akano AO Sam O Goedert JJ Hainaut P Hall AJ Whittle H Montesano R The Gambia Liver Cancer Study: Infection with hepatitis B and C and the risk of hepatocellular carcinoma in West Africa Hepatology 2004 39 211 219 14752840 10.1002/hep.20027
Kirk GD Camus-Randon AM Mendy M Goedert JJ Merle P Trepo C Brechot C Hainaut P Montesano R Ser-249 p53 mutations in plasma DNA of patients with hepatocellular carcinoma from The Gambia J Natl Cancer Inst 2000 92 148 153 10639517 10.1093/jnci/92.2.148
Mbaye PS Renaudineau Y Diallo A Haudrechy D Sane M Michel G Raphenon G Klotz F Hepatitis C virus and chronic hepatopathies in Dakar: case-control study Med Trop 2000 60 47 52
Ferrer F Candela MJ Garcia C Martinez L Rivera J Vicente V A comparative study of two third-generation anti-hepatitis C virus ELISAs Haematologica 1997 82 690 693 9499668
Fleiss JL Statistical methods for rates and proportions 1981 2 New York: John Wiley 23 48
Huh K Choi SY Whang YS Lee DS Prevalence of viral hepatitis markers in Korean patients with hepatocellular carcinoma J Korean Med Sci 1998 13 306 310 9681811
Ka MM Herve P Leguenno B Ndiaye MF Diop TM Diop B Dangou JM Bao O Brechot C Low prevalence of ant-hepatitis C virus antibodies in hepatocellular Carcinoma in Senegal Ann Gastroenterol Hepatol (Paris) 1996 31 329 332 8920074
Lu SN Lee CM Changchien CS Chen CJ Excess mortality from hepatocellular carcinoma in an HCV-endemic township of an HBV-endemic country (Taiwan) Trans R Soc Trop Med Hyg 1999 93 600 6022 10717743 10.1016/S0035-9203(99)90063-9
Madhava V Burgess C Drucker E Epidemiology of chronic hepatitis C virus infection in sub-Saharan Africa Lancet Infect Dis 2003 2 293 302 12062995 10.1016/S1473-3099(02)00264-5
Ogimoto I Shibata A Kurozawa Y Nose T Yoshimura T Suzuki H Iwai N Sakata R Fujita Y Ichikawa S Fukuda K Tamakoshi A Japan Collaborative Cohort Study Group. Risk of death due to hepatocellular carcinoma among drinkers and ex-drinkers. Univariate analysis of JACC study data Kurume Med J 2004 51 59 70 15150901
|
16202160
|
PMC1262750
|
CC BY
|
2021-01-04 16:38:37
|
no
|
Int Semin Surg Oncol. 2005 Oct 4; 2:20
|
utf-8
|
Int Semin Surg Oncol
| 2,005 |
10.1186/1477-7800-2-20
|
oa_comm
|
==== Front
J Exp Clin Assist ReprodJournal of Experimental & Clinical Assisted Reproduction1743-1050BioMed Central London 1743-1050-2-131619755410.1186/1743-1050-2-13ResearchFollicular fluid levels of vascular endothelial growth factor and early corpus luteum function during assisted reproductive technology cycles Coppola F [email protected] B [email protected] L [email protected] V [email protected] MC [email protected] G [email protected] Center for Reproductive Medicine – Department of Obstetrics, Gynecology and Neonatology – University of Parma – 43100 Parma – Italy2005 30 9 2005 2 13 13 7 6 2005 30 9 2005 Copyright © 2005 Coppola et al; licensee BioMed Central Ltd.2005Coppola et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The relation between vascular endothelial growth factor (VEGF) and early luteal function has rarely been proven in humans. The purpose of this study was to define the relation between follicular fluid concentrations of VEGF (FF VEGF) and early luteal function at the preimplantation stage during assisted reproductive technology (ART) cycles.
Methods
71 women were divided into two groups, based on reproductive outcome: women who became pregnant after embryo transfer (ET) (n = 18, Group A) and non-pregnant women (n = 53, Group B). Serum progesterone (Se P) and inhibin A on ET day, and FF VEGF levels were measured in all women. Data were expressed as mean ± standard deviation. Statistical analysis was performed using Excel Office 98 for Student's t-test, linear regression test and chi-square test. A p value of < 0.05 was considered statistically significant.
Results
The groups were comparable for age, ovarian reserve, number and quality of the oocytes retrieved and of the embryos obtained and transferred. FF VEGF levels were increased (4235 ± 1433 vs 3432 ± 1231 pg/ml), while Se P and inhibin A levels were significantly reduced (83.1 ± 34.1 vs 112.0 ± 58.8 ng/ml and 397.4 ± 223 vs 533.5 ± 283 pg/ml, respectively) in the non-pregnant group and were negatively correlated with FF VEGF (r = -0.482, p < 0.05; r = -0.468, p < 0.05) only in pregnant women.
Conclusion
Much has to be learned about the regulation and role of VEGF during the early luteal phase. We advance the hypothesis that the existence of a negative correlation between FF VEGF/Se P and FF VEGF/inhibin A in pregnant women might indicate the existence of a normal VEGF-mediated paracrine response when Se P and inhibin A levels are decreased. Excess production of FF VEGF and the absence of a correlation between FF VEGF/Se P and FF VEGF/inhibin A in non-pregnant women may be a paracrine reaction to immature luteal vasculature, resulting in luteal dysfunction.
==== Body
Background
Several aspects of reproduction, from folliculogenesis to corpus luteum function, from embryo implantation to placental function, are related to angiogenesis [1-3]. The relation between vascular endothelial growth factor (VEGF) and luteal vascularization has been demonstrated in a large number of studies on animals [1-3], but it has rarely been proven in humans [4] and during assisted reproductive technology (ART) cycles [5]. This growth factor, which is produced by granulosa and thecal cells in response to FSH, LH and hCG production, and to hypoxia, is a potent angiogenic cytokine. It plays a prominent role in the development of an extensive, mature vascular network around the follicles and within the corpus luteum [4]. In a group of 13 women undergoing ART and regardless of reproductive outcome, Lee A. et al. [5] found that VEGF concentration in the follicular fluid (FF VEGF) at the time of oocyte retrieval (day 0) was positively correlated with serum (Se P) and follicular fluid (FF P) progesterone. This in turn indicates a correlation between FF VEGF and follicular luteinization at a very early stage, in a process that can be described as the first wave of angiogenesis. At 11 to 14 days after ET (day +11 to +13), the same authors [5] reported that the pregnant recipients of autologous fresh embryos had higher serum VEGF levels than both non-pregnant recipients of autologous fresh embryos and pregnant recipients of donor eggs. These data suggest that at onset of gestation, endogenous hCG produced by the embryo stimulates ovarian VEGF production, a process also known as the second wave of angiogenesis.
The purpose of this study was to define the relation between FF VEGF levels and early luteal function at the preimplantation stage – 3 days after oocyte retrieval (day +3), before endogenous hCG produced by the embryo stimulates ovarian VEGF production – and to verify whether dysfunction in VEGF production may interfere with luteal function during ART.
Methods
Patients
The study was performed on follicular fluid and blood samples routinely collected on oocyte pick-up day and 3 days after oocyte retrieval, respectively, in 71 infertile women referred for IVF-ET to the Center for Reproductive Medicine of the University of Parma Department of Obstetrics, Gynecology and Neonatology between September 2002 and March 2003. Inclusion criteria were: normogonadotropic women regularly menstruating with normal ovulatory function and normal corpus luteum processes, age under 40, and baseline FSH levels lower than 10 IU/L. Exclusion criteria were: genital conditions that might interfere with implantation, e.g. fibroma, hydrosalpinx, uterine malformations, polycystic ovary (PCO), endometriosis, and such complications as ovarian hyperstimulation syndrome (OHSS), chemical or clinical abortion. The patients were divided into two groups, based on reproductive outcome following ART: ongoing pregnancy after ET (Group A, n = 18) and non-pregnancy (Group B, n = 53). Permission for use of the follicular fluid and blood samples for this study was given by the University of Parma Ethics Committee. All patients gave their informed consent to IVF-ET.
Stimulation protocol
All patients underwent controlled ovarian hyperstimulation (COH) following pituitary desensitization with a GnRH agonist (leuprolide 0.5 mg/day s.c. by Takeda Italia Farmaceutici, Rome, Italy) administered on day 21 of the cycle. The women were first monitored for inhibition of all ovarian activity and then stimulated with a standard dose of 225 IU of recombinant FSH (rFSH) (Gonal F 75 by Serono, Rome, Italy). The dose was later adjusted according to each patient's response. Ovarian response was monitored by serum estradiol (Se E2) assay and transvaginal sonography every other day. Finally, the patients received a dose of 10,000 IU of hCG (Profasi HP by Serono, Rome, Italy) when at least two follicles were > 16 mm and Se E2 levels were > 1000 pg/ml. The thickness of the endometrium was assessed by transvaginal sonography on the day of hCG administration.
Oocyte retrieval and ET
Oocytes were retrieved for IVF 35 hours after hCG administration by transvaginal ultrasound-guided aspiration. The oocytes thus obtained were tested in the laboratory and classified into four maturation stages depending on the maturity of the oocyte-cumulus-corona complex. For the purpose of this study, oocytes were considered mature when they showed an extensive dispersal of the investing granulosa cells and an expanded cumulus and corona, while the zona pellucida was distinct and the ooplasm clear (Type 1). Oocytes of intermediate maturity manifested a slightly denser corona and a dispersed cumulus, and lacked a nuclear membrane (Type 2). Immature oocytes were surrounded by a compact corona and few cell layers in the cumulus (Type 3). Atretic or postmature oocytes displayed a dark, irregular ooplasm (Type 4). Embryos were classified according to the criteria proposed in 1994 by C. Staessen et al. [6]: Type 1, or excellent embryos (blastomers of equal size or, if not of equal size, without anucleate fragments); Type 2, or good embryos (blastomers of equal size or, if not of equal size, with up to 20% the volume of the embryo filled with anucleate fragments); and, Type 3, or fair embryos (anucleate fragments present in 20–50% the volume of the embryo). ET was performed 3 days after oocyte retrieval (110 hours after hCG administration). ET was considered difficult when there was blood on the catheter after transfer. The luteal phase was supported with 200 mg micronized progesterone (Esolut by Angelini S.p.A, Rome, Italy) administered daily (100 mg in the morning and 100 mg in the evening) by vaginal route, starting on the day of oocyte pick-up and continuing until the day of the pregnancy test.
Pregnancy was considered ongoing if serum levels of beta hCG (Se βhCG) at day 16 after oocyte pick-up were above 112 mIU/mL (10th percentile of our trend curve) and confirmed by ultrasound at day 35 after oocyte pick-up.
Follicular fluid and blood sampling and assays
The patients' blood samples were tested: i) on day 3 of the cycle, 3 to 6 months prior to ovarian stimulation (baseline FSH); ii) on day 6 of ovarian stimulation (Se LH); iii) on hCG day (Se E2); and, iv) on ET day (Se P and inhibin A).
Se E2 (pg/ml) and Se P (ng/ml) levels were determined in duplicate using a chemoluminescence kit (Medical System, Genoa, Italy). Se FSH (IU/L) and Se LH
(IU/L) were determined in duplicate by chemoluminescence using Coat-A-Count FSH and LH IRMA kits (manufactured by Euro/Diagnostic Products Corporation, Witney, Oxfordshire, UK).
Inhibin A measurement (pg/ml) was performed by solid-phase sandwich enzyme-linked immunosorbant assay (ELISA) using a DSL kit manufactured by Webster, Texas, USA, and distributed by Pantec, Turin, Italy.
Inhibin A reflects ovarian function and is a product not only of the fetoplacental unit but also of the corpus luteum in early pregnancy [7-11]. It would then represent a more reliable index of corpus luteum function than Se P in women concomitantly receiving vaginal micronized progesterone [7].
The follicular fluid pool from each patient was centrifuged at 900 g/min for 15 minutes to remove cellular and blood contamination and then kept at -70°C for the subsequent determination in duplicate of FF VEGF. The assay technique used in this study was: ELISA for total FF VEGF (pg/ml) (R&D System, Minneapolis, Minnesota, USA).
Intra- and interassay parameter variations were 4% and 4.2% for E2, 2.2% and 4% for FSH, 1% and 2.2% for LH, 6.2% and 6.7% for progesterone, 6.0% and 7.4% for inhibin A, respectively, and < 8% for VEGF.
Statistical analysis
Data were expressed as mean (M) ± standard deviation (SD). Statistical analysis was performed using Excel Office 98 for Student's t-test, linear regression test and chi-square test. A p value of < 0.05 was considered statistically significant.
Results
The characteristics of the patients under study – number, age, ovarian reserve, as well as Se LH levels on day 6 of ovarian stimulation (+6), and Se E2 levels on hCG day – are reported in Table 1.
Table 1 Patient population characteristics.
Group A Group B p
Cases (no.) 18 53
Age (M ± SD) 33.9 ± 3.6 34.0 ± 3.13 n.s.
Baseline Se FSH (IU/L) (M ± SD) 7.1 ± 2.4 6.8 ± 2.5 n.s.
Se LH, day +6 (IU/L) (M ± SD) 1.2 ± 1.0 1.3 ± 1.1 n.s.
Se E2, hCG day (pg/ml) (M ± SD) 1401 ± 597 1467 ± 716 n.s.
rFSH (IU/pt) (M ± SD) 3098 ± 949 3562 ± 1256 n.s.
As no significant differences emerged between women who became pregnant after ET (Group A) and non-pregnant women (Group B), the two groups could be considered comparable in view to the examined parameters. A comparison of data for the IVF-ET procedure in the two groups (Table 2) did not show any differences in the number and quality of the oocytes retrieved, in the embryos obtained and transferred in utero and in difficult ET. Measurement of endometrial thickness did not differ in the two groups, either (Table 2).
Table 2 IVF-ET data. Values are given as M ± SD
Group A Group B p
Total oocytes/pt 8.4 ± 4.0 8.1 ± 4.3 n.s.
Oocytes type 1,1–2/pt 7.9 ± 3.6 7.0 ± 4.2 n.s.
Total embryos/pt 4.2 ± 2.0 3.7 ± 1.8 n.s.
Embryo type 1/pt 2.6 ± 1.8 2.0 ± 1.9 n.s.
Endometrial thickness (mm) 11.1 ± 2.9 10.9 ± 2.7 n.s.
Embryos transferred/pt 3.1 ± 0.8 3.0 ± 1.0 n.s.
Difficult ET (no./%) 1/18 5,5% 2/53 3,7% n.s.
FF VEGF levels – an expression of the follicular paracrine environment – were significantly elevated (p = 0.039) in non-pregnant women (4235 pg/ml ± 1443) compared with women who became pregnant after ET (3432 pg/ml ± 1231) (Table 3).
Table 3 Follicular environment and luteal function. Values are given as M ± SD
Group A Group B p
FF VEGF (pg/ml) 3432 ± 1231 4235 ± 1443 0.039
Se P (ng/ml) 112.0 ± 58.8 83.1 ± 34.1 0.013
Inhibin A (pg/ml) 533.5 ± 283 397.4 ± 223 0.042
FF VEGF vs Se P r = -0.482 (a) r = -0.178 (b) (a) < 0.05; (b) n.s.
FF VEGF vs inhibin A r = -0.468 (c) r = -0.092 (d) (c) < 0.05; (d) n.s.
(a) and (b) are correlation coefficients of FF VEGF to FF P
(c) and (d) are correlation coefficients of FF VEGF to inhibin A
Se P and inhibin A levels measured in the early luteal phase were significantly (p = 0.013 and p = 0.04) elevated in Group A (112.0 ng/ml ± 58.8 and 533.5 pg/ml ± 283) versus Group B (83.1 ng/ml ± 34.1 and 397.4 pg/ml ± 223) (Table 3) and were negatively correlated with FF VEGF (r = -0.482, p < 0.05; r = -0.468, p < 0.05) only in Group A; no correlation (r = -0.179, p n.s.; r = -0.09, p n.s.) was found in Group B (Table 3). Despite the concomitant administration of vaginal micronized progesterone, Se P was well correlated (r = 0.646, p < 0.001) with inhibin A (Fig 1). This reduced the degree of inaccuracy that Se P might exhibit versus inhibin A in luteal function expression.
Figure 1 Correlation between Se P and inhibin A.
Discussion
In our study, the two patient groups were comparable, not only for clinical characteristics, but also for the number and quality of the oocytes retrieved and of the embryos obtained and transferred in each group. Se P and inhibin A levels were significantly (p = 0.013 and p = 0.042) more elevated in pregnant women versus women who failed to conceive, while the angiogenic reaction was more marked (p = 0.03) in the latter. The cause of this excess angiogenic reaction is not yet clear. Under normal conditions, hCG administered to trigger ovulation induces the formation of a vascular network (first wave of angiogenesis) that helps ensure the correct development and functioning of the corpus luteum as a primary source of circulating progesterone [12] and inhibin A. In our study, the existence of a negative correlation (r = -0.482 and r = -0.468, p < 0.05) between FF VEGF/Se P and FF VEGF/inhibin A in pregnant women enabled us to demonstrate the existence of a normal VEGF-mediated paracrine reaction when Se P and inhibin A levels were decreased.
Beyond a given limit, as is the case with non-pregnant patients in whom there is no such correlation (r = -0.178 and r = -0.092, p n.s.), excess VEGF production is no longer enough to induce the formation of a mature vascular network in the corpus luteum through a paracrine mechanism and consequently to sustain adequate production of progesterone and inhibin A. This is also confirmed by studies on animals [13,1,14], revealing that luteal function can be impaired by anti-VEGF treatments down to a 50% reduction in Se P concentrations [2]. A large part of endothelial cells in the vascular network of the corpus luteum depend on VEGF support. Several variables, such as oxygen tension [15,16], insulin-like growth factor [3], aging [15], and poor response to ovarian stimulation [17], modulate the expression of this angiogenic factor. The pre-ovulatory follicle is a relatively large avascular multicellular structure [18]. Using mathematical formulas, it has been calculated that O2 content is potentially below the normal-range threshold (underoxygenation) [19], since the amount of O2 within mature follicles has been documented to range from less than 1% to about 5.5% [20]. Luteal expression of VEGF occurs primarily in specific perivascular cells, including arteriolar smooth muscle and capillary pericytes, and is regulated primarily by oxygen levels [12].
The reduced effect of exogenous hCG given to trigger ovulation on follicular luteinization in the women who failed to conceive and the consequent failure to achieve pregnancy in spite of supplemental progesterone administration in the luteal phase raises the question of whether exogenous hCG is capable of sustaining an adequate corpus luteum vascularization and whether progesterone can effectively substitute for luteal function. Our data seem to indicate that luteal function before ET is improved only in women who become pregnant. This might be due not only to improved corpus luteum vascularization, but also to a consequently adequate production of other local factors [21] of luteal origin (inhibin, oxitocin, growth factor, cytokines, prostaglandins and leucotrienes), which act in an autocrine or paracrine manner to modulate and perhaps mediate the development and function of the corpus luteum. Further studies are needed to investigate this aspect and lead to an in-depth understanding of the corpus luteum regulation mechanisms. The achievement of pregnancy following ART depends on a variety of factors (immunologic, genetic as well as endocrine, paracrine and vascular). In particular, getting to know with certainty why a woman fails to conceive is a crucial part of the process. Based on our study that evaluated only the maternal side prior to implantation in a selected sample of patients, we can hypothesize that suboptimal early luteal function – a condition related to immature vasculature – may play an important role in ART failure.
Conclusion
Much has to be learned about the regulation and role of VEGF during the final stage of follicular development and the early luteal phase. We advance the hypothesis that the negative correlation between FF VEGF and the early luteal phase in pregnant women and elevated FF VEGF levels in non-pregnant women could be ascribed to a paracrine response compensating for perifollicular hypoxia and that luteal malfunction in women that fail to conceive could be the expression of an immature luteal vasculature. This mechanism is likely to occur when there is not enough angiogenic compensation and represents a clinical application of what has already been found in tests on non-human primates. However, further studies are needed to determine the role played by angiogenic factors or other substances, such as recombinant hCG, in controlling the growth and maturation of perifollicular and luteal vasculature, in order to find new therapeutic strategies that may help treat sterility by manipulating angiogenesis.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
FC have made substantial contributions to conception, design and interpretation of data. BF have made substantial contributions to interpretation of data. LB performed the statistical analysis. VC have made substantial contributions to acquisition of data.
MCS have made substantial contributions to acquisition of data. GP have been involved in revising the article critically.
Acknowledgements
The authors would like to acknowledge Mario Rossi, L.T., Rinaldo Spallanzani L.T., Maria Grazia Ziveri B.Sc., and medical student Manuela Calabrese for their precious assistance in data collection.
==== Refs
Fraser HM Wulff C Angiogenesis in the primate ovary Reprod Fertil Dev 2001 13 557 66 11999306 10.1071/RD01055
Dickson SE Bicknell R Fraser HM Mid-luteal angiogenesis and function in the primate is dependent on vascular endothelial growth factor J Endocrinol 2001 168 409 16 11241172 10.1677/joe.0.1680409
Stouffer RL Martinez-Chequer JC Molskness TA Xu F Hazzard TM Regulation and action of angiogenic factors in the primate ovary Arch Med Res 2001 32 567 75 11750732 10.1016/S0188-4409(01)00323-X
Wulff C Dickson SE Duncan WC Fraser HM Angiogenesis in the human corpus luteum: simulated early pregnancy by HCG treatment is associated with both angiogenesis and vessel stabilization Hum Reprod 2001 16 2515 24 11726568 10.1093/humrep/16.12.2515
Lee A Christenson LK Stouffer RL Burry KA Patton PE Vascular endothelial growth factor levels in serum and follicular fluid of patients undergoing in vitro fertilization Fertil Steril 1997 68 305 11 9240261 10.1016/S0015-0282(97)81520-8
Staessen C Van den Abbeel E Janssenwillen C Devroey P Van Steirteghem AC Controlled comparison of Earle's balanced salt solution with Menezo B2 medium for human in vitro fertilization performance Human Reprod 1994 9 1915 19
Treetampinich C O'Connor AE MacLachlan V Groome NP de Kretser DM Maternal serum inhibin A concentrations in early pregnancy after IVF and embryo transfer reflect the corpus luteum contribution and pregnancy outcome Hum Reprod 2000 15 2029 2032 10.1093/humrep/15.9.2028
Petraglia F Garuti GC Calza L Roberts V Giardino L Genazzani AR Vale W Meunier H Inhibin subunit in human placenta: localization and messenger ribonucleic acid concentration during pregnancy Am J Obstet Gynecol 1991 165 750 758 1892206
Rombauts L Verhoven G Meuleman C Koninckx PR Poncelet E Franchimont P Dimeric inhibin A and alpha-subunit immunoreactive material in maternal serum during spontaneous and in-vitro fertilization pregnancies J Clin Endocrinol Metab 1996 81 985 989 8772561 10.1210/jc.81.3.985
Santoro N Schneyer AL Ibrahim J Schmidt CL Gonadotropin and inhibin concentrations in early pregnancy in women with and without corpora lutea Obstet Gynecol 1992 79 579 585 1553181
Yohkaichiya T Polson DW Hughes EG MacLachlan V Robertson DM Healy DL de Kretser DM Serum immunoreactive inhibin concentrations in early pregnancy after in vitro fertilization and embryo transfer: evidence for an ovarian source of inhibin and assessment of the inhibin measurements in predicting outcome of pregnancy Fertil Steril 1993 5 1081 1089 8486178
Reynolds LP Grazul-Bilska AT Redmer DA Angiogenesis in the corpus luteum Endocrine 2000 12 1 9 10855683 10.1385/ENDO:12:1:1
Fraser HM Lunn SF Angiogenesis and its control in the female reproductive system Br Med Bull 2000 56 787 97 11255562 10.1258/0007142001903364
Torry DS Holt VJ Keenan JA Harris G Caudle MR Torry RJ Vascular endothelial growth factor expression in cycling human endometrium Fertil Steril 1996 66 72 80 8752614
Friedman CI Danforth DR Herbosa-Encarnación C Arbogast L Alak BM Seifer DB Follicular fluid vascular endothelial growth factor concentrations are elevated in women of advanced reproductive age undergoing ovulation induction Fertil Steril 1997 68 607 12 9341598 10.1016/S0015-0282(97)00278-1
Koga K Osuga Y Tsutsumi O Momoeda M Suenaga A Kugu K Evidence for the presence of angiogenin in human follicular fluid and the up-regulation of its production by human chorionic gonadotropin and hypoxia J Clin Endocrinol Metab 2000 85 3352 55 10999833 10.1210/jc.85.9.3352
Battaglia C Genazzani AD Regnani G Primavera MR Petraglia F Volpe A Perifollicular Doppler flow and follicular fluid vascular endothelial growth factor concentrations in poor responders Fertil Steril 2000 74 809 12 11020528 10.1016/S0015-0282(00)01517-X
Neeman M Abramovitch R Schiffenbauer YS Tempel C Regulation of angiogenesis by hypoxic stress: from solid tumours to the ovarian follicle Int J Exp Pathol 1997 78 57 70 9203980 10.1046/j.1365-2613.1997.d01-247.x
Gosden RG Byatt-Smith JG Oxygen concentration gradient across the ovarian follicular epithelium: model, predictions and implications Hum Reprod 1986 1 65 68 3558757
Van Blerkom J Antczak M Schrader R The developmental potential of the human oocyte is related to the dissolved oxygen content of follicular fluid: association with vascular endothelial growth factor levels and perifollicular blood flow characteristics Hum Reprod 1997 12 1047 55 9194664 10.1093/humrep/12.5.1047
Stouffer RL Adash EY, Rock JA, Rosenwaks Z Corpus luteum formation and demise Reproductive Endocrinology, Surgery, and Technology 1996 1 Lippincott – Raven Philadelphia 251 269
|
16197554
|
PMC1262751
|
CC BY
|
2021-01-04 16:40:17
|
no
|
J Exp Clin Assist Reprod. 2005 Sep 30; 2:13
|
utf-8
|
J Exp Clin Assist Reprod
| 2,005 |
10.1186/1743-1050-2-13
|
oa_comm
|
==== Front
J Immune Based Ther VaccinesJournal of Immune Based Therapies and Vaccines1476-8518BioMed Central London 1476-8518-3-71618149410.1186/1476-8518-3-7Original ResearchMycobacterial immune reconstitution inflammatory syndrome in HIV-1 infection after antiretroviral therapy is associated with deregulated specific T-cell responses: Beneficial effect of IL-2 and GM-CSF immunotherapy Pires A [email protected] M [email protected] AL [email protected] M [email protected] B [email protected] F [email protected] N [email protected] Department of Immunology Imperial College London, Chelsea and Westminster Hospital, 369 Fulham Road, London. UK2 Department of HIV/GU Medicine, Chelsea and Westminster Hospital, 369 Fulham Road, London, UK3 Department of HIV/GU Medicine, Royal Sussex County Hospital, Brighton, UK2005 25 9 2005 3 7 7 6 4 2004 25 9 2005 Copyright © 2005 Pires et al; licensee BioMed Central Ltd.2005Pires et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
With the advent of antiretroviral therapy (ART) cases of immune reconstitution inflammatory syndrome (IRIS) have increasingly been reported. IRIS usually occurs in individuals with a rapidly rising CD4 T-cell count or percentage upon initiation of ART, who develop a deregulated immune response to infection with or without reactivation of opportunistic organisms. Here, we evaluated rises in absolute CD4 T-cells, and specific CD4 T-cell responses in 4 HIV-1+ individuals presenting with mycobacterial associated IRIS who received in conjunction with ART, IL-2 plus GM-CSF immunotherapy.
Methods
We assessed CD4 T-cell counts, HIV-1 RNA loads, phenotype for naïve and activation markers, and in vitro proliferative responses. Results were compared with those observed in 11 matched, successfully treated asymptomatic clinical progressors (CP) with no evidence of opportunistic infections, and uninfected controls.
Results
Median CD4 T-cell counts in IRIS patients rose from 22 cells/μl before initiation of ART, to 70 cells/μl after 8 months of therapy (median 6.5 fold increase). This coincided with IRIS diagnosis, lower levels of naïve CD4 T-cells, increased expression of immune activation markers, and weak CD4 T-cell responses. In contrast, CP had a median CD4 T-cell counts of 76 cells/μl at baseline, which rose to 249 cells/μl 6 months post ART, when strong T-cell responses were seen in > 80% of patients. Higher levels of expression of immune activation markers were seen in IRIS patients compared to CP and UC (IRIS > CP > UC). Immunotherapy with IL-2 and GM-CSF paralleled clinical recovery.
Conclusion
These data suggest that mycobacterial IRIS is associated with inadequate immune reconstitution rather than vigorous specific T-cell responses, and concomitant administration of IL-2 and GM-CSF immunotherapy with effective ART may correct/augment T-cell immunity in such setting resulting in clinical benefit.
Immune reconstitutionT cellsHIV-1Mycobacterial infectionMAC
==== Body
Background
The degree of immune reconstitution observed in HIV-1+ individuals following initiation of antiretroviral therapy (ART), is variable [1-4]. Although seen even in late-stage disease, it is more prominent in patients who commence treatment during early HIV-1 infection before substantial damage to the immune system, where robust responses are often seen after treatment [5-7]. Such responses likely reflect effective immune surveillance, mimicking the beneficial T-cell responses seen in untreated long-term non-progressors, HIV-1 exposed but seronegative individuals, and after therapeutic vaccination of asymptomatic patients [8-10].
It has been postulated that after treatment of late-stage HIV-1 infection, recovery and augmentation of immune function, and responses to previous sub-clinical infections with existing pathogens such as Mycobacterium spp, hepatitis B and hepatitis C viruses, or cytomegalovirus (CMV) may result in exacerbated inflammatory diseases [11-19]. This phenomenon, described by others as immune reconstitution inflammatory syndrome (IRIS), is mostly seen in profoundly immunosuppressed patients with CD4 T-cell nadirs of less than 100 cells/μl, who upon receiving ART rapidly achieve an undetectable plasma viremia, and experience a very rapid increase in CD4 T cells [12,14]. This complex syndrome presents with either active opportunistic infections, or recurrence of previous infections.
We investigated the quality and breadth of lymphoproliferative responses in a group of HIV-1+ patients on stable ART for > 6 months with suppressed viremia, diagnosed with Mycobacterium avium complex (MAC) associated IRIS, who were unresponsive to conventional anti-MAC therapy. We showed that these patients lacked pathogen-specific in vitro T-cell responses suggesting that the degree and quality of immune reconstitution following ART is inadequate to eliminate underlying opportunistic infections. Furthermore, immunotherapy with IL-2 and GM-CSF in combination with effective ART appears to accelerate augmentation of specific CD4 T-cell responses and increase the rapidity of immune recovery allowing underlying opportunistic infections to be cleared and leading to a better immediate outcome and resolution of IRIS.
Methods
Subjects studied
Fifteen HIV-1+ patients at the Chelsea and Westminster Hospital, London, UK were studied. Four presented with MAC-associated IRIS, and where acid-fast bacilli were detected, patients were given anti-MAC therapy as soon as diagnosed. These patients had a median CD4 T-cell count of 22 cells/μl (interquartile range (IQR) 6.3–50.3) before initiation of ART, rising to 70 cells/μl (IQR 63–123) after 8 months of therapy (Table 1). For clarity these subjects will be referred to as IRIS patients. Previous reports define IRIS as a syndrome occurring in individuals with a rising CD4 T-cell count or percentage upon initiation of ART, who develop new clinical pathologies with either a new clinical presentation or reactivation of opportunistic organisms [15,16]. Viral load was undetectable in all patients at presentation of IRIS. IRIS patients (n = 4) received immunotherapy as salvage therapy consisting of IL-2 (Chiron Therapeutics, Uxbridge, UK) at 5 million units twice daily subcutaneously for 5 days, in three cycles 4 weeks apart. During the third cycle of IL-2, concomitant GM-CSF (Novartis, Schering-Plough, Camberley, UK) was administered subcutaneously 150 μg daily for 5 days. The remaining patients were asymptomatic clinical progressors (n = 11) receiving ART for 6 months, with a median CD4+ T-cell count of 76 cells/μl (IQR 22.5–90) at baseline, rising to 249 cells/μl (IQR 187.5–303.5) 6 months post ART, and with viral load levels from undetectable (80% of patients) to 127 HIV-1 RNA copies/ml plasma. These patients developed no secondary effects following treatment, and had no evidence of opportunistic infections/exacerbated immune responses. Sixteen healthy HIV uninfected donors were used as controls. Informed consent was obtained from all patients for the administration of immunotherapy and investigations carried out, and ethics committee approval was obtained for the studies described.
Table 1 Clinical features of IRIS patients
Patient CD4 T cell count before ART cells/μl CD4 T cell count at presentation of IRIS cells/μl Fold change in CD4 T cell counts from baseline to IRIS presentation CD4 T cell count after remission of IRIS cells/μl HIV-1 RNA at presentation of IRIS copies/ml Reason for admission Time on therapy* Therapy
1 7 69 9.86 202 U/D MAC 8 d4T+ddI+NFV+ImRx
2 37 70 1.89 140 U/D MAC 12 AZT+3TC+IDV+ImRx
3 4 45 11.25 93 U/D MAC 18 d4T+ddI+NFV+ImRx
4 90 280 3.11 601 U/D MAC 8 AZT+3TC+EFV+ImRx
*Time in months from initiation of potent ART until diagnosis of IRIS. Drugs used in ART regime: Nucleoside analogues; Stavudine (d4T), Didanosine (ddI), Lamivudine (3TC) and Zidovudine (AZT) Protease inhibitors; Nelfinavir (NFV), Indinavir (IDV), or Non-nucleoside reverse transcriptase inhibitor; Efavirenz (EFV); ImRx = immunotherapy; MAC = Mycobacterium avium complex.
Plasma viral RNA assay
Viral load was measured at each time point of sample collection using the Bayer HIV-1 RNA 3.0 assay (bDNA) (Bayer Diagnostics, Newbury, UK) with lower detection limit of 50 HIV-1 RNA copies/ml plasma.
Lymphocyte subset quantification
The Epics XL-MCL (Beckman Coulter, High Wycombe, UK) was used for four-colour flow cytometric analysis. Anti-human CD3, CD4, CD8, and CD45 were used to analyze T cell subsets. Leukocytes were analysed on the Epics XL-MCL flow cytometer using system II software in conjunction with control reagents (Beckman Coulter) which provide automated colour compensation, light scatter and colour intensities.
T cell proliferation assay
Peripheral blood mononuclear cells (PBMC) were cultured in triplicate with HIV-1 or other recall/viral antigen in round-bottomed microtiter plates (Greiner, Gloucester, UK) for 5 days as described previously [20-22]. The antigens used were: Herpes simplex virus (HSV), purified avian protein derivative of tuberculin (PPD), tetanus toxoid (TTox), Varicella-Zoster virus (VZV), Candida (CAN), and Cytomegalovirus (CMV) as described in reference 21. HIV-1 recombinant antigens were obtained from the Medical Research Council Centralised Facility for AIDS Reagents (National Institute for Biological Standards and Controls, Potters Bar, UK) and comprised: recombinant HIV-1-nef, recombinant HIV-1-gp120 and recombinant HIV-1-p24 (all used at 10 μg/ml final concentration) [22]. Adjuvant-free Remune and its native-p24 antigens were a generous gift from Dr Ronald Moss (Immune Response Corporation, Carlsbad, USA) and were used at 3 μg/ml to ensure that anti-HIV-1 responses were not overlooked due to clade variability. On day 5, 100 μl of supernatant was collected from each well and stored at -20°C for subsequent cytokine measurement, cells were pulsed with [3H]thymidine (Amersham International, Amersham, UK) and 16 h later cells were harvested onto glass fiber filtermats and counted (Wallac Oy, Turku, Finland). Results are expressed as stimulation indices (SI) with a positive response defined as an SI of 3 or more and Δ counts per minute (CPM) > 600 as described previously [21-23]. Control wells, for calculation of background activity, contained PBMC only.
Measurement of IL-4 production
Fifty μl of supernatant from proliferative cultures was transferred to 96-well round-bottomed plates in triplicate for quantification of IL-4 on the indicator cell line CT.h4S (a generous gift of W. Paul, Bethesda, MD) as previously described [20-22]. Briefly, CT.h4S (5 × 103 cells/well), were added in 50 μl to 50 μl of supernatant to give a final volume of 100 μl. After 24 h in culture, wells were pulsed with [methyl-3H]thymidine, and cells were harvested as described above. Results are expressed as the mean cpm for triplicate cultures, with an error of the mean of ± 15%. A positive result is defined as significant proliferation above the background activity and detection threshold. In all experiments, a standard titration of indicator cell proliferation to a range of recombinant IL-4 from 0.01 to 100 U/ml was included. Control wells for calculation of background activity contained indicator cells only.
Phenotypic analysis of lymphocytes
PMBC were incubated with a panel of murine anti-human mAbs (all Beckman Coulter), for 30 minutes at 4°C. Directly conjugated antibodies used were: Fluorescein isothiocyanate (FITC)-CD8, CD45RA; Phycoerytherin (PE)-CD38, HLA-DR, CD27, and CD45RA; PE-cyanine (PC-5)-CD4, all used according to the manufacturer's instructions. Cells were washed and fixed in PBS containing 2% paraformaldehyde (Sigma). On acquisition, a gate was set around the lymphocyte population on a forward scatter versus side scatter dot plot, and 10,000 gated events collected for each sample. Data analysis was performed using CELLQuest™ Software (Becton Dickinson, Oxford, UK). Appropriate isotype matched controls were run in parallel for each sample.
Statistical analysis
Computer software (Statview 5.01; Abacus, Berkeley, CA) was used for all statistical calculations. Data are presented as median (inter-quartile range IQR). Analysis of data between the different groups was performed using a Mann Whitney-U-test and intra-group variations were compared using the Wilcoxon signed rank test. P values below 0.05 were considered significant.
Results
Patients
We studied both IRIS and CP patients who had been receiving effective ART for similar periods. IRIS was diagnosed at a median 10 months (IQR 8–13) after initiation of ART (Table 1). This is in agreement with previous reports [24]. The patients did not recover from the underlying MAC infection despite receiving conventional anti-MAC treatment, and were given IL-2 and GM-CSF in conjunction with ART as salvage therapy as detailed in materials and methods and as previously described [9,21]. Increases in CD4 T-cell counts from baseline to IRIS diagnosis were observed in all patients. Viral load reached BDL in all patients and remained undetectable throughout the study.
IRIS patients receiving ART plus immunotherapy
Immunotherapy was initiated for severely immuno-compromised patients with exacerbated underlying MAC infection (IRIS) who were unable to achieve remission after receiving ART and anti-mycobacterial therapy. Four patients with median CD4+ T-cell counts of 70 cells/μl (IQR 63–123) after a median 10 months on ART, with persistent MAC infection, received IL-2 and GM-CSF (see Table 1 for drug regimen). PPD-specific T-cell responses and responses to other recall/viral antigens were absent in all patients before immunotherapy (Fig 1a). After administration of immunotherapy, we saw an increase in median CD4+ T-cell counts to 171 cells/μl (IQR 128–302) (Table 1). Moreover, we observed robust antigen-specific T-cell responses to a panel of antigens including PPD. Such responses were sometimes more vigorous than those observed in patients on ART alone (Fig 1a), and were paralleled by remission from the underlying MAC infection. Furthermore, immune reconstitution characterised by a rise in CD4 T-cell counts and constant undetectable plasma viremia was achieved. Patient 4 was admitted with localised MAC associated lymphadenitis of the neck and lacked in vitro proliferative responses to PPD and other recall antigens despite a CD4 T-cell count of 280 cells/l μblood. After administration of immunotherapy we observed a rapid recovery in immune function (Fig 2a). The parallel clinical manifestations depicted an improvement in the neck lesion after IL-2 therapy and complete remission post administration of IL-2 plus GM-CSF (Fig 2b–d). We also observed an increase in CD4 T-cell counts from 280 to 601 cells/μl during this period (Table 1). No significant changes were seen in HIV-1-specific T-cell responses, which remained undetectable throughout the study (Fig 1b).
Figure 1 (a – top) Lymphoproliferative responses to a panel of recall antigens in IRIS patients during IRIS manifestation and post remission. Open bars denote T cell responses during IRIS manifestation and hatched bars represent T cell responses after immunotherapy with IL-2 plus GM-CSF in conjunction with ART and resolution of IRIS. Data are shown as median SI values with interquartile ranges. X-axis depicts the recall antigens tested. (b- bottom) HIV-1-specific lymphoproliferative responses in IRIS patients. Data depicted are before immunotherapy (solid bars), 4 weeks after IL-2 administration (crossed bars) and 4 weeks after IL-2 plus GM-CSF (hatched bars). Data are shown as median values with interquartile ranges.
Figure 2 (a – top) Specific lymphoproliferative responses to recall antigens of patient 4. Data are at IRIS presentation (white bars), 4 weeks after IL-2 administration (hatched bars) and 4 weeks post final IL-2 and GM-CSF dosing (solid bars). Photographs depict the clinical manifestation of MAC lymphadenitis of the neck in patient 4, at IRIS presentation (b – bottom left), 4 weeks after IL-2 administration (c- bottom centre), and 4 weeks after IL-2 plus GM-CSF administration (d – bottom right).
In 3/4 IRIS patients we carried out IL-4 bioassays in culture supernatants, rather than ELISA, in order to assess the levels of bioactive cytokine being produced. We were able to detect production of IL-4 in cultures with antigens to which the patients had been previously exposed including anti-PPD responses in 2/3 patients (Fig 3). Upon initiation of immunotherapy there was a decrease in IL-4 production, which was paralleled by restoration of proliferative specific-anti-PPD T-cell responses.
Figure 3 IL-4 production in culture supernatants of PBMC from 2 patients. Data are from 5 days stimulation with/without PPD and p24 antigens, before immunotherapy (white bars) and 4 weeks after administration of IL-2 plus GM-CSF (hatched bars). Data are expressed as the mean cpm (proliferation of cell line CT.h4S) for triplicate cultures, with an error of the mean of ± 15%.
Lymphoproliferative T-cell responses to recall/viral antigens in asymptomatic clinical progressors and seronegative controls
We assessed T-cell proliferation in CP and uninfected controls (UC) and compared these with the responses seen in IRIS patients. CP presented a median 3.9 fold increase in CD4 T-cell counts 6 months post initiation of ART from 76 cells/μl (IQR 22.5–90) to 249 cells/μl (IQR 187.5–303.5) (p < 0.001) (Table 1). These patients remained clinically asymptomatic and had detectable specific T-cell responses to at least one recall antigen (Fig 4). All UC showed vigorous responses to recall antigens (Fig 4).
Figure 4 Box-plots depicting specific lymphoproliferative responses to different recall antigens, during manifestation of IRIS and comparison with clinical progressors with no IRIS and uninfected controls. The antigens used are shown on the x-axis. Bars denote median responses with interquartile ranges. White boxplots represent IRIS patients receiving ART and immunotherapy. Non-IRIS patients are depicted by dotted boxplots and uninfected controls by hatched boxplots.
Flow cytometry revealed higher levels of CD38 and HLA-DR expression and lower levels of naïve CD4 T cells in IRIS patients than in asymptomatic clinical progressors
Compared to CP, IRIS patients showed significantly higher percentages of CD4+HLA-DR+ T lymphocytes (p < 0.005), and significantly higher percentages of CD8+CD38+ T cells (p < 0.05) (Table 2). When activation was quantified on a per cell basis, IRIS patients showed higher levels of activation of both CD4 and CD8 T cells compared to CP (p < 0.02 and p < 0.01, respectively), as demonstrated by analysing the mean fluorescent intensity levels of CD38 expression (Table 2). Furthermore, the median percentage of naïve CD4+CD45RA+CD27+ T cells in IRIS patients was significantly lower than in CP and UC (p < 0.005). These observations are not surprising as IRIS patients were more immuno-compromised when therapy was initiated suggesting that IRIS may be associated with persistent hyperactivation of both CD4 and CD8 T lymphocytes, and associated with a lack of naïve CD4 T cells possibly due to absence of thymic function.
Table 2 Percentages of different lymphocyte subsets in the CD4+ and CD8+ T cell population, in immune reconstitution inflammatory syndrome patients, clinical progressors and uninfected controls.
CD4 CD8
IRIS p value
HLA-DR 27.4 (16–38.4) 28.7 (24.6–32) < 0.005/n.s
CD38 77 (70–78.5) 82.6 (75.3–90.7) n.s/< 0.05
CD38 MFI 1118 (1253-780) 493 (548-297) < 0.02/< 0.01
Naïve 19.4 (6.9–23.9) 25 (7–42) < 0.005/n.s
Memory 85 (44–96) 53.8 (37.5–76) n.s/n.s
CP
HLA-DR 7.3 (5.4–9.2) 30.5 (22–40)
CD38 66 (60.7–71.7) 73 (61.7–80)
CD38 MFI 173 (85–211) 133 (105–146)
Naïve 35 (30–45) 32 (24.6–42)
Memory 59 (54–67) 38 (26–46)
UC
HLA-DR 4 (2.8–5.3) 7.6 (6.9–9.9)
CD38 70 (68–75.5) 66 (66–70)
CD38 MFI 97 (71–116) 85 (58–87)
Naïve 40.8 (40–49) 60.4 (51.2–61.5)
Memory 30 (29.2–40) 22 (20.8–30.3)
IRIS- Immune reconstitution inflammatory syndrome patients; CP- asymptomatic clinical progressors; UC- uninfected controls. Data are shown as median percentage, and CD38 mean fluorescent intensity (MFI) (interquartile range). p values shown between IRIS and CP patients for CD4/CD8 data respectively.
Discussion
Immune reconstitution after initiation of ART may be concurrent with both an increase in immuno-pathological responses against opportunistic pathogens and with the induction of IRIS [11-19,24]. The IRIS phenomenon has been ascribed to vigorous immune responses specific to underlying pathogens, with clinical manifestations related to the immune response elicited against such pathogens. Typically, IRIS patients have an undetectable viral load, and CD4 T-cell counts that have rapidly increased, by 3 or 4 fold, shortly after initiation of ART. Previous studies have used delayed type hypersensitivity (DTH) tests to assess the cell-mediated immunity of these patients [12,14,15]. In contrast, we used the thymidine incorporation assay to evaluate lymphocyte proliferation. This allows visualisation of in vitro immune function of T lymphocytes in peripheral blood and direct comparison with asymptomatic HIV-1+ subjects as well as uninfected controls. Some reports have shown the lack of correlation between these two assays [25], as functionally T-cell proliferation and DTH responses can diverge [26]. Therefore, by utilising the thymidine incorporation assay, we demonstrate the correlation between in vitro functional data and clinical evidence.
There were two important findings in this study. Firstly, patients admitted with IRIS lacked antigen-specific T-cell responses. Secondly, administration of IL-2 and GM-CSF appeared to rapidly introduce these responses and was associated with clinical recovery in patients with advanced HIV-1 infection.
Data from uninfected controls and treated asymptomatic clinical progressors, revealed the presence of lymphoproliferative responses to recall/viral antigens, compared to IRIS patients, confirming that functional specific T lymphocytes are associated with the control of opportunistic infections. Thus, the clinical picture presented by our cohort of IRIS patients is likely to be associated with the lack of lymphoproliferation and IL-2 production rather than with robust antigen-specific T cell responses. This suggests an alternative/additional mechanism for IRIS, distinct from previous hypotheses which suggest that IRIS is caused by pathogen-specific responses induced by successful ART.
Of note, are the generally weak proliferative responses observed in response to other pathogens such as CMV, which recover after initiation of ART [27,28]. Although complete immune impairment appears to be restricted to HIV-1 and PPD antigens, other studies have reported CMV-specific responses to remain generally unchanged in HIV-1+ patients, regardless of the patients' HIV viral loads and clinical state [29]. Different reasons may explain this observation: CMV viraemia is often very low or undetectable and thus fails to induce robust T cell responses [29]; in addition CMV does not target the antigen presenting cells such as dendritic cells and macrophages hence enabling the uninfected antigen presenting cells to efficiently carry out processing and presentation and to generate specific T-cell responses – unlike mycobacteria and HIV-1 which are both known to target antigen presenting cells. In addition, our group has previously shown that in general, HSV- and CMV-specific T-cell responses appear to be more robust in both CP and LTNP [22,28].
The mechanisms behind IRIS still remain elusive. Through cell-to-cell contact, various complex molecular interactions and the production of cytokines, CD4 T helper lymphocytes modulate the activity of all cells involved in both innate and acquired immunity, including virus-specific cytotoxic CD8 T lymphocytes [30-33]. Although anti-HIV-1 CD8 T cells are present in chronically infected individuals, they lack specific functional properties, most likely because CD4 T cell help is impaired [34-37]. The HIV-1-induced defect of CD4 T-cell responses is likely to be an underlying mechanism causative of anergy and subsequently IRIS, due to defective antigen presentation and lack of T-cell help.
The use of immunotherapy in severely immuno-compromised patients has been shown to have beneficial effects in partially reversing the CD4 T-cell defects exerted by HIV-1, when administered concomitantly with ART [9,21]. Such data concurs with our previous findings that concomitant administration of ART and IL-2 plus GM-CSF reversed the type 2 anti-proliferative cytokine environment, decreasing IL-4 levels and inducing pathogen-specific proliferative responses. We have previously shown that such regimens induce HIV-1-specific proliferative CD4 and IFN-γ secreting CD8 T-cells. In these studies, HIV-1-specific proliferative responses paralleled by increased IL-2 production, responsiveness and up-regulated expression of IL-2-specific mRNA were associated with remission of disease [21]. It is important to note that, in previous studies, GM-CSF was administered at doses twice the level of those described here, thus possibly inducing HIV-1-specific immunity. This may explain our inability to induce detectable HIV-1-specific T-cell responses despite a decline in IL-4 production. Regardless, in our cohort of immunotherapy recipients, remission from MAC occurred rapidly despite profound immuno-suppression.
During HIV-1 infection immune hyperactivation is accompanied by up-regulation of surface expression of CD38 and HLA-DR on CD4 and CD8 T cells [38-41]. Higher T-cell activation in the IRIS cohort compared to CP is not surprising, as T-cell specific responses were much lower in IRIS patients. This is in concordance with data from Caruso et al describing the occasional lack of correlation between the percentage of T cells expressing activation markers and thymidine incorporation [42]. It is suggested that hyperactivation by HIV-1 and other underlying pathogens is associated with activation induced cell death and/or anergy [43]. Furthermore, high HLA-DR levels are suggestive of increased T cell↔T cell antigen presentation, which is associated with induction of T-cell anergy/dysfunction and increased IL-4 production [44,45]. This is in agreement with our previous findings that anergised antigen-specific T cells are present at baseline in immunocompromised patients but lack proliferative/functional ability [21]. The lower levels of naïve T cells observed in IRIS patients compared to CP may be due to a more immunocompromised state and possibly reflect thymic dysfunction/inactivity.
Our data suggests that the degree of immune reconstitution achieved with potent ART alone is dependent on the clinical stage of the patient when therapy was initiated. Furthermore, we hypothesise that in some late-stage patients ART may elicit the expansion of abnormal/anergic T cell clones responsible for an erratic immune response. Concomitant administration of IL-2 and GM-CSF may be associated with provision of proliferative signals to immature thymocytes and/or rescue of anergic CD4 T cells to generate fully functional T-cell responses. GM-CSF acts directly on antigen presenting cells, including macrophages, which may induce and contribute to a more rapid remission from intracellular infection [46-48].
Conclusion
In conclusion, ART induced immune reconstitution restores specific responses, which likely help in the clearance of opportunistic pathogens. However, inadequate immune responses observed in treated late-stage disease, in addition to other possible confounders, may be causative of IRIS. Administration of IL-2 and GM-CSF concomitantly with ART in late-stage patients may result in proliferation of pathogen-specific CD4 T-lymphocytes, which likely enable a more rapid clearance of intracellular pathogens such as Mycobacterium avium. Such responses may be associated with a better outcome and, possibly quicker recovery, in immuno-compromised patients who fail to achieve immune reconstitution with ART alone, indicating that this therapeutic approach as salvage immuno-therapy may have an impact on short-term mortality. The small number of patients is noteworthy- this is nonetheless an interesting and provocative finding that deserves further prospective exploration in larger numbers of similar patients.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
AP carried out the proliferation assays, phenotypic analysis, data analysis, participated in the study design and wrote the manuscript. MN, ALP, MF and BG recruited patients and participated in the study design. FG participated in the study design and participated in the drafting of the manuscript. NI conceived the study, carried out proliferation assays and bioassays, data analysis and the design, coordination, the draft and finalisation of the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We would like to thank all patients and staff at Chelsea & Westminster Hospital who participated in this study; Ron Moss from the Immune Response Corp., Carlsbad San Diego CA, for the whole HIV-1 antigen and the 'native' clade G p24. This work was supported by the Wellcome Trust (Grant number: 058700) and the AVIP EU Programme (Grant number: LSHP-CT-2004-503487).
==== Refs
Landay AL Bettendorf D Chan E Spritzler J Schmitz JL Bucy RP Gonzalez CJ Schnizlein-Bick CT Evans T Squires KE Evidence of immune reconstitution in antiretroviral drug-experienced patients with advanced HIV disease AIDS Res Hum Retroviruses 2002 18 95 102 11839142 10.1089/08892220252779638
Autran B Carcelain G Li TS Blanc C Mathez D Tubiana R Katlama C Debre P Leibowitch J Positive effects of combined antiretroviral therapy on CD4+ T cell homeostasis and function in advanced HIV disease Science 1997 277 112 6 9204894 10.1126/science.277.5322.112
Li TS Tubiana R Katlama C Calvez V Ait Mohand H Autran B Long-lasting recovery in CD4 T-cell function and viral-load reduction after highly active antiretroviral therapy in advanced HIV-1 disease Lancet 1998 351 1682 6 9734884 10.1016/S0140-6736(97)10291-4
Palella FJ JrDelaney KM Moorman AC Loveless MO Fuhrer J Satten GA Aschman DJ Holmberg SD Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. HIV Outpatient Study Investigators N Engl J Med 1998 338 853 60 9516219 10.1056/NEJM199803263381301
Musey LK Krieger JN Hughes JP Schacker TW Corey L McElrath MJ Early and persistent human immunodeficiency virus type 1 (HIV-1)-specific T helper dysfunction in blood and lymph nodes following acute HIV-1 infection J Infect Dis 1999 180 278 84 10395840 10.1086/314868
Rosenberg ES Altfeld M Poon SH Phillips MN Wilkes BM Eldridge RL Robbins GK D'Aquila RT Goulder PJ Walker BD Immune control of HIV-1 after early treatment of acute infection Nature 2000 407 523 6 11029005 10.1038/35035103
Oxenius A Fidler S Brady M Dawson SJ Ruth K Easterbrook PJ Weber JN Phillips RE Price DA Variable fate of virus-specific CD4(+) T cells during primary HIV-1 infection Eur J Immunol 2001 31 3782 8 11745399 10.1002/1521-4141(200112)31:12<3782::AID-IMMU3782>3.0.CO;2-#
Rowland-Jones S Sutton J Ariyoshi K Dong T Gotch F McAdam S Whitby D Sabally S Gallimore A Corrah T HIV-specific cytotoxic T-cells in HIV-exposed but uninfected Gambian women Nat Med 1995 1 59 64 7584954 10.1038/nm0195-59
Imami N Hardy G Burton C Pires A Pido-Lopez J Moss R Gazzard B Gotch F Immune responses and reconstitution in HIV-1 infected individuals: impact of anti-retroviral therapy, cytokines and therapeutic vaccination Immunol Lett 2001 79 63 76 11595291 10.1016/S0165-2478(01)00267-X
Rosenberg ES Billingsley JM Caliendo AM Boswell SL Sax PE Kalams SA Walker BD Vigorous HIV-1-specific CD4+ T cell responses associated with control of viremia Science 1997 278 1447 50 9367954 10.1126/science.278.5342.1447
French MA Antiretroviral therapy. Immune restoration disease in HIV-infected patients on HAART AIDS Read 1999 9 548 9 11082733
French MA Lenzo N John M Mallal SA McKinnon EJ James IR Price P Flexman JP Tay-Kearney ML Immune restoration disease after the treatment of immunodeficient HIV-infected patients with highly active antiretroviral therapy HIV Med 2000 1 107 15 11737333 10.1046/j.1468-1293.2000.00012.x
French MA Mallal SA Dawkins RL Zidovudine-induced restoration of cell-mediated immunity to mycobacteria in immunodeficient HIV-infected patients Aids 1992 6 1293 7 1472334
French MA Price P Immune restoration disease in HIV-infected patients after antiretroviral therapy Clin Infect Dis 2001 32 325 6 11170934 10.1086/318465
Price P Mathiot N Krueger R Stone S Keane NM French MA Immune dysfunction and immune restoration disease in HIV patients given highly active antiretroviral therapy J Clin Virol 2001 22 279 87 11564593 10.1016/S1386-6532(01)00200-1
Shelburne SA Hamill RJ Rodriguez-Baradas MC Greenberg SB Atmar RL Musher DW Gathe JC JrVisnegarwala F Trautner BW Immune reconstitution inflammatory syndrome: emergence of a unique syndrome during highly active antiretroviral therapy Medicine (Baltimore) 2002 81 213 27 11997718 10.1097/00005792-200205000-00005
Narita M Ashkin D Hollender ES Pitchenik AE Paradoxical worsening of tuberculosis following antiretroviral therapy in patients with AIDS Am J Respir Crit Care Med 1998 158 157 61 9655723
Race EM Adelson-Mitty J Kriegel GR Barlam TF Reimann KA Letvin NL Japour AJ Focal mycobacterial lymphadenitis following initiation of protease-inhibitor therapy in patients with advanced HIV-1 disease Lancet 1998 351 252 5 9457095 10.1016/S0140-6736(97)04352-3
Cooney EL Clinical indicators of immune restoration following highly active antiretroviral therapy Clin Infect Dis 2002 34 224 33 11740712 10.1086/323898
Imami N Larche M Ritter MA Inhibition of alloreactivity by mAb MR6: differential effects on IL-2- and IL-4- producing human T cells Int Immunol 1994 6 1575 84 7826948
Imami N Hardy GA Nelson MR Morris-Jones S Al-Shahi R Antonopoulos C Gazzard B Gotch FM Induction of HIV-1-specific T cell responses by administration of cytokines in late-stage patients receiving highly active anti-retroviral therapy Clin Exp Immunol 1999 118 78 86 10540163 10.1046/j.1365-2249.1999.01012.x
Imami N Pires A Hardy G Wilson J Gazzard B Gotch F A balanced type 1/type 2 response is associated with long-term nonprogressive human immunodeficiency virus type 1 infection J Virol 2002 76 9011 23 12186885 10.1128/JVI.76.18.9011-9023.2002
Characterisation of HIV-1-specific CD4+ T helper epitopes
Lawn SD Bekker LG Miller RF Immune reconstituion disease associated with mycobacterial infections in HIV-infected individuals receiving antiretrovirals Lancet Infect Dis 2005 5 361 73 15919622 10.1016/S1473-3099(05)70140-7
Sei S Sandelli SL Theofan G Ratto-Kim S Kumagai M Loomis-Price LD Cox JH Jarosinski P Walsek CM Brouwers P Preliminary evaluation of human immunodeficiency virus type 1 (HIV-1) immunogen in children with HIV-1 infection J Infect Dis 1999 180 626 40 10438349 10.1086/314944
Schrier RD Wiley CA Spina C McCutchan JA Grant I Pathogenic and protective correlates of T cell proliferation in AIDS. HNRC Group. HIV Neurobehavioral Research Center J Clin Invest 1996 98 731 40 8698865
Li TS Tubiana R Katlama C Calvez V Ait Mohand H Autran B Long-lasting recovery in CD4 T-cell function and viral-load reduction after highly active antiretroviral therapy in advanced HIV-1 disease Lancet 1998 351 1682 6 9734884 10.1016/S0140-6736(97)10291-4
Hardy GA Imami N Sullivan AK Pires A Burton CT Nelson MR Gazzard B Gotch FM Reconstitution of CD4+ T cell responses in HIV-1 infected individuals initiating highly active antiretroviral therapy (HAART) is associated with renewed interleukin-2 production and responsiveness Clin Exp Immunol 2003 134 98 106 12974761 10.1046/j.1365-2249.2003.02256.x
Day CL Walker BD Progress in defining CD4 Helper Responses in chronic viral infections J Exp Med 2003 198 1773 77 14676292 10.1084/jem.20031947
Ridge JP Di Rosa F Matzinger P A conditioned dendritic cell can be a temporal bridge between a CD4+ T-helper and a T-killer cell Nature 1998 393 474 8 9624003 10.1038/30989
Janssen EM Lemmens EE Wolfe T Christen U Von Herrath MG Schoenberger SP CD4(+) T cells are required for secondary expansion and memory in CD8(+) T lymphocytes Nature 2003 421 852 6 12594515 10.1038/nature01441
Sun JC Bevan MJ Defective CD8 T cell memory following acute infection without CD4 T cell help Science 2003 300 339 42 12690202 10.1126/science.1083317
Shedlock DJ Shen H Requirement for CD4 T cell help in generating functional CD8 T cell memory Science 2003 300 337 9 12690201 10.1126/science.1082305
Shedlock DJ Whitmire JK Tan J MacDonald AS Ahmed R Shen H Role of CD4 T cell help and costimulation in CD8 T cell responses during Listeria monocytogenes infection J Immunol 2003 170 2053 63 12574376
Appay V Nixon DF Donahoe SM Gillespie GM Dong T King A Ogg GS Spiegel HM Conlon C Spina CA HIV-specific CD8(+) T cells produce antiviral cytokines but are impaired in cytolytic function J Exp Med 2000 192 63 75 10880527 10.1084/jem.192.1.63
Kalams SA Walker BD The critical need for CD4 help in maintaining effective cytotoxic T lymphocyte responses J Exp Med 1998 188 2199 204 9858506 10.1084/jem.188.12.2199
Wilson JD Imami N Watkins A Gill J Hay P Gazzard B Westby M Gotch FM Loss of CD4+ T cell proliferative ability but not loss of human immunodeficiency virus type 1 specificity equates with progression to disease J Infect Dis 2000 182 792 8 10950773 10.1086/315764
Gougeon ML Lecoeur H Dulioust A Enouf MG Crouvoiser M Goujard C Debord T Montagnier L Programmed cell death in peripheral lymphocytes from HIV-infected persons: increased susceptibility to apoptosis of CD4 and CD8 T cells correlates with lymphocyte activation and with disease progression J Immunol 1996 156 3509 20 8617980
Kumagai M Coustan-Smith E Murray DJ Silvennoinen O Murti KG Evans WE Malavasi F Campana D Ligation of CD38 suppresses human B lymphopoiesis J Exp Med 1995 181 1101 10 7869031 10.1084/jem.181.3.1101
Morra M Zubiaur M Terhorst C Sancho J Malavasi F CD38 is functionally dependent on the TCR/CD3 complex in human T cells Faseb J 1998 12 581 92 9576485
Todisco E Suzuki T Srivannaboon K Coustan-Smith E Raimondi SC Behm FG Kitanaka A Campana D CD38 ligation inhibits normal and leukemic myelopoiesis Blood 2000 95 535 42 10627459
Caruso A Licenziati S Corulli M Canaris AD De Francesco MA Fiorentini S Peroni L Fallacara F Dima F Balsari A Flow cytometric analysis of activation markers on stimulated T cells and their correlation with cell proliferation Cytometry 1997 27 71 6 9000587 10.1002/(SICI)1097-0320(19970101)27:1<71::AID-CYTO9>3.0.CO;2-O
Mohri H Perelson AS Tung K Ribeiro RM Ramratnam B Markowitz M Kost R Hurley A Weinberger L Cesar D Increased turnover of T lymphocytes in HIV-1 infection and its reduction by antiretroviral therapy J Exp Med 2001 194 1277 87 11696593 10.1084/jem.194.9.1277
Lombardi G Hargreaves R Sidhu S Imami N Lightstone L Fuller-Espie S Ritter M Robinson P Tarnok A Lechler R Antigen presentation by T cells inhibits IL-2 production and induces IL-4 release due to altered cognate signals J Immunol 1996 156 2769 75 8609395
Pichler WJ Wyss-Coray T T cells as antigen-presenting cells Immunol Today 1994 15 312 5 7522009 10.1016/0167-5699(94)90078-7
Burgess AW Metcalf D The nature and action of granulocyte-macrophage colony stimulating factors Blood 1980 56 947 58 7002232
Burgess AW Metcalf D Watt SM Regulation of hemopoietic cell differentiation and proliferation J Supramol Struct 1978 8 489 500 309973 10.1002/jss.400080411
Metcalf D Johnson GR Burgess AW Direct stimulation by purified GM-CSF of the proliferation of multipotential and erythroid precursor cells Blood 1980 55 138 47 6965346
|
16181494
|
PMC1262752
|
CC BY
|
2021-01-04 16:37:45
|
no
|
J Immune Based Ther Vaccines. 2005 Sep 25; 3:7
|
utf-8
|
J Immune Based Ther Vaccines
| 2,005 |
10.1186/1476-8518-3-7
|
oa_comm
|
==== Front
J Inflamm (Lond)Journal of Inflammation (London, England)1476-9255BioMed Central London 1476-9255-2-101620738010.1186/1476-9255-2-10ResearchRegulation of IκBα expression involves both NF-κB and the MAP kinase signaling pathways Zhang Ning [email protected] Muhammad H [email protected] Lingyun [email protected] Lidia C [email protected] Anthony F [email protected] David B [email protected] Xenogen Corporation, 860 Atlantic Avenue, Alameda, California 94501, USA2005 5 10 2005 2 10 10 24 3 2005 5 10 2005 Copyright © 2005 Zhang et al; licensee BioMed Central Ltd.2005Zhang et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
IκBα is an inhibitor of the nuclear transcription factor NF-κB. Binding of IκBα to NF-κB inactivates the transcriptional activity of NF-κB. Expression of IκBα itself is regulated by NF-κB, which provides auto-regulation of this signaling pathway. Here we present a mouse model for monitoring in vivo IκBα expression by imaging IκBα-luc transgenic mice for IκBα promoter driven luciferase activity. We demonstrated a rapid and systemic induction of IκBα expression in the transgenic mice following treatment with LPS. The induction was high in liver, spleen, lung and intestine and lower in the kidney, heart and brain. The luciferase induction in the liver correlated with increased IκBα mRNA level. Pre-treatment with proteasome inhibitor bortezomib dramatically suppressed LPS-induced luciferase activity. The p38 kinase inhibitor SB203580 also showed moderate inhibition of LPS-induced luciferase activity. Analysis of IκBα mRNA in the liver tissue showed a surprising increase of the IκBα mRNA after bortezomib and SB203580 treatments, which could be due to increased IκBα mRNA stability. Our data demonstrate that regulation of IκBα expression involves both the NF-κB and the p38 signaling pathways. The IκBα-luc transgenic mice are useful for analyzing IκBα expression and the NF-κB transcriptional activity in vivo.
IkappaBNF-κBMAP kinasebortezomiblipopolysaccharidebioluminescent imaging
==== Body
Introduction
IκBα is an inhibitor of nuclear transcription factor NF-κB, which regulates the expression of proinflammatory and cytotoxic genes [1]. In nonstimulated cells NF-κB proteins are present in the cytoplasm in association with specific inhibitors IκBα, IκBβ and IκBγ. Stimulation by extra-cellular inducers results in the phosphorylation and degradation of IκB through a ubiquitin-proteasome pathway, allowing NF-κB to translocate into the nucleus to activate the transcription of target genes [2,3]. The IκBα gene contains functional NF-κB sites in the promoter region. Transcriptional activation of IκBα expression by NF-κB leads to rapid re-synthesis of IκBα protein and blockade of NF-κB nuclear translocation [4,5]. This auto-regulatory loop is both sensitive to and rapidly influenced by NF-κB activating stimuli [6]. In addition, phosphorylation of IκB kinase and the activation of NF-κB also involve the MAP kinase signaling pathways [7].
In this paper we describe and characterize an IκBα-luc transgenic mouse that was used for monitoring IκBα expression through bioluminescent imaging. We tested the effect of bortezomib and several MAP kinase inhibitors on LPS-induced IκBα expression. The results that follow suggest that, in addition to NF-κB, the MAP kinase signaling pathway is involved in controlling IκBα expression.
Materials and methods
Construction of pIκBα-luc vector and generation of IκBα-luc transgenic mice
A mouse BAC clone containing the mouse IκBα gene was isolated from a CT7 mouse BAC library (Invitrogen, Carlsbad, CA). A 11.0 kb promoter fragment containing sequences 5' to the first ATG for the mouse IκBα gene was obtained by the RED cloning method [8] and cloned upstream of the firefly luciferase gene in the pGL3-Basic vector (Promega, Madison, WI). A 0.8 kb human β-globin intron 2 was placed between the IκBα promoter and the luciferase gene to optimize the luciferase expression in transgenic mice. The transgene cassette was separated from the vector backbone sequences and used for pronuclear injection into Balb/C mouse strain embryos. These steps yielded the transgenic model henceforth designated Balb/C-Tg(IκBα-luc)Xen and abbreviated in the text as IκBα-luc.
Reagents
We purchased bacterial lipopolysaccharide (LPS, from Salmonella abortus equi), PD098580 from Sigma-Aldrich Chemical Co., (St. Louis, MO), Bortezomib (VALCADE, PS-341) from Millennium Pharmaceuticals, Inc. (Cambridge, MA), SB203580 from EMD Biosciences, Inc. (La Jolla, CA) and SP600125 from A.G. Scientific, Inc. (San Diego, CA).
In vivo imaging of luciferase activity
In vivo imaging was performed using an IVIS® Imaging System 100 Series (Xenogen Corp., Alameda, CA). IκBα-luc transgenic mice were anesthetized with isoflurane and injected intraperitoneally with 150 mg/kg of luciferin (Biosynth, A.G., Switzerland). Ten minutes after the luciferin injection, mice were imaged for 1–10 seconds. Photons emitted from specific regions were quantified using Living Image® software (Xenogen Corp.). In vivo luciferase activity is expressed as photons/second/cm2.
Study of in vivo IκBα gene regulation using IκBα-luc transgenic mice
IκBα-luc transgenic mice of 3–6 months of age were injected with LPS (1 mg/kg, i.p.). Control mice were injected with saline. At selected time points, mice were imaged for the luciferase signal. To test the effect of various compounds, mice were pre-treated with bortezomib (1 mg/kg, i.v.), PD098059 (10 mg/kg, i.v.), SP600125 (20 mg/kg, i.v.), or SB203580 (5 mg/kg, i.v.) 1 hour prior to the LPS injection.
Tissue luciferase activity
Selected organs were removed and homogenized in 3 volumes of PBS containing a protease inhibitor cocktail (Roche Applied Science, Indianapolis, IN) and lysed with passive lysis buffer (Promega). After centrifugation at 14,000-rpm for 10 min at 4°C, the supernatant was collected. Luciferase activity was assayed using the Luciferase Assay System (Promega) and a Turner Design, TD 20/20, Luminometer (Sunnyvale, CA). Protein concentration was estimated with Bradford reagent (Sigma-Aldrich).
Northern blot analysis
Total RNA was isolated from mouse tissue using RNAwiz (Ambion, Austin, TX) and further purified using the RNAeasy kit (Qiagen Inc., Valencia, CA). A total of 2 μg of RNA sample was analyzed by Northern blot using a NorthernMax system (Ambion). A 482 nt IκBα cDNA fragment was amplified (forward primer: 5'- GCTCTAGAGCAATCATCCACGAAGAGAAGC-3'; reverse primer: 5'- CGGAATTCGCCCCACATTTCAACAAGAGC-3') and cloned into the pBlueScript SK vector (Stratagene, La Jolla, CA) that was linearized with XbaI and EcoRI. Single strand antisense IκBα RNA probe was prepared by transcription with T7 polymerase using a Strip-EZ kit (Ambion). After hybridization, the signal was detected using a BrightStar BioDetect kit (Ambion)
Statistics
Nonparametric tests for significance were used to test whether changes in luciferase signal from baseline were significantly greater than zero within groups (sign test) and whether the changes from baseline were significantly different between treatment groups (Mann-Whitney test). Values are presented as means ± one standard error in the graphs and text unless otherwise noted. For some statistical tests genders were combined to increase sample number in each group. All significance levels are two-sided.
Results
Induction of IκBα expression by LPS
We generated IκBα-luc transgenic mice and screened for their response to LPS treatment through bioluminescent imaging of luciferase activity. Transgenic mice from all founder lines showed inducible luciferase expression after LPS treatment. One transgenic line was selected for this study. In untreated IκBα-luc mice, basal luciferase signal was detected throughout the entire body. Male and female mice showed similar levels of basal luciferase signal. After LPS treatment, an induction of luciferase signal was observed at 2 hours after treatment. The signal remained highly induced at 4 hours and started to decline at 7 hours. By 24 hours, the signal declined to near baseline levels (Figure 1A). Anatomically, the induction was higher in hepatic and intestinal regions of the abdomen than that in other parts of the body.
Figure 1 Imaging analysis of luciferase expression in IκBα-luc transgenic mice treated with LPS. A. IκBα-luc transgenic mice were imaged at T = 0, 2, 4, 7 and 24 hours after treatment with LPS (1 mg/kg, i.p., n = 4 for males, n = 6 for females). Representative mice from each treatment group are shown. The color overlay on the image represents the photons/second emitted from the mouse body in accord with the pseudo-color scale shown on the right of the images. Red represents the highest photons/sec while blue represents the lowest photons/sec. B. Quantification of the luciferase signal from the abdominal region of the body. Data are means luciferase activity (billion photon/second) ± SE. Statistical analysis was done for male and female combined data. * indicates a significant induction of luciferase signal by LPS (P = 0.002). C. Northern blot analysis of IκBα mRNA in the liver tissue. Liver tissue was harvested from saline (control) or LPS treated IκBα-luc female mice at 4 hours after treatment and processed for RNA isolation. A total of 2 μg of RNA was analyzed by Northern blot. Equal loading was demonstrated by 28S rRNA.
Luciferase signals from the abdominal region of LPS-treated mice were quantified using the Living Image® software to produce the data shown in Figure 1B. At the peak of induction 2 to 4 hours after injection, the luciferase signals were increased 6 to 10-fold by LPS as compared with basal luciferase signal at T = 0 hour. At 24 hours, the luciferase signal was still 2 to 3-fold greater than basal levels.
IκBα expression is induced in multiple tissues after LPS treatment
Table 1 displays the luciferase activity in selected organs in IκBα-luc mice. In untreated mice, ex vivo luciferase activity was detected in all the dissected organs of both sexes. The pattern of luciferase expression of the male tissues was similar to that of the female tissues. The luciferase activity was the highest in liver, spleen and lung, lowest in heart, and intermediate in intestine, kidney and brain. In LPS treated mice, all the examined organs showed a significant induction of the luciferase activity. Liver, spleen, lung and intestine showed dramatically higher luciferase expression than that in kidney, heart and brain. As calculated from the mean of the control mice, LPS treatment caused 19-to 23-fold luciferase induction in the liver, 19- to 28-fold in the spleen, 8-fold in the lung, 19- to 52-fold in the intestine, 6-to 11-fold in the kidney, 54- to 63-fold in the heart, 5- to 7-fold in the brain.
Table 1 Ex vivo measurement of luciferase activity (Unit/μg protein). Selected organs were harvested from IkBα-luc mice that were untreated (control, n = 3) or treated with LPS (1 mg/kg, i.p., n = 3) at 4 hours prior to the harvesting.
Mean ± SE
MALE FEMALE
ORGANS Control LPS Control LPS
Liver 157 ± 30 3651 ± 48*§ 157 ± 2 2933 ± 69*
Spleen 363 ± 69 6906 ± 878* 218 ± 58 6203 ± 1414*
Lungs 430 ± 112 3549 ± 291* 348 ± 52 2718 ± 452*
Intestine 89 ± 39 4640 ± 601* 73 ± 9 1367 ± 598*
Kidney 65 ± 7 709 ± 62* 67 ± 5 414 ± 26*
Heart 15 ± 2§ 951 ± 141* 7 ± 2 405 ± 8*
Brain 72 ± 13 513 ± 84* 73 ± 9 384 ± 52*
*Difference from controls significant at P ≤ 0.05 by Mann-Whitney nonparametric test.
We further attempted to establish a correlation between luciferase activity and IκBα mRNA expression. In the liver tissue of un-treated mice, IκBα mRNA expression was detectable. Following LPS treatment, an induction of IκBα mRNA expression was observed (Figure 1C), which correlated with the increase of luciferase activity in the liver.
Bortezomib inhibited LPS-induced IκBα expression
Using the IκBα-luc model, we tested the effect of bortezomib on LPS-induced IκBα expression in vivo. As shown in Figure 2A, pre-treatment of the IκBα-luc mice with bortezomib significantly inhibited LPS-induced luciferase expression in the whole body, especially in liver and intestine where the luciferase signal was highly induced. Quantification of the luciferase signal showed that inhibition of luciferase activity by bortezomib was significant at all the time points in both male and female mice (Figure 2B, C). At the peak of induction at 2–4 hours, bortezomib inhibited 70–80% of LPS-induced luciferase activity in the abdominal region.
Figure 2 Effect of bortezomib on LPS-induced luciferase expression. A. IκBα-luc transgenic mice were pre-treated with bortezomib (1 mg/kg, i.v. n = 5) at 1 hour prior to the LPS treatment. The positive control mice (n = 4 for males, n = 6 for females) were pre-injected with saline. All the mice were imaged at T = 0, 2, 4, 7 and 24 hours after the LPS treatment. B, C. Quantification of the luciferase signal from the abdominal region of the body for male and female mice respectively. Data are expressed as billion photons/second. Nonparametric significance levels for the difference between treatment groups were determined by a Mann-Whitney test and are presented above the bars.
Bortezomib inhibited LPS-induced IκBα expression in all the organs except the brain
We examined the effect of bortezomib on LPS-induced IκBα expression in selected organs (Figure 3A, B). In comparison to the LPS-treated mice, mice pre-treated with bortezomib showed significant inhibition of luciferase induction in all organs examined except the brain. The inhibition ranges from 50% to 80% in examined tissues excluding the brain.
Figure 3 Effect of bortezomib pre-treatment on the LPS-induced luciferase activity in selected tissues in IκBα-luc male (A) and female (B) mice (n = 3 for both genders). Mice were injected with bortezomib (1 mg/kg, i.v.) 1 hour prior to the LPS treatment (1 mg/kg, i.p.). Mice treated with LPS alone were used as positive controls. Organs were harvested from all the mice at 3 hours after the LPS injection and processed for luciferase activity.* indicates a significant reduction in signal by bortezomib (P = 0.05). C. Northern blot analysis of IκBα mRNA in the liver tissue. IκBα-luc transgenic mice were sacrificed at 3 hours after LPS injection. Liver tissue was harvested and processed for RNA isolation. A total of 2 μg of RNA was analyzed by Northern blot. Equal loading was demonstrated by 28S rRNA.
We further examined the effect of bortezomib on IκBα mRNA induction by LPS. In both male and female mice, pre-treatment with bortezomib increased LPS-induced IκBα mRNA level in the liver tissue (Figure 3C).
Effect of the MAP kinase inhibitors on IκBα induction by LPS
We examined the effect of MAP kinase inhibitors SB203580, PD098059 and SP600125 on LPS-induced IκBα expression. The bioluminescent images and the quantification are presented in Figure 4A and 4B respectively. Pre-treatment of the IκBα-luc mice with SB203580 moderately inhibited LPS-induced luciferase expression. PD098059 pre-treated mice also had lower luciferase activity as compared to the LPS-treated positive control mice. However, the difference was significant at 7 hours only (Figure 4B). SP600125 failed to affect LPS-induced luciferase expression.
Figure 4 Effect of MAP kinase inhibitors on LPS-induced luciferase expression. A. Female IκBα-luc transgenic mice were pre-treated with SB203580 (5 mg/kg, i.v., n = 5), PD098059 (10 mg/kg, i.v., n = 5), or SP600125 (20 mg/kg, i.v., n = 8) at 1 hour prior to the LPS treatment. The positive control mice were pre-injected with DMSO (n = 8). All the mice were imaged at T = 0, 2, 4, 7 and 24 hours after LPS treatment. Representative mice are shown for each group. B. Quantification of the luciferase signal from liver region and the data were expressed as photons/second/cm2.
We further analyzed the luciferase activity in selected organs harvested from SB203580-pre-treated mice at 3 hours after the LPS injection. As shown in Figure 5A, SB203580 significantly inhibited LPS-induced luciferase activity in liver, lung, and intestine, but not in the spleen, brain, kidney or heart.
Figure 5 Ex vivo measurement of the effect of SB203580 on LPS-induced luciferase expression. A. Selected organs were harvested from SB203580 pre-treated mice and LPS treated control mice at 4 hours after the LPS injection. * indicates a significant difference between vehicle (DMSO) + LPS and SB203580 + LPS (p = 0.05; sign test). B. Northern blot analysis of IκBα mRNA in the liver tissue. IκBα-luc transgenic mice were sacrificed at 3 hours after LPS injection. Liver tissue was harvested and processed for RNA isolation. A total of 2 μg of RNA was analyzed by Northern blot. Equal loading was demonstrated by 28S rRNA.
The effect of SB203580 on IκBα mRNA induction by LPS is shown in Figure 5B. Pre-treatment with SB203580 increased LPS-induced IκBα mRNA level in the liver tissue of the IκBα-luc mice.
Discussion
The mouse IκBα promoter contains 6 putative NF-κB binding sites that mediate the NF-κB regulation [9]. Induction of IκBα-luc expression in the early stage of the LPS response is consistent with a tight auto-regulation of the NF-κB signaling pathway by IκBα [6]. By reflecting NF-κB transcriptional activity, the luciferase signal in the IκBα-luc mouse provides a convenient approach for in vivo monitoring of NF-κB activation.
It has been shown previously that LPS treatment causes degradation of IκBα protein within 40 minutes, followed by induction of IκBα mRNA that results in rapid recovery of the IκBα protein by 3 hours. As a result, maximal NF-κB activation occurred 1 hour after LPS treatment but started to decline at 3–6 hours post treatment [10]. In agreement, our in vivo imaging data demonstrated an induction of luciferase activity at 2 to 4 hours after treating the IκBα-luc mice with LPS, followed by decline of the luciferase activity at 7 and 24 hours. In addition, we also observed a slight gender difference of the kinetics of NF-κB activation following LPS treatment. Male mice showed a peak of induction at 4 hours, followed by a sharp decrease at 7 hours. Female mice showed a peak of induction at 2 hours, followed by a sequential decrease at 7 and 24 hours. This indicates that LPS-induced inflammation process may be sustained longer in female mice than in male mice.
Ex vivo analysis of selected tissues of IκBα-luc mice showed baseline luciferase expression in liver, spleen and lung, with lower expression in intestine, kidney, heart and brain. Significant induction of luciferase expression was observed in all of these organs in both male and female mice after LPS treatment, with higher luciferase activity observed in liver, spleen and intestine as compared to other tissues (Table 1). This is consistent with the bioluminescent imaging analysis of luciferase activity in the live mice that shows higher luciferase signals were present in both hepatic and intestinal regions than other parts of the body (Figure 1A). High extent of luciferase induction in the liver, spleen, lung and intestine by LPS is consistent with IκBα degradation and NFκB activation in these organs in response to endotoxemia [11-13]. When male and female mice are compared, the luciferase signal in intestine was significantly higher in the LPS-treated male mice as compared with the female mice. The difference could be due to the difference of the kinetics of luciferase induction between male and female mice or simply due to a relatively small sample number used for this study.
Bortezomib inhibited LPS-induced luciferase activity by 70–80% in the IκBα-luc mice, which is confirmed by a broad suppression of luciferase activity in all the analyzed tissues except the brain. Bortezomib is an inhibitor of proteasome activity that is required for IκB degradation and subsequent nuclear translocation of NF-κB [14]. In addition, bortezomib can also inhibit other cell signaling pathways, such as mitogen-activated protein kinase growth signaling, causing inhibition of cell proliferation and induction of cell apoptosis [15,16]. Analysis of the IκBα mRNA showed that bortezomib pre-treatment caused a further increase of LPS-induced IκBα mRNA in the liver. Since the transcriptional activity of the IκBα promoter was suppressed bortezomib, we suspect that the increase of IκBα mRNA after bortezomib treatment should be due to an increase of IκBα mRNA stability. These data suggest that inhibition of NF-κB mediated inflammation by bortezomib may be due to a broad range of effects, affecting processes such as IκB protein degradation and IκBα mRNA stability.
Several MAP kinase inhibitors were tested for their effect on LPS-induced NF-κB activation. We demonstrated that pre-treatment with p38 MAP kinase inhibitor SB203580 at a dose of 5 mg/kg partially inhibited LPS-induced luciferase expression in the IκBα-luc mice in liver, lung and intestine. It has been reported that SB203580 inhibits inflammatory cytokine production in vivo in both mice and rat with IC50 value of 15 to 25 mg/kg [17]. In another report, it was shown that SB203580 at 5, 10 and 20 mg/kg produced a dose dependent inhibition on TNF-alpha production in vivo [18]. Therefore, it is likely that the SB203580 dose used in our study had an inhibitory effect on p38 MAP kinase activation. We also showed that the ERK MAP kinase inhibitor PD098059 at 10 mg/kg partially inhibited LPS-induced luciferase expression at 7 hours. At this dose, PD098059 was able to suppress ERK1/2 phosphorylation in vivo [19]. We further showed that JNK kinase inhibitor SP600125 at 20 mg/kg had no effect on LPS-induced luciferase expression. At this dose, SAPK/JNK MAP kinase phosphorylation can be totally inhibited in the liver tissue [20].
In summary, we have produced a transgenic mouse in which luciferase expression is driven by the IκBα promoter. We observed a ubiquitous expression and induction of IκBα in the IκBα-luc transgenic mice by LPS. We demonstrated involvement of both the NF-κB and the p38 MAP kinase signaling pathways in the induction of IκBα expression by LPS.
Clinically, NF-κB activation is involved in many chronic disease conditions, such as rheumatoid arthritis, atheroscleorosis, asthma and tumor development [21,22]. The luciferase activity in the IκBα-luc mice could be used as a sensor for monitoring the NF-κB activation and to further understand how NF-κB activation contributes to the initiation and progression of these disease conditions. In addition, IκBα-luc mice could also be used for testing or even screening of novel NF-κB inhibitors for therapeutic potential.
Acknowledgements
We thank Paul T. Williams for consulting on the statistical analyses of the data.
==== Refs
Karin M Delhase M JNK or IKK, AP-1 or NF-kappaB, which are the targets for MEK kinase 1 action? Proc Natl Acad Sci 1998 95 9067 9 9689033 10.1073/pnas.95.16.9067
Baldwin AS Jr The NF-kappa B and I kappa B proteins: new discoveries and insights Annu Rev Immunol 1996 14 649 83 8717528 10.1146/annurev.immunol.14.1.649
Li Q Verma IM NF-kappaB regulation in the immune system Nat Rev Immunol 2002 2 975 10.1038/nri910
Sun SC Ganchi PA Ballard DW Greene WC NF-kappa B controls expression of inhibitor I kappa B alpha: evidence for an inducible autoregulatory pathway Science 1993 259 1912 5 8096091
Arenzana-Seisdedos F Thompson J Rodriguez MS Bachelerie F Thomas D Hay RT Inducible nuclear expression of newly synthesized I kappa B alpha negatively regulates DNA-binding and transcriptional activities of NF-kappa B Mol Cell Biol 1995 15 2689 2696 7739549
Pando MP Verma IM Signal-dependent and -independent degradation of free and NF-kappa B-bound IkappaBalpha J Biol Chem 2000 275 21278 86 10801847 10.1074/jbc.M002532200
Dong C Davis RJ Flavell RA MAP kinases in the immune response Annu Rev Immunol 2002 20 55 72 11861597 10.1146/annurev.immunol.20.091301.131133
Lee EC Yu D Martinez de Velasco J Tessarollo L Swing DA Court DL Jenkins NA Copeland NG A highly efficient Escherichia coli-based chromosome engineering system adapted for recombinogenic targeting and subcloning of BAC DNA Genomics 2001 73 56 65 11352566 10.1006/geno.2000.6451
Rupec RA Poujol D Grosgeorge J Carle GF Livolsi A Peyron JF Schmid RM Baeuerle PA Messer G Structural analysis, expression, and chromosomal localization of the mouse IκBα gene Immunogenetics 1999 49 395 403 10199915 10.1007/s002510050512
Velasco M Diaz-Guerra MJ Martin-Sanz P Alvarez A Bosca L Rapid up-regulation of IB-β and abrogation of NF-B activity in peritoneal macrophages stimulated with lipopolysaccharide J Biol Chem 1997 272 23025 23030 9287299 10.1074/jbc.272.37.23025
Pritts TA Moon R Fischer JE Salzman AL Hasselgren PO Nuclear factor-kappaB is activated in intestinal mucosa during endotoxemia Arch Surg 1998 133 1311 5 9865648 10.1001/archsurg.133.12.1311
Szabo G Romics L JrFrendl G Liver in sepsis and systemic inflammatory response syndrome Clin Liver 2002 6 1045 66 10.1016/S1089-3261(02)00058-2
Aldridge AJ Role of the neutrophil in septic shock and the adult respiratory distress syndrome Eur J Surg 2002 168 204 14 12440757 10.1080/11024150260102807
Lightcap ES McCormack TA Pien CS Chau V Adams J Elliott PJ Proteasome inhibition measurements: clinical application Clin Chem 2000 46 673 683 10794750
Hideshima T Richardson P Chauhan D Palombella VJ Elliott PJ Adams J Anderson KC The proteasome inhibitor PS-341 inhibits growth, induces apoptosis, and overcomes drug resistance in human multiple myeloma cells Cancer Res 2001 61 3071 6 11306489
Sunwoo JB Chen Z Dong G Yeh N Crowl Bancroft C Sausville E Adams J Elliott P Van Waes C Novel proteasome inhibitor PS-341 inhibits activation of nuclear factor-kappa B, cell survival, tumor growth, and angiogenesis in squamous cell carcinoma Clin Cancer Res 2001 7 1419 28 11350913
Badger AM Bradbeer JN Votta B Lee JC Adams JL Griswold DE Pharmacological profile of SB 20 a selective inhibitor of cytokine suppressive binding protein/p38 kinase, in animal models of arthritis, bone resorption, endotoxin shock and immune function J Pharmacol Exp Ther 3580 279 1453 61 8968371
Slomiany BL Slomiany A Delay in oral mucosal ulcer healing by aspirin is linked to the disturbances in p38 mitogen-activated protein kinase activation J Physiol Pharmacol 2001 52 185 94 11453099
Clemons AP Holstein DM Galli A Saunders C Cerulein-induced acute pancreatitis in the rat is significantly ameliorated by treatment with MEK1/2 inhibitors U0126 and PD98059 Pancreas 2002 25 251 9 12370536 10.1097/00006676-200210000-00007
Zhang N Ahsan MH Zhu L Sambucetti LC Purchio AF West DB NF-kappaB and not the MAPK signaling pathway regulates GADD45beta expression during acute inflammation J Biol Chem 2005 280 21400 8 15797874 10.1074/jbc.M411952200
Barnes PJ Karin M Nuclear factor-kappaB: a pivotal transcription factor in chronic inflammatory diseases N Engl J Med 1997 336 1066 71 9091804 10.1056/NEJM199704103361506
Yamamoto Y Gaynor RB Therapeutic potential of inhibition of the NF-kappaB pathway in the treatment of inflammation and cancer J Clin Invest 2001 107 135 42 11160126
|
16207380
|
PMC1262753
|
CC BY
|
2021-01-04 16:36:23
|
no
|
J Inflamm (Lond). 2005 Oct 5; 2:10
|
utf-8
|
J Inflamm (Lond)
| 2,005 |
10.1186/1476-9255-2-10
|
oa_comm
|
==== Front
J NeuroinflammationJournal of Neuroinflammation1742-2094BioMed Central London 1742-2094-2-211617429410.1186/1742-2094-2-21Research5-aminoimidazole-4-carboxamide-1-beta-4-ribofuranoside (AICAR) attenuates the expression of LPS- and Aβ peptide-induced inflammatory mediators in astroglia Ayasolla Kamesh R [email protected] Shailendra [email protected] Avtar K [email protected] Inderjit [email protected] Department of Pediatrics, Medical University of South Carolina, Charleston, South Carolina, 29425, USA2 Department of Pathology, Medical University of South Carolina, Charleston, South Carolina, 29425, USA3 Department of Obstetrics & Gynaecology, Medical University of South Carolina, Charleston, South Carolina, 29425, USA4 Department of Pathology, Ralph H. Johnson VA Medical Center, Charleston, South Carolina 29425, USA2005 20 9 2005 2 21 21 21 7 2005 20 9 2005 Copyright © 2005 Ayasolla et al; licensee BioMed Central Ltd.2005Ayasolla et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Alzheimer's disease (AD) pathology shows characteristic 'plaques' rich in amyloid beta (Aβ) peptide deposits. Inflammatory process-related proteins such as pro-inflammatory cytokines have been detected in AD brain suggesting that an inflammatory immune reaction also plays a role in the pathogenesis of AD. Glial cells in culture respond to LPS and Aβ stimuli by upregulating the expression of cytokines TNF-α, IL-1β, and IL-6, and also the expression of proinflammatory genes iNOS and COX-2. We have earlier reported that LPS/Aβ stimulation-induced ceramide and ROS generation leads to iNOS expression and nitric oxide production in glial cells. The present study was undertaken to investigate the neuroprotective function of AICAR (a potent activator of AMP-activated protein kinase) in blocking the pro-oxidant/proinflammatory responses induced in primary glial cultures treated with LPS and Aβ peptide.
Methods
To test the anti-inflammatory/anti-oxidant functions of AICAR, we tested its inhibitory potential in blocking the expression of pro-inflammatory cytokines and iNOS, expression of COX-2, generation of ROS, and associated signaling following treatment of glial cells with LPS and Aβ peptide. We also investigated the neuroprotective effects of AICAR against the effects of cytokines and inflammatory mediators (released by the glia), in blocking neurite outgrowth inhibition, and in nerve growth factor-(NGF) induced neurite extension by PC-12 cells.
Results
AICAR blocked LPS/Aβ-induced inflammatory processes by blocking the expression of proinflammatory cytokine, iNOS, COX-2 and MnSOD genes, and by inhibition of ROS generation and depletion of glutathione in astroglial cells. AICAR also inhibited down-stream signaling leading to the regulation of transcriptional factors such as NFκB and C/EBP which are critical for the expression of iNOS, COX-2, MnSOD and cytokines (TNF-α/IL-1β and IL-6). AICAR promoted NGF-induced neurite growth and reduced neurite outgrowth inhibition in PC-12 cells treated with astroglial conditioned medium.
Conclusion
The observed anti-inflammatory/anti-oxidant and neuroprotective functions of AICAR suggest it as a viable candidate for use in treatment of Alzheimer's disease.
==== Body
Background
Alzheimer's disease (AD) is a neurological disorder and the brain pathology is characterized by the presence of senile plaques rich in insoluble aggregates of beta amyloid (1–40) and (1–42) peptides, degradation products of the larger amyloid precursor protein (APP) [1,2]. All major pro-inflammatory cytokines with the exception of IFN-γ (TNF-α, IL-1 and IL-6) have been detected in AD brain suggesting that an inflammatory immune reaction also plays a role in the pathogenesis of AD [3,4]. The deposited Aβ peptides have also been implicated in oxidative stress-induced responses, via NADPH oxidase activation and superoxide anion generation [5].
The astroglial population has a major role in neuroinflammatory disease processes, and has been implicated in various neurological disorders including AD [6]. Though we still do not know what endogenous ligands may trigger an inflammatory response in AD, several studies have reported that LPS/Aβ treatment of glia serves as a good cell culture model for mimicking the inflammatory conditions in AD [7-10]. In vitro treatment of glial cells with LPS/Aβ peptides induces cytokines (TNF-α, IL-1β), and also leads to the release of NO by induction of iNOS, as a function of innate immune response (for a detailed review see [6,11]). COX-2, an enzyme in the PLA-2 cascade, involved in the arachidonic acid metabolic pathways for the synthesis of prostaglandins, is yet another enzyme that is expressed along with other inflammatory mediators in these glial cells [6]. Its expression has been observed to be coincident with the onset of expression of apoptotic neuronal cell death markers, due to excitotoxic neurotoxicity. The expression of iNOS leading to production of nitric oxide and, as a result, generation of peroxynitrite (a reaction product of the superoxide anion and nitric oxide) under oxidative stress conditions has also been implicated in the extensive neuronal damage of several neurological disorders including AD [12,13]. Therefore the mechanisms of pro-inflammatory cytokine-mediated oxidative stress (or vice versa) may be the potential target(s) for AD therapeutics.
AMP-activated protein kinase [14] (AMPK) is a member of the family of serine/threonine kinases and is activated by cellular increases in AMP concentrations under conditions of nutritional/metabolic stress [15].
This is thus often referred to as the fuel gauge of the cell, since it protects the cell against ATP depletion and boosts the energy generation pathways [16,17]. AMPK is activated by AMP-dependent phosphorylation by an upstream kinase, i.e. AMPK kinase (AMPKK; recently recognized as LKB1 [16,17]. AICAR is also reported to activate AMPK in the cell following its conversion to ZMP (a non-degradable AMP analog) and thus mimics the activity of AMP for activation of AMPK [18]. Recently, we reported anti-inflammatory properties of AICAR through activation of AMPK [19] in glial cells. AICAR was found to inhibit expression of pro-inflammatory cytokines and of iNOS in glial cells and in macrophages in cell culture as well as in rats treated with a sublethal dose of LPS [19] by attenuating NFκB and C/EBP pathways.
Aβ peptides are known to alter cellular redox, thereby triggering down stream kinase cascades leading to inflammation [12,20]. Hence this study was designed to evaluate the anti-oxidant/anti-inflammatory functions of AICAR in blocking LPS/Aβ-mediated down-stream signaling cascades leading to transcription factor activation and inflammatory cytokine release and iNOS and COX-2 expression. This study describes AICAR-mediated activation of AMPK and downregulation of LPS/Aβ-induced expression of inflammatory mediators in astrocyte-enriched glial cell cultures, possibly via reduction/regulation of cellular redox.
Methods
Reagents
DMEM and fetal bovine serum were obtained from Life Technologies Inc., Gaithersburg MD, USA, and LPS (Escherichia coli) from Calbiochem. Antibodies against iNOS and MnSOD were obtained from Transduction Labs, and antibody to COX-2 was from Cayman chemicals, Ann Arbor, MI. β-Actin and β-amyloid peptide (25–35) fragment as well as the reverse peptide (35–25), β-αmyloid peptides (1–40) and (1–42) were from Sigma. Antibodies for p65; p50; IB kinase (KKK); CCAAT/enhancer-binding proteins (C/EBP)-α, -β, and -δ; and oligonucleotides for NF-κB and C/EBP were from Santa Cruz. Recombinant tumor necrosis factor (TNF-α) and interleukin (IL)-1; and ELISA kits for TNF-α, IL-1, IL-6, and IFN-γ were from R & D Systems. Trizol and transfection reagents (Lipofectamine-2000, Lipofectamine-Plus, and Oligofectamine) were from Invitrogen. Chloramphenicol acetyltransferase ELISA, -galactosidase (-gal), 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT), and lactic dehydrogenase (LDH) kits were obtained from Roche. The enhanced chemiluminescence-detecting reagents were purchased from Amersham Biosciences. The luciferase assay system was from Promega. Antibodies against phosphospecific as well as nonphospho-p42/44, and -AMPK were from Cell Signaling Technology. NF-κB-luciferase was provided by Dr. Hanfang Zhang (Medical College of Georgia, Augusta, GA).
Cell culture and treatment of rat primary glial cultures and astrocytes
Astroglial cells were isolated from rat cerebral tissue as described by McCarthy and DeVellis [21]. Astrocytes were isolated and maintained as described earlier [12]. Cells were maintained in DMEM containing 10% fetal bovine serum. Glial cells were stimulated with either LPS (125 μg/ml), cytokines, or with sphingomyelinase (SMase) with or without β-amyloid peptide in serum-free DMEM and were harvested after 18 h unless stated otherwise. AICAR (1 mM), NAC (15 mM), Vitamin E (20 μM), or other substances were added 4 hr prior to stimulation with LPS/cytokines and were again added at the time of addition of stress stimuli.
Preparation of aged Aβ (1–40), (1–42) and (25–35) and induction of cells with β-amyloid peptide
The Aβ peptides (25–35), (1–40), (1–42) and the reverse peptide Aβ (40–1) were all purchased from Sigma. They were solubilized in phosphate-buffered saline (PBS) at a concentration of 1 mM, incubated in a capped vial at 37°C for 4 days [22], and stored frozen at -20°C until use. They were used at a final concentration of 7.5 μM or in higher amounts, as indicated.
Assay for NO synthesis
Synthesis of NO was determined by assaying culture supernatants for nitrite, a stable reaction product of NO with molecular oxygen [19]. Briefly, 400 μl of culture supernatant was allowed to react with 200 μl of Griess reagent and incubated at room temperature for 15 min. The optical density of the assay samples was measured spectrophotometrically at 570 nm. Fresh culture media served as the blank in all experiments. Nitrite concentrations were calculated from a standard curve derived from the reaction of NaNO2 in the assay.
Fluorescence measurements for superoxide production using hydroethidine
Hydroethidine (HE) or dihdroethidium (DHE), a redox sensitive probe, have been widely used to detect intracellular superoxide anion. The oxidation of HE in a superoxide generating system was performed by spectrofluorimetry, essentially according to the method described by Zhao, et al [23] with slight modifications. Briefly, following treatment of cells with LPS, with or without Aβ ± AICAR (1 mM), for 6 h, the cultures were rinsed in PBS and the medium was replaced with fresh medium containing 50 μM HE (stock solution 5 mM in dimethyl sulfoxide) in DMEM/high glucose-containing medium. Following incubation for 60 min at 37°C, cells were rinsed twice in phosphate-buffered saline (PBS) to remove any unbound dye and then lysed in buffer containing 0.1 N NaOH in 50% MetOH and vortexed for 20 min on a shaker. Generation of ROS was measured by a fluorescence plate reader, at an excitation wavelength of 510 nm, and emission at 595 nm (gain 10). The blank values consisted of wells containing no cells but loaded with HE and identically processed. Equal volumes of PBS or NaOH-MetOH were added for cell lysis, before fluorescence measurement.
Immunoblot analysis
These were performed essentially as described earlier [12,19]. Briefly, glial cells (2 × 106/ml), after incubation in the presence or absence of different stimuli, cell lysates was prepared in 0.5 ml of buffer containing 20 mM HEPES, pH 7.4, 2 mM EDTA, 250 mM NaCl, 0.1% Nonidet, P-40, 0.1% Triton-X (100), 2 μg/ml leupeptin, 2 μg/ml aprotinin, 1 mM phenylmethylsulfonyl fluoride, 0.5 μg/ml benzamidine, and 1 mM dithiothreitol. The lysate was briefly centrifuged at 500 rpm for 10 min, and the supernatant was collected. Cell extract protein (50 μg) was then resolved on 4–10% SDS-PAGE, electrotransferred onto a nitrocellulose membrane, blotted with indicated antibodies, and then detected by chemiluminescence (ECL; Amersham Pharmacia Biotech).
Preparation of nuclear extracts and electrophoretic mobility shift assay (EMSA)
Nuclear extracts from treated or untreated cells (1 × 107) were prepared using the method of Dignam et al, [24] with slight modification. Cells were harvested, washed twice with ice-cold PBS, and lysed in 400 μl of buffer A (10 mM HEPES, pH 7.9; 10 mM KCl; 2 mM MgCl2; 0.5 mM dithiothreitol; 1 mM phenylmethylsulfonyl fluoride; 5 μg/ml aprotinin; 5 μg/ml pepstatin A; and 5 μg/ml leupeptin) containing 0.1% Nonidet P-40 for 15 min on ice, vortexed vigorously for 15 s, and centrifuged at 14,000 rpm for 30 s. The pelleted nuclei were resuspended in 40 μl of buffer B (20 mM HEPES, pH 7.9; 25% (v/v) glycerol; 0.42 M NaCl; 1.5 mM MgCl2; 0.2 mM EDTA; 0.5 mM dithiothreitol; 1 mM phenylmethylsulfonyl fluoride; 5 μg/ml aprotinin; 5 μg/ml pepstatin A; and 5 μg/ml leupeptin). After 30 min on ice, the lysates were centrifuged at 14,000 rpm for 10 min. Supernatants containing the nuclear proteins were diluted with 20 μl of modified buffer C (20 mM HEPES, pH 7.9; 20% (v/v) glycerol; 0.05 M KCl; 0.2 mM EDTA; 0.5 mM dithiothreitol; and 0.5 mM phenylmethylsulfonyl fluoride) and stored at -70°C until further use. Nuclear extracts were used for the electrophoretic mobility shift assay using the NFκB DNA-binding protein detection system kit (Life Technologies, Inc.) according to the manufacturer's protocol. Briefly, the protein-binding DNA sequences (previously labeled with 32P) of C/EBP, NFκB, AP-1 and CREB were incubated with nuclear extracts prepared after various treatments of glial cells. The DNA-protein binding reactions were performed at room temperature for 20 min in 10 mM Trizma base pH 7.9, 50 mM NaCl, 5 mM MgCl2, 1 mM EDTA, and 1 mM dithiothreitol plus 1 μg of poly (dI-dC), 5% (v/v) glycerol, and ~0.3 pmol of 32P labeled either C/EBP, NFκB, AP-1 or CREB (all from Santa Cruz Biotechnology). Protein DNA complexes were resolved from protein-free DNA in 5% polyacrylamide gels at room temperature in 50 mM Tris, pH 8.3, 2 mM EDTA and were detected by autoradiography. For Supershift analysis, 1 μg of the respective antibody (wherever indicated) was included in the DNA protein-binding reaction.
Real-time PCR
Real time PCR was performed as described previously [25,26]. Briefly, total RNA from cells was isolated with Trizol (Gibco) according to the manufacturer's protocol. Real-time PCR was conducted using a Bio-Rad iCycler (iCycler iQ Multi-Color Real-Time PCR Detection System; Bio-Rad, Hercules, CA). Single stranded cDNA was synthesized from RNA isolated from untreated, LPS/β-amyloid-treated cells in the presence or absence of AICAR using the Superscript preamplification system for first-strand cDNA synthesis (Life Technologies, Gaithersburg, MD). Total RNA (5 μg) was treated with 2 U DNase I (bovine pancreas; Sigma) for 15 min at room temperature in 18 μl volume containing 1× PCR buffer and 2 mM MgCl2. It was then inactivated by incubation with 2 μl of 25 mM EDTA at 65°C for 15 min. Random primers were added (2 μl) and annealed to the RNA according to the manufacturer's instructions. cDNA was prepared using poly-dT as a primer and Moloney murine leukemia virus reverse transcriptase (Promega) according to manufacturer's instructions. The primer sets used were designed and synthesized by Integrated DNA technologies (IDT, Coralville, IA). The primer sequences are: for glyceraldehyde-3-phosphate dehydrogenase (GAPDH), forward 5'-CCTACCCCCAATGTATCCGTTGTG-3' and reverse 5'-GGAGGAATGGGAGTTGCTGTTGAA-3'; IL-1β, forward 5'-GAGAGACAAGCAACGACAAAATCC-3' and reverse 5'-TTCCCATCTTCTTCTTTGGGTATTG-3'; TNFα, forward 5'-CTTCTGTCTACTGAACTTCGGGGT-3' and reverse 5'-TGGAACTGATGAGAGGGAGCC-3'; and iNOS, forward 5'-GGAAGAGGAACAACTACTGCTGGT-3' and reverse 5'-GAACTGAGGGTACATGCTGGAGC-3'. IQTM SYBR Green Supermix was purchased from Bio-Rad. Thermal cycling conditions were as follows: activation of iTaq DNA polymerase at 95°C for 10 min, followed by 40 cycles of amplification at 95°C for 30 sec and 55–57.5°C for 30 sec. Levels were expressed as arbitrary units normalized to expression of the target gene relative to GAPDH.
Cytokine assay
The levels of TNF-α, IL-1β, and IL-6 were measured in culture supernatant by ELISA using protocols supplied by the manufacturer (R & D Systems).
Transcriptional assays
Primary astrocytes were transiently transfected with NF-κB-, or C/EBP-luciferase reporter gene with β-galactosidase by Lipofectamine-2000 (Invitrogen) according to the manufacturer's instructions. pcDNA3 was used to normalize all groups to equal amounts of DNA. Luciferase activity was determined using a luciferase kit (Promega).
Cell viability
Cytotoxic effects of various treatments were determined by measuring the metabolic activity of cells with MTT and LDH release assay (Roche).
Studies on phaeochromocytoma (PC-12) cell neurite extension
Rat phaeochromocytoma (PC-12) cells were plated on 60-mm Petri dishes precoated with 10 mg/ml poly-D-lysine and cultured in Kaighn's modified medium containing 20% Horse serum and 2% FBS, 100 U/ml penicillin, and 100 mg/ml streptomycin (All from GIBCO-BRL) for ~12 h. The cells were then incubated in low-serum media (2% horse serum and 1% bovine calf serum) containing NGF (50 ng/ml) for 48 h before challenging them again with NGF (50 ng/ml) either in the presence or absence of astroglial LPS-conditioned medium and/or AICAR. The cells were then evaluated after 4 days of stimulation by phase contrast microscopy (Olympus). The images obtained were adjusted to set to a color background for clarity using Adobe Photoshop software (version 7). Scoring for neurite outgrowth of PC-12 cells was performed as described previously by Dikic et al.[27]. Briefly, neurite lengths greater than 100 μM were taken into consideration and were scored and compared with relevant controls.
Statistical analysis of the data
All data are expressed as means + SEM. All necessary comparisons were carried out using the Tukey-Kramer multiple comparison test. Statistical differences at p < 0.05 were considered significant. The densitometric data for iNOS and MnSOD, and for all phosphorylation blots are expressed on an arbitrary scale.
Results
AICAR attenuates LPS- and Aβ peptide-induced expression of cytokines and iNOS, and NO production in glial cells
It has been suggested that, in the CNS, activated microglia and astrocytes are linked to neurodegeneration as a result of expression of inflammatory mediators by these glial cells [6,28,29]. Major cytokines implicated in AD (with the major exception of IFN-γ), include TGF-β, TNF-α, IL-1, IL-2, IL-6, IL-10 and IL-12 [3]. In addition to cytokine expression and release, rat primary glial cells are known to express iNOS as well as COX-2. As mentioned earlier, LPS has been routinely used to stimulate/induce the inflammatory cytokine responses in glial cells [7,8,10]. Hence, to mimic the inflammatory responses, rat primary glial cell cultures were treated with LPS ± Aβ (1–42) peptide. As evident from (Fig. 1a,c,e and 1g) and 2, Aβ significantly upregulated the LPS-induced production of cytokines TNF-α, IL-1β, IL-6 and nitric oxide (NO) in glial cells, which is further supported by increases in the expression of mRNA for iNOS, TNF-α, IL-1β and IL-6 (Fig. 1b,d,f and 1h). AICAR attenuated the LPS/Aβ-induced production of TNF-α, IL-1β, IL-6 and NO, and of their mRNA expression, in a dose-dependent manner (Fig. 1a–h).
Figure 1 AICAR inhibits LPS- and Aβ peptide-induced cytokine production. Astrocyte-enriched glial cells (mixed glial cells) were pre-incubated with different concentrations of AICAR (as indicated) for 4 h and were stimulated with 1 ng/ml LPS ± Aβ peptide (1–42) (15 μM) as shown. After 18 h of incubation, concentrations of NO, TNF-α, IL-1β, and IL-6 released into the culture medium were measured using ELISA (left panel figs. a, c, e and g). Alternatively the cells were harvested for RNA by extraction with Trizol (see methods) and the levels of mRNA for cytokines were measured (See right panel figs. b, d, f and h) by real time-PCR (RT-PCR). Data are expressed as the mean ± SD of three different experiments. *P < 0.001 was considered significant.
Figure 2 AICAR treatment inhibits LPS + Aβ-stimulated iNOS gene expression and nitric oxide release in glial cells. Cell cultures were pre-incubated with 1 mM AICAR following stimulation by LPS ± Aβ peptides (1–40) or (1–42) in concentrations as indicated. The corresponding reverse peptide (40–1) in lane 3 and 9 served as a positive control in this assay. The production of NO (top) and expression of iNOS, COX-2, and MnSOD was determined in cell lysates, 18 h following treatment, by immunoblot analysis (bottom). Experiments were performed in triplicate and data are means (±SEM). P < 0.05 compared to relative control value was considered significant. However, P value for histograms in lane 8 and 9 (* and **) not significant.
We have previously reported that SMase-activated ceramide release is redox sensitive and that ceramide-mediated induction of MnSOD and reactive oxygen species (ROS) generation is central to inflammatory responses in glial cells [12,30-32]. Hence expression of MnSOD was routinely evaluated as a ROS-induced stress sensor protein [33]. Figure 2 shows the expression of iNOS, MnSOD and COX-2 in glial cells. Aβ peptide upregulated the LPS-mediated expression of MnSOD. Aβ peptides (1–40) and (1–42) induced the expression of iNOS, COX-2, and MnSOD; but not Aβ (40–1) peptide in reverse sequence. Cells responded to both Aβ peptides (1–40) and (1–42). However, with these two peptides in combination, at equimolar concentrations (7.5 μM each), Aβ (1–42) induced approximately twice the amount of nitric oxide release and correspondingly higher iNOS expression as compared to Aβ (1–40) (lane 11 vs. lane 14). At higher concentrations (15 μM each) of these peptides, the differences in iNOS expression or nitrite production (lane 12 Vs 15) were no longer evident. AICAR treatment blocked the iNOS, COX-2 expression, as well as a nearly normalized expression of MnSOD.
The treatment of glial cells with LPS and Aβ peptide (25–35) elicited a similar inflammatory response in terms of cytokine release and iNOS, COX-2 and MnSOD expression (Fig. 3 and 4) as well as a dose-dependent inhibition with AICAR. Figure 3A shows the dose-dependent expression of TNF-α, IL-1β and IL-6 by Aβ peptide (25–35) and figure 3B shows the dose-dependent inhibition of these cytokines by AICAR. Fig 4 shows inhibition of iNOS and MnSOD expression by 0.5 mM AICAR to a fixed concentration of LPS (125 μg/ml) with various concentrations of Aβ peptide.
Figure 3 AICAR inhibits LPS- and Aβ (25–35) peptide-induced cytokine production in glial cells stimulated with 125 ng/ml LPS ± 7.5 μM Aβ peptide (25–35). After 18 h of incubation, concentrations of TNF-α, IL-1β, and IL-6 released into the culture medium were measured using ELISA. Fig. 3A. Shows dose-response curves, using LPS and various concentrations of Aβ (25–35) peptide in stimulating cytokine release in glial cells. Fig. 3B shows a dose-response inhibition of cytokine release with various concentrations of AICAR (0.25 to 1 mM) following stimulation with LPS + Aβ (25–35) peptide.
Figure 4 AICAR inhibits LPS- and Aβ- (25–35) induced expression of iNOS, COX-2, and MnSOD in astrocyte-enriched glial cells. Cells were preincubated with 1 mM AICAR for 4 h prior to treatment with either LPS or Aβ (25–35) in concentrations indicated earlier. After 18 h incubation, an aliquot of the medium was used for nitrite measurement as described under Methods. Data are mean ± SD of three different experiments. (A) Cell homogenates were used for western-immunoblot analysis of iNOS, COX-2 and MnSOD. Western immunoblots for iNOS (B) and MnSOD (C) upon treatment of glial cell cultures to LPS and to various concentrations of Aβ (25–35) either in the presence or absence of 0.5 mM AICAR (lane 1 is control, lane 2 LPS alone, and in lanes 3 to 6, Aβ was added to final concentrations of 7.5, 15, 30 or 45 μM, respectively). The protein bands were scanned on a densitometric scanner and represented as a graph (bottom). Experiments were performed in triplicate and data are means (±SEM). *P < 0.05 compared to relative control value was considered significant.
Taken together, these studies indicate that Aβ (25–35) peptide induces proinflammatory responses similar to those observed with Aβ (1–40 or 1–42) peptide. Hence, Aβ (25–35) peptide was used in the rest of this study. Cell viability was tested under experimental conditions as described in this study but no toxicity was evident in MTT or in LDH-release assays.
The observed expression of cytokines TNF-α and IL-1β by activated glial cells (Figures 1 and 3), is consistent with expression of these cytokines in brains of experimental animal models of Alzheimer's and in the brain of Alzheimer's disease patients [3]. We have reported previously that Aβ also upregulates TNF-α/IL-1β-induced iNOS expression and nitrite release [12]. Hence, to study autocrine/paracrine effects, astrocytes in culture were treated with TNF-α/IL-1β. As shown in figure 5, TNF-α/IL-1β treatment led to increased iNOS and COX-2 expression, and nitrite production which was further upregulated by the addition of Aβ (25–35) peptide. The induction of these pro-inflammatory mediators was also significantly attenuated by AICAR (Fig. 5). Further, the increased expression of anti-oxidant enzyme MnSOD in response to TNF-α/IL-1β ± Aβ (25–35), was also markedly reduced upon pre-treatment of glial cells with AICAR (Fig. 5C). From these studies we conclude that AICAR attenuates LPS/cytokine- and Aβ peptide-induced inflammatory cytokine release; and iNOS, COX-2 and MnSOD expression.
Figure 5 AICAR inhibits TNF-α-, and/or IL-1β- and/or Aβ- (25–35) stimulated iNOS expression and nitric oxide release in astrocytic cellcultures. Cells were pre-incubated with 1 mM AICAR prior to stimulation. Fig. A shows nitric oxide released into the medium upon treatment with cytokines +/- Aβ (25–35) with relevant controls. Figure B shows nitric oxide released into the medium and corresponding western-immunoblot for iNOS, COX-2, and MnSOD, after stimulation with cytokines (TNF-α + IL-1β ± Aβ peptide), either in the presence or absence of AICAR in concentrations used in figure A. Figure C shows a dose-dependent inhibition of nitric oxide production and expression of iNOS, COX-2 and MnSOD proteins, on stimulation with cytokines and Aβ and after pre-incubation with increasing amounts of AICAR, as shown. The increase in p-AMPK protein band (Thr-172) indicates activation of AMPK with increasing concentrations of AICAR.
Anti-oxidant functions of AICAR on LPS, Aβ-induced oxidative stress responses
Earlier studies from our laboratory [12,30], as well as others [20,34] have reported cytokine- or LPS- and Aβ-induced alterations in cellular redox activate the sphingomyelin(SM)-ceramide (Cer) signal-transduction cascade by conversion of sphingomyelin to ceramide in glial cells in culture. This pro-inflammatory cascade of events could be blocked by anti-oxidants such as NAC and vitamin E as well as by neutral sphingomyelinase inhibitor (3-o-methyl sphingomyelin) [12,19,30]. The elevated expression of MnSOD, Cu/ZnSOD, reactive oxygen species (ROS), and reduction in glutathione, indicate altered redox balance upon LPS, Aβ treatment, which was attenuated by vitamin E treatment [35]. Quantification of production of ROS, after treatment of glial cells with LPS/Aβ peptide, using a fluorescent dye-based assay (HE fluorescence) showed an increase in ROS generation, which was blocked by AICAR pre-treatment (Fig. 6A). This inhibition of ROS generation by AICAR treatment possibly blocks the down-stream targets thereby inhibiting the inflammatory gene expression. The generation of ceramide from sphingomyelin was reported to be redox sensitive [30] and ceramide generated by exogenous sphingomyelinase upregulated the expression of iNOS [30]. We observed that SMase – [the enzyme that degrades sphingomyelin (SM) to ceramide (cer)] and Aβ-treatment of glial cells also leads to increased iNOS expression and NO production which is inhibited by preincubating the cells with AICAR (Fig. 6B). This also confirmed our previous observations of the involvement of SM-ceramide cascade-signaling in expression of iNOS and cytokines [12]. These observed alterations of SM-Cer- and ROS-mediated signaling, with LPS/Aβ-induced expression of proinflammatory mediators, by antioxidant activity of AICAR are consistent with our previous observations that LPS/Aβ-induced expression of iNOS and production of NO are blocked by anti-oxidants (vitamin E or NAC) (Fig. 6C) and thus support the conclusion that AICAR functions in blocking the generation of ROS and in turn the SM-ceramide cascade as a suppressor of pro-oxidant activity [12]. Moreover, intracellular glutathione and mercaptans (which includes total cellular thiol group compounds) levels, which showed a decrease with LPS/SMase and/or Aβ peptide treatment, were restored to significant levels with AICAR treatment (fig. 7), thereby confirming AICAR's potential to balance the cellular redox status.
Figure 6 AICAR down-regulates Aβ ± LPS- or sphingomyelinase-induced generation of reactive oxygen species (ROS) and expression of iNOS, COX-2 and MnSOD. Figure A shows LPS ± Aβ- (25–35) induced ROS generation in glial cell cultures. Pre-treatment of glial cells with AICAR inhibits LPS- and Aβ-mediated ROS generation. Cells were pre-incubated in the culture medium with 1 mM AICAR for 4 h prior to treatment with LPS (0.125 μg/ml) and/or 7.5 μM Aβ (25–35). ROS generation was measured by incubating the cells with the fluorescent dye Hydro Ethidene (HE) as described under Methods. Columns where AICAR was added have been shown diagramatically for brevity. Figure B: anti-oxidants (NAC and vitamin E) mediated inhibition of LPS, Aβ-stimulated nitric oxide release in glial cultures. Cells were pretreated with either AICAR (1 mM), vitamin E (20 μM) or N-acetyl cysteine (NAC) (10 mM) 4 h prior to stimulation with 125 ng/ml LPS and Aβ 7.5 μM (25–35). Nitric oxide released into the medium was measured as described under methods. Figure C shows AICAR mediated inhibition of both SMase- and Aβ-stimulated nitric oxide release (top) as well as the expression of iNOS, COX-2 and MnSOD in astrocyte enriched glial cultures (bottom). Cultures were pre-incubated with AICAR 4 h prior to stimulation with SMase (5 units/ml) ± 7.5 μM Aβ (25–35). Nitric oxide produced in the culture medium was measured using 'Greis reagent'. The cell lysates were western-immunoblotted for iNOS, COX-2 and MnSOD expression.
Figure 7 Effect of AICAR in normalization of LPS-, or SMase- and/or Aβ peptide- (25–35) induced decreases in cellular glutathione and total mercaptans (thiol group containing compounds). Figure A shows a histogram plot of NO released after treatment with LPS or SMase and/or Aβ peptide, either in the presence or absence of AICAR (1 mM). Figure B shows corresponding levels of glutathione (in light colored bars) and total mercaptans (dark colored bars).
AICAR treatment upregulates phosphorylation of AMPK, and possibly down-regulates the Pkb/Akt cascade
Recent reports from Jhun et al., [36] and a previous study by Morrow et al., [37] reported the involvement of Pkb/Akt kinases via activation of PI-3 kinase in the nitric oxide release pathways in macrophages and in endothelial cells. Hence, we tested the phosphorylation status of Akt upon stimulation with LPS/Aβ, with or without treatment with AICAR. There was an increase in phosphorylation of Ser-473 of p-Akt on stimulation of cells with LPS/Aβ, which was significantly reduced in AICAR-treated cells (Fig. 8A).
Figure 8 AICAR inhibits LPS-and Aβ-induced activation of Pkb/Akt kinase activity, but activates AMP kinase activity, in astrocyte-enriched glial cell cultures. Cultures pretreated with AICAR or untreated cells were stimulated with LPS (0.125 μg/ml) and 7.5 μM Aβ for the indicated time periods following which cell homogenates were western-immunoblotted for phosphorylated forms of AMPK and Pkb/Akt. Figure 8A shows immunoblots for p-Pkb/Akt proteins. Samples (1 and 7) correspond to control, (2 and 8), (3 and 9), (4 and 10), (5 and 11), (6 and 12) correspond to cells treated with LPS + Aβ +/- 1 mM AICAR for 15 min, 30 min, 45 min, 4 h or 12 h respectively (as shown). Figure B shows phosphor-AMP K (Thr-172 of the α1/α2 subunits) and Figure C to phosphorylated AMPK (Ser 108) of AMPK β subunit. In Figures B and C samples (1) control (2 and 5), (3 and 6) and (4 and 7) correspond to cells treated with LPS + Aβ peptide (for 15 min, 30 min or 60 min respectively) with or without AICAR pre-treatment. Fig D, shows AICAR-mediated inhibition of TNF-α/IL-1β- and Aβ-induced activation of ERK and activation of AMPK. Cells were pre-incubated with AICAR (1 mM) for 4 h prior to treatment with cytokine and Aβ (25–35). Cell homogenates were prepared at indicated time points and western immunoblotted for either phosphorylated or nonphosphorylated iso forms as shown.
We previously reported that AICAR mediates its effects via activation of AMPK and that activated AMPK downregulates pro-inflammatory responses by downregulation of the IKK cascades [19]. Inside the cell (in vivo) AICAR is converted to ZMP (an analog of AMP) which activates AMP kinase kinase (AMPKK) which in turn activates AMP kinase (AMPK) by phosphorylation on residues Thr 172 of the α1/α2 subunits and on Ser 108 of the β subunit of AMPK. AICAR treatment of glial cells activated AMPK as evident from the enhanced intensities of the phospho-specific protein bands of this AMPK at Ser-108 and Thr-172 (Fig. 8B and 8C). Immunoblot analysis of cytokine- (TNF-α/IL-1β) treated cells showed significantly increased ERK phosphorylation (MAP kinase activation) and Aβ treatment further upregulated this MAP kinase activation (Fig. 8D). AICAR treatment down-regulated cytokine/Aβ-induced activation of MAP kinases. These observations indicate that AICAR activation of AMP kinase by phosphorylation of its catalytic subunits (Thr-172 of α1/α2 subunits) may possibly down-regulate MAP kinase activation and inhibition of proinflammatory gene expression. However, at present it is not clear how the activation of AMP kinase cascade would mediate reduced activation of the MAP kinases.
AICAR inhibits LPS- or SMase- and Aβ (25–35)-induced NFκB, AP-1, C/EBP and CREB binding activity
Transcription factors such as NFκB, AP-1, CREB and C/EBP are often the downstream targets of MAP kinase signaling cascades, for the transactivation of genes expressed under proinflammatory conditions [12,38,39]. These transcription factors have consensus sequences in the promoter regions of proinflammatory genes such as iNOS, COX-2, MnSOD as well as those of cytokines [38]. Therefore, we investigated the possible role of these transcription factors in AICAR-mediated regulation of expression of proinflammatory genes. Cell cultures transiently transfected with expression vectors for NFκB-luciferase or C/EBP-luciferase, upon stimulation with LPS/Aβ (1–42) peptide upregulated luciferase activity, a reflection of the activation of these transcription factors. Treatment with AICAR showed dose-dependent attenuation of these luciferase activities (Fig. 9A). To further confirm these observations, we performed EMSA for activation of these transcription factors. Aβ treatment upregulated the LPS- or SMase-induced binding of these transcription factors and this enhanced binding activity was blocked by AICAR treatment (Fig. 8B). NFκB/IKK-mediated transcriptional activity involves different subunits of NFκB (RelA/p65, c-Rel and Rel B or the NFκB p50 and p52 heteromeric combination) [38,40]. Supershift analysis of NFκB using antibodies to various subunits of NFκB demonstrated the possible participation/involvement of p65, p52 and Rel B subunits in the NFκB complexes [Fig. 9(c-i)]. C/EBP β and δ are reported to be the key regulators in the pro-inflammatory cascades in glial cells surrounding the amyloid plaques in Alzheimer's disease brains [12,39]. Similar analysis using antibodies to several different C/EBP subunits revealed the involvement of C/EBP α, β and δ components in LPS/Aβ peptide-induced activation of C/EBP (Fig. 9C-ii).
Figure 9 (A) AICAR inhibits LPS- and Aβ-induced activation of NFκB, AP-1 CREB and C/EBP transcription factors. NFkB luciferase and C/EBP luciferase activities in glial cells transfected for NFkB luciferase (i) or C/EBP luciferase following stimulation with LPS and Aβ (1–42) peptide (figure A) were measured as described under legends to figure 1. Experiments were performed in triplicate and data are expressed as mean ± SEM. *P < 0.05 compared to relative control value was considered significant. AICAR inhibited the LPS- or SMase- and Aβ-induced NFκB, C/EBP, CREB and AP-1 binding activity as seen by EMSA (figure B). EMSA was carried out using the nuclear extracts prepared from astrocyte-enriched glial cells after treatment with LPS or SMase and Aβ for 1 h either in the presence or absence of AICAR. In case of EMSA for NFκB, for clarity, the dried gel was exposed for autoradiography either for longer (24 h) or for shorter (6 h) periods. The top shows the picture of the autoradiogram of the shorter exposure time (Sx) and the bottom shows the longer exposure time (Lx). In figure B(i) polyclonal IgGs specific for NFκB subunits -p65, p52, p50, RelB or cRel were used for supershift experiments with nuclear extracts from LPS, Aβ-treated (1 h) glial cells for binding to γ-32P-labeled NFκB oligomer. Note the supershifted complexes in lanes 2, 4 and 5 (correspond to -p65, p50 and -RelB proteins). In figure (ii) lanes 2–7, polyclonal IgGs specific for C/EBP α, β, p-β, δ and ε were used in supershift experiments with nuclear extracts from LPS, Aβ-treated (1 h) glial cells for binding to γ-32P-labeled C/EBP oligomer. ε antibodies from two different stocks were tested (in lanes 6 and 7). Note the supershifted complexes in lanes 2, 3 and 5 corresponding to -α, β and -δ subunits of C/EBP.
AICAR attenuates the inhibition of neurite outgrowth in PC-12 cells by astroglial conditioned medium obtained from glia following stimulation with LPS and Aβ (25–35)
Accumulation of amyloid-β protein, a pathological hallmark of AD, also contributes to many alterations of neuronal structure leading to axonal loss and consequent to neuronal cell loss [41,42]. To further investigate the role of AICAR as a neuroprotective agent against cytokines and inflammatory mediators released by glial cells (cytokines, prostanoids and nitric oxide), we studied the effect of conditioned media from LPS/Aβ peptide-treated glial cells on NGF-induced neurite extension of PC-12 cells. The scheme of treatment is described in figure 10A; and under 'Methods'.
Figure 10 AICAR promotes PC-12 cell neurite extension following stimulation with NGF in the presence of glial conditioned medium. Figure A shows the scheme of treatment using glial cell-conditioned medium (see Methods) either in the presence or absence of AICAR (1 mM). Figures B shows phase contrast micrographs of PC-12 cells of control untreated cells and NGF differentiated PC-12 cells. Where figure B(i) shows control untreated cells (40× magnification) and treated with NGF (50 ng/ml) form neurites reminiscent of axonal network in neurons as shown in figure B (ii) at 40× magnification. AICAR protection against glial conditioned medium is demonstrated in figure 10C where figure (i) is vehicle (control) NGF + 1 mM AICAR treatment. Inhibition of NGF-induced neurite outgrowth observed on treatment with glia-conditioned medium (from LPS, Aβ treatment) showing aggregation of cells (ii) and its reversal in AICAR pre-treated and NGF- and conditioned medium-challenged cells (iii), (all at 20× magnification). Aβ (25–35) (1 μM) treatment was performed as one of the controls (iv) 20× magnification. Experiments were carried out in triplicate. Figure D shows a histogram plot of neurite lengths greater than or equal to 100 μm in the above experiment in vehicle-treated, condition medium-treated, AICAR-treated or Aβ- (25–35) treated PC-12 cells. Data shown are mean ± SEM. # p < 0.5 was considered significant. Figure E shows the effects of different concentrations of AICAR in the presence of fixed amount of conditioned medium on neurite outgrowth. AICAR was tried at three different concentrations against condition medium challenged cells. Values shown are relative percentage values. The neurite outgrowth was recorded after 96 h of NGF stimulation. Quantitative analysis of neurite outgrowth is from ~300 cells. The data was plotted as mean ± SD from three experiments. *p < 0.5 was considered significant. P values for Conditioned medium/AICAR treated samples * and ** significant.
PC-12 cells (Fig. 10), following stimulation with NGF, showed extensive neuronal growth forming a network reminiscent of neuronal extensions (Fig 10B). They tend to form extensive neurites that appeared to be connected well to adjacent cells. Challenging the cells with pre-conditioned medium from glial cells led to extensive loss of NGF-induced neurite extension and also to cell clustering and cell death (Fig. 10C). AICAR treated cells (vehicle) without the conditioned media on average showed 20% fewer neurites than control untreated cells (Fig. 10D). However, conditioned medium-challenged cells showed greater than 80% loss in neurites on the average, whereas, AICAR treatment reduced this loss to 25% loss of neurites. To rule out the possibility that Aβ (25–35) peptide from the conditioned medium may be a contributing factor in the observed loss of neurites, Aβ (25–35) (1 μM), treated PC-12 cells showed ~18% loss in neurites as compared to greater than 80% loss in conditioned medium treated cells. AICAR treatment at concentrations of 0.25, 0.5 and 1 mM led to 40-, 53- and 62-percent protection against neurite loss, respectively (Fig. 10E). However, higher concentrations of AICAR (2 mM) were found to be toxic (data not shown). In a similar set of experiments we observed similar protection of neurite extension by AICAR against cytokines (TNF-α and IL-1β)/Aβ (7.5 μM) (data not shown here).
Discussion
We have previously reported that LPS/Aβ-induced expression of proinflammatory mediators (e.g. iNOS) is mediated via the SM-ceramide-associated cellular redox signaling cascade [12,32]. This study reports activation of AMPK by AICAR, and its possible down-regulation of LPS- or cytokine/Aβ-mediated signaling events associated with cellular oxidative stress and inflammatory activity. These conclusions are based on the following observations. The anti-oxidant functions of AICAR are evident from the observations that AICAR blocks LPS- and LPS ± Aβ-induced ROS generation (Fig. 6) as well as nitric oxide release, very similar to that seen with other anti-oxidants such as NAC or Vitamin E [12,43]. Moreover, MnSOD expression (the mitochondrial oxidative stress sensor), was upregulated along with proinflammatory cytokines, iNOS, and COX-2 in LPS/Aβ-stimulated cells and their expression was down-regulated following AICAR treatment (Figs 2, 3, 4, 5, 6). These antioxidant and anti-inflammatory functions of AICAR (Figs 1, 2, 3), its associated protective effects, and its promotion of neurite outgrowth extension by PC-12 cells (Fig. 10) exposed to glia-conditioned media (and hence to inflammatory mediators secreted by activated glial cells) indicate that AICAR may provide protection against inflammatory mediators (cytokines, NO and O2•-) and Aβ-mediated toxicity to neurons in AD. In cell culture studies (in vitro) AICAR is effective in attenuating the inflammatory process at 0.5–1 mM, and up to 2 mM with no toxicity observed. We have also tested the efficacy of AICAR in attenuation of ischemia-reperfusion injury in a canine model of autologous renal transplantation at 50 mg/kg body weight [44], an animal model of experimental autoimmune encephalomyelitis [45], an animal model of multiple sclerosis at 100–500 mg/kg body weight, and LPS-induced neurotoxicity in rats at 100 mg/kg body weight [19], with no side effects. This points to the in vivo beneficial effects of this compound against inflammatory immune response mechanisms.
AICAR, upon internalization into the cells is immediately phosphorylated by adenosine kinase to AICA riboside monophosphate ZMP (a purine nucleotide) which mimics the effect of AMP without altering the cellular ratio of ATP/AMP and thus activates AMP kinase kinase (AMPKK), and in turn AMPK [16,46]. AMPK is known to regulate glucose transport [47] and its metabolism [48], lower blood pressure, boost liver insulin action [49], ameliorate insulin resistance induced by free fatty acids [18], and regulate protein synthesis [50] and forkhead transcription factor FKHR (FOXO1a) [51]. It is also reported to down-regulate the synthesis of fatty acids as well as cholesterol [52]. These observations indicate participation of AMPK in the regulation of cellular energy metabolism. We have earlier reported on the anti-inflammatory role of AICAR via activation of AMPK in quenching LPS-induced pro-inflammatory responses by blockade of MAP kinase and IKK α/β-signaling cascades [19]. AICAR treatment activated AMP kinase activity, and antisense oligonucleotides for AMPKKα as well as expression of dominant negative cDNA of AMPKα in glia reversed the AICAR-mediated inhibition of iNOS gene expression in response to LPS treatment [19]. These above studies demonstrate AICAR attenuation of LPS- or cytokine/Aβ-induced expression of inflammatory mediators (e.g. iNOS, COX-2 and cytokines) by inhibiting the activation of transcription factors (NFκB and C/EBP) required for induction of the inflammatory process. Moreover, the current study highlights yet another novel function of AICAR in protection of neurite outgrowth against the toxicity of inflammatory mediators secreted by activated microglia and astrocytes. This protection of neurite growth may in part be mediated via the energy-(ATP) saving mechanisms of AMPK since activation of AMPK is known to shut or slow energy consuming reactions in the cell [16].
Alterations in cellular redox appear to be central to inflammatory events associated with amyloid toxicity (Fig. 11) [12,20,43]. Cytokines as well as Aβ peptides are known to perturb the intracellular redox state via generation of reactive nitrogen species (RNS; NO, ONOO-) and reactive oxygen species (ROS; O2-, OH- and H2O2) [53,54] and, more importantly, by reducing cellular thiols [glutathione and other mercaptans (total thiol group containing compounds)] [35]. Glial cells treated with LPS and Aβ showed significantly reduced intracellular levels of glutathione and mercaptans. However, AICAR treatment restored intracellular thiols (Fig. 7). Similarly, MnSOD expression was nearly normalized in AICAR-treated cells (Figs 2, 3, 4, 5, 6). These findings support the idea of antioxidant/ anti-inflammatory functions of AICAR and thereby the potential of AICAR as possible therapy for inflammatory disease processes.
Figure 11 Figure A: Possible anti-oxidant mechanisms involved in the LPS, Aβ-induced ceramide generation leading to superoxide anion formation, nitric oxide release and peroxynitrite generation. Figure B: inflammation is possibly triggered as a result of imbalance in the radical generating systems and radical scavenger systems creating an oxidative stress, thus leading to the formation of nitric oxide and superoxide anion generation, and thereby depleting cellular anti-oxidants. Figure C. Putative role of ceramide in eicosanoid synthesis. Ceramide generated as a result of LPS- and Aβ-peptide induced SMase activation, leads to the release of eicosanoids. Eicosanoids (leukotrienes and prostaglandins) thus generated may perhaps potentially enhance or ameliorate the cytokine-induced pro-inflammatory responses or vice versa. Non steroidal anti-inflammatory drugs (NSAIDs) as well as COX inhibitors block these responses. Figure D. Overall scheme in the LPS, Aβ-induced pro-inflammatory signaling cascade involving cytokine release thereby leading to the expression of iNOS, COX-2 and MnSOD. The anti-inflammatory effect of AICAR is perhaps a result of its multiple regulatory roles. However, AICAR blockade of ROS generation keeps the redox balance in check, thereby inhibiting the inflammatory signaling cascade.
Details of signal transduction pathways that mediate the neurotoxic effects of β-amyloid on neurons and on glia remain elusive. However, glial biology in relation to neuro-inflammatory responses is important for the following reasons: a) Glial cells out number neurons; b) Glia are involved in the upregulation of cytokines and iNOS and thus may participate in chronic β-amyloid-induced activation of astrocytes observed in AD [55]. Astroglia-released cytokines can further activate surrounding astrocytes which may be necessary to phagocytose excessively generated amyloid. The possible role of NO in neuronal damage is supported by the protection observed with NOS inhibitor (Ng-nitro-L-arginine methyl ester [L-NAME]) in Aβ- (1–42) induced selective loss of cholinergic neurons [13,29,56]. Furthermore, induction of iNOS following direct injection of β-amyloid into rat brain also supports a role for NO-induced toxicity in Aβ-mediated neurotoxicity [55]. Release of inflammatory cytokines, iNOS, ROS and NO may cause direct stress to neurons. However ROS and NO generation in the same environment can have potentially detrimental effects, due to the formation of peroxynitrite radicals which have the potential to cause neuronal stress and apoptosis (Fig. 11A and [57]). In rodent model studies, astrocytes are reported to damage neurons through NO production [3]. Hence our findings described in this study, documenting inhibition of production of both NO and ROS by AICAR, suggest AICAR/AMPK-mediated protection against cytokine/Aβ-induced oxidative stress/neurotoxicity in AD.
Several studies have indicated that use of non-steroidal anti-inflammatory drugs (NSAIDs) may delay the onset and/or slow the cognitive decline in AD [58,59]. COX-2 is an important enzyme in the PLA2 pathways for the synthesis of various eicosanoids (Fig. 11 and [55]). There is evidence that COX-2 may exacerbate neuronal injury in a variety of diseases [58]. It has been reported that ceramide generated by activated SMase activates cPLA2 cascades leading to enhanced COX-2 expression and hence to the release of eicosanoids [55,60]. ROS are yet another by-product of the conversion of arachidonic acid to prostanoids (prostaglandins and leukotrienes), and perhaps one of the leading contributors of neuronal cell death [61]. COX-2 over-expression has been reported in apoptotic neuronal cell death, and inhibition of COX-2 activity has been reported to protect neurons against excitoxicity in ischemia- and seizure-induced injury [58,62]. Specific COX-2 inhibitors have also been reported to suppress COX-2 activity and to reduce neuronal cell death in the CNS of animal models of cerebral ischemia [63,64]. Upregulation of COX-2 expression in an Alzheimer's mouse model [65] and in cell culture studies has been reported in response to Aβ toxicity [66], indicating the potential of selective COX-2 inhibitors as neuroprotective agents in AD [58,59]. Since, iNOS and COX-2 are important components of the post-lesion inflammatory cascade in various types of brain damage [67], the observed suppression of Aβ and LPS/cytokine-induced COX-2/iNOS expression in glial cell cultures indicates the potential of AICAR to protect against Aβ-induced inflammatory disease process.
Conclusion
The major themes of ROS and RNS formation associated with the neuroinflammatory processes, and the suppression of these stress mechanisms by antioxidants, continue to yield promising leads for new therapies. Anti-oxidants have been reported to have beneficial effects against Alzheimer's disease [6,20]. Numerous studies in various experimental paradigms of neuronal cell death both in vitro and in vivo, have shown protection by free radical scavengers including vitamin E, estrogen, ebselen, flavanoids, N-acetyl cysteine, glutathione, α-lipoic acid, etc [20]. The fact that Aβ peptide-associated oxidative damage leads to neuroinflammation, which is effectively attenuated/blocked by AICAR treatment, provides strong evidence that altered redox equilibrium processes are directly related to neuroinflammation.
Disease progression in Alzheimer's disease (AD) often causes massive neuronal stress, contributing to the loss of cognitive function observed in the disease. Many brain regions in patients with AD show changes in axonal and dendritic fields, dystrophic neurites, synapse loss, as well as neuronal loss [41]. Accumulation of amyloid-β protein and tau-induced changes (in the form of 'neurofibrillary tangles') are pathological hallmarks of the disease and are believed to contribute to many of these alterations of neuronal structures [42]. More so, areas of the brain displaying high degrees of plasticity are particularly vulnerable to degeneration in Alzheimer's disease. Perhaps this reflects a loss in the regenerative capacity of the brain, relative to renewed axonal growth, or perhaps a reduced capability of pluripotent stem cells to replace dystrophic neurites. Hence AICAR's potential to aid neurite outgrowth in PC-12 cells challenged with toxic mediators suggests that it may prove beneficial in AD, perhaps leading to functional recovery in these patients. In conclusion, the observed anti-inflammatory and anti-oxidant and neuroprotective functions of AICAR point to the multiple regulatory and therapeutic potentials of this drug for AD.
List of abbreviations used
AICAR (5-aminoimidazole-4-carboxamide-1-beta-4-ribofuranoside); Aβ (beta amyloid peptide); ROS (Reactive oxygen species); RNS (Reactive nitrogen species); NGF (Nerve growth factor); MnSOD (manganese superoxide dismutase); SDS-PAGE (sodium dodecyl sulfate-polyacrylamide gel electrophoresis); MTT (methylthiazoletetrazolium); EMSA (Electrophoretic mobility shift assay); COX-2 (Cycloxygenase-2); TNF-α (tumor necrosis factor alpha); SMase (Sphingomyelinase); ROS (reactive oxygen species); NFκB (Nuclear factor kappa B); C/EBP (CCAAT enhancer binding protein).
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
KRA carried out the various experiments, participated in the design of the study and helped draft the manuscript.; AKS and IS participated in the design of the study and helped to draft the manuscript. All authors read and approved the final manuscript.
Acknowledgements
The authors wish to thank Dr. Shailendra Giri for his immense help during various stages of experimentation and in the process of preparation of the manuscript. We also thank Ramandeep Rattan for laboratory assistance, Dr. Manu Jatana for his assistance with microscopy and Ms. Joyce Bryan for secretarial help. This work was supported in parts by grants (NS-22576, NS-34741, NS-40144, NS-40810 and NS-37766) from National Institutes of Health.
==== Refs
Selkoe DJ Alzheimer's disease results from the cerebral accumulation and cytotoxicity of amyloid beta-protein J Alzheimers Dis 2001 3 75 80 12214075
Abramov AY Canevari L Duchen MR Beta-amyloid peptides induce mitochondrial dysfunction and oxidative stress in astrocytes and death of neurons through activation of NADPH oxidase J Neurosci 2004 24 565 575 14724257 10.1523/JNEUROSCI.4042-03.2004
McGeer EG McGeer PL Inflammatory processes in Alzheimer's disease Prog Neuropsychopharmacol Biol Psychiatry 2003 27 741 749 12921904 10.1016/S0278-5846(03)00124-6
Akiyama H Barger S Barnum S Bradt B Bauer J Cole GM Cooper NR Eikelenboom P Emmerling M Fiebich BL Finch CE Frautschy S Griffin WS Hampel H Hull M Landreth G Lue L Mrak R Mackenzie IR McGeer PL O'Banion MK Pachter J Pasinetti G Plata-Salaman C Rogers J Rydel R Shen Y Streit W Strohmeyer R Tooyoma I Van Muiswinkel FL Veerhuis R Walker D Webster S Wegrzyniak B Wenk G Wyss-Coray T Inflammation and Alzheimer's disease Neurobiol Aging 2000 21 383 421 10858586 10.1016/S0197-4580(00)00124-X
Tsukamoto E Hashimoto Y Kanekura K Niikura T Aiso S Nishimoto I Characterization of the toxic mechanism triggered by Alzheimer's amyloid-beta peptides via p75 neurotrophin receptor in neuronal hybrid cells J Neurosci Res 2003 73 627 636 12929130 10.1002/jnr.10703
Minagar A Shapshak P Fujimura R Ownby R Heyes M Eisdorfer C The role of macrophage/microglia and astrocytes in the pathogenesis of three neurologic disorders: HIV-associated dementia, Alzheimer disease, and multiple sclerosis J Neurol Sci 2002 202 13 23 12220687 10.1016/S0022-510X(02)00207-1
Cordle A Landreth G 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitors attenuate beta-amyloid-induced microglial inflammatory responses J Neurosci 2005 25 299 307 15647473 10.1523/JNEUROSCI.2544-04.2005
Kim SH Smith CJ Van Eldik LJ Importance of MAPK pathways for microglial pro-inflammatory cytokine IL-1 beta production Neurobiol Aging 2004 25 431 439 15013563 10.1016/S0197-4580(03)00126-X
Monsonego A Imitola J Zota V Oida T Weiner HL Microglia-mediated nitric oxide cytotoxicity of T cells following amyloid beta-peptide presentation to Th1 cells J Immunol 2003 171 2216 2224 12928365
Rampe D Wang L Ringheim GE P2X7 receptor modulation of beta-amyloid- and LPS-induced cytokine secretion from human macrophages and microglia J Neuroimmunol 2004 147 56 61 14741428 10.1016/j.jneuroim.2003.10.014
Monsonego A Weiner HL Immunotherapeutic approaches to Alzheimer's disease Science 2003 302 834 838 14593170 10.1126/science.1088469
Ayasolla K Khan M Singh AK Singh I Inflammatory mediator and beta-amyloid (25-35)-induced ceramide generation and iNOS expression are inhibited by vitamin E Free Radic Biol Med 2004 37 325 338 15223066 10.1016/j.freeradbiomed.2004.04.007
Tran MH Yamada K Nakajima A Mizuno M He J Kamei H Nabeshima T Tyrosine nitration of a synaptic protein synaptophysin contributes to amyloid beta-peptide-induced cholinergic dysfunction Mol Psychiatry 2003 8 407 412 12740598 10.1038/sj.mp.4001240
Salt IP Johnson G Ashcroft SJ Hardie DG AMP-activated protein kinase is activated by low glucose in cell lines derived from pancreatic beta cells, and may regulate insulin release Biochem J 1998 335 ( Pt 3) 533 539 9794792
Blazquez C Geelen MJ Velasco G Guzman M The AMP-activated protein kinase prevents ceramide synthesis de novo and apoptosis in astrocytes FEBS Lett 2001 489 149 153 11165240 10.1016/S0014-5793(01)02089-0
Hardie DG Carling D The AMP-activated protein kinase--fuel gauge of the mammalian cell? Eur J Biochem 1997 246 259 273 9208914 10.1111/j.1432-1033.1997.00259.x
Woods A Johnstone SR Dickerson K Leiper FC Fryer LG Neumann D Schlattner U Wallimann T Carlson M Carling D LKB1 is the upstream kinase in the AMP-activated protein kinase cascade Curr Biol 2003 13 2004 2008 14614828 10.1016/j.cub.2003.10.031
Olsen GS Hansen BF AMP kinase activation ameliorates insulin resistance induced by free fatty acids in rat skeletal muscle Am J Physiol Endocrinol Metab 2002 283 E965 70 12376323
Giri S Nath N Smith B Viollet B Singh AK Singh I 5-aminoimidazole-4-carboxamide-1-beta-4-ribofuranoside inhibits proinflammatory response in glial cells: a possible role of AMP-activated protein kinase J Neurosci 2004 24 479 487 14724246 10.1523/JNEUROSCI.4288-03.2004
Behl C Moosmann B Antioxidant neuroprotection in Alzheimer's disease as preventive and therapeutic approach Free Radic Biol Med 2002 33 182 191 12106814 10.1016/S0891-5849(02)00883-3
Pahan K Sheikh FG Namboodiri AM Singh I Lovastatin and phenylacetate inhibit the induction of nitric oxide synthase and cytokines in rat primary astrocytes, microglia, and macrophages J Clin Invest 1997 100 2671 2679 9389730
Pike CJ Burdick D Walencewicz AJ Glabe CG Cotman CW Neurodegeneration induced by beta-amyloid peptides in vitro: the role of peptide assembly state J Neurosci 1993 13 1676 1687 8463843
Zhao H Kalivendi S Zhang H Joseph J Nithipatikom K Vasquez-Vivar J Kalyanaraman B Superoxide reacts with hydroethidine but forms a fluorescent product that is distinctly different from ethidium: potential implications in intracellular fluorescence detection of superoxide Free Radic Biol Med 2003 34 1359 1368 12757846 10.1016/S0891-5849(03)00142-4
Dignam JD Lebovitz RM Roeder RG Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei Nucleic Acids Res 1983 11 1475 1489 6828386
Khan M Sekhon B Jatana M Giri S Gilg AG Sekhon C Singh I Singh AK Administration of N-acetylcysteine after focal cerebral ischemia protects brain and reduces inflammation in a rat model of experimental stroke J Neurosci Res 2004 76 519 527 15114624 10.1002/jnr.20087
Paintlia AS Paintlia MK Singh AK Stanislaus R Gilg AG Barbosa E Singh I Regulation of gene expression associated with acute experimental autoimmune encephalomyelitis by Lovastatin J Neurosci Res 2004 77 63 81 15197739 10.1002/jnr.20130
Dikic I Schlessinger J Lax I PC12 cells overexpressing the insulin receptor undergo insulin-dependent neuronal differentiation Curr Biol 1994 4 702 708 7953556 10.1016/S0960-9822(00)00155-X
Minagar A Shapshak P Heyes M Sheremata WA Fujimara R Ownby R Goodkin K Eisdorfer K Microglia and astrocytes in neuro-AIDS, alzheimers disease, and multiple sclerosis ScientificWorldJournal 2001 1 69
Gasic-Milenkovic J Dukic-Stefanovic S Deuther-Conrad W Gartner U Munch G Beta-amyloid peptide potentiates inflammatory responses induced by lipopolysaccharide, interferon -gamma and 'advanced glycation endproducts' in a murine microglia cell line Eur J Neurosci 2003 17 813 821 12603271 10.1046/j.1460-9568.2003.02506.x
Singh I Pahan K Khan M Singh AK Cytokine-mediated induction of ceramide production is redox-sensitive. Implications to proinflammatory cytokine-mediated apoptosis in demyelinating diseases J Biol Chem 1998 273 20354 20362 9685387 10.1074/jbc.273.32.20354
Pahan K Sheikh FG Khan M Namboodiri AM Singh I Sphingomyelinase and ceramide stimulate the expression of inducible nitric-oxide synthase in rat primary astrocytes J Biol Chem 1998 273 2591 2600 9446561 10.1074/jbc.273.5.2591
Won JS Im YB Khan M Singh AK Singh I The role of neutral sphingomyelinase produced ceramide in lipopolysaccharide-mediated expression of inducible nitric oxide synthase J Neurochem 2004 88 583 593 14720208 10.1046/j.1471-4159.2003.02165.x
Kamata H Honda S Maeda S Chang L Hirata H Karin M Reactive oxygen species promote TNFalpha-induced death and sustained JNK activation by inhibiting MAP kinase phosphatases Cell 2005 120 649 661 15766528 10.1016/j.cell.2004.12.041
Cutler RG Kelly J Storie K Pedersen WA Tammara A Hatanpaa K Troncoso JC Mattson MP Involvement of oxidative stress-induced abnormalities in ceramide and cholesterol metabolism in brain aging and Alzheimer's disease Proc Natl Acad Sci U S A 2004 101 2070 2075 14970312 10.1073/pnas.0305799101
Ayasolla K Singh AK Singh I Vitamin E restores amyloid peptide (25-35) induced changes in cellular redox in C6 glioma cells; implications to Alzheimer's disease. Special Edition of free Radical Biology & Medicine; Proceedings of 'XI Biennial meeting of the society of free Radical Research International'. Monduzzi Editore S.p.A-MEDIMOND Inc. C7160487/503
Jhun BS Jin Q Oh YT Kim SS Kong Y Cho YH Ha J Baik HH Kang I 5-Aminoimidazole-4-carboxamide riboside suppresses lipopolysaccharide-induced TNF-alpha production through inhibition of phosphatidylinositol 3-kinase/Akt activation in RAW 264.7 murine macrophages Biochem Biophys Res Commun 2004 318 372 380 15120611 10.1016/j.bbrc.2004.04.035
Morrow VA Foufelle F Connell JM Petrie JR Gould GW Salt IP Direct activation of AMP-activated protein kinase stimulates nitric-oxide synthesis in human aortic endothelial cells J Biol Chem 2003 278 31629 31639 12791703 10.1074/jbc.M212831200
Lee AK Sung SH Kim YC Kim SG Inhibition of lipopolysaccharide-inducible nitric oxide synthase, TNF-alpha and COX-2 expression by sauchinone effects on I-kappaBalpha phosphorylation, C/EBP and AP-1 activation Br J Pharmacol 2003 139 11 20 12746218 10.1038/sj.bjp.0705231
Cardinaux JR Allaman I Magistretti PJ Pro-inflammatory cytokines induce the transcription factors C/EBPbeta and C/EBPdelta in astrocytes Glia 2000 29 91 97 10594926 10.1002/(SICI)1098-1136(20000101)29:1<91::AID-GLIA9>3.0.CO;2-I
Li X Massa PE Hanidu A Peet GW Aro P Savitt A Mische S Li J Marcu KB IKKalpha, IKKbeta, and NEMO/IKKgamma are each required for the NF-kappa B-mediated inflammatory response program J Biol Chem 2002 277 45129 45140 12221085 10.1074/jbc.M205165200
Spires TL Hyman BT Neuronal structure is altered by amyloid plaques Rev Neurosci 2004 15 267 278 15526551
Mandelkow EM Stamer K Vogel R Thies E Mandelkow E Clogging of axons by tau, inhibition of axonal traffic and starvation of synapses Neurobiol Aging 2003 24 1079 1085 14643379 10.1016/j.neurobiolaging.2003.04.007
Gilgun-Sherki Y Melamed E Offen D Oxidative stress induced-neurodegenerative diseases: the need for antioxidants that penetrate the blood brain barrier Neuropharmacology 2001 40 959 975 11406187 10.1016/S0028-3908(01)00019-3
Lin A Sekhon C Sekhon B Smith A Chavin K Orak J Singh I Singh A Attenuation of ischemia-reperfusion injury in a canine model of autologous renal transplantation Transplantation 2004 78 654 659 15371664 10.1097/01.TP.0000131664.18670.17
Nath N Giri S Prasad R Salem ML Singh AK Singh I 5-aminoimidazole-4-carboxamide ribonucleoside: a novel immunomodulator with therapeutic efficacy in experimental autoimmune encephalomyelitis J Immunol 2005 175 566 574 15972693
Kemp BE Stapleton D Campbell DJ Chen ZP Murthy S Walter M Gupta A Adams JJ Katsis F Van Denderen B Jennings IG Iseli T Michell BJ Witters LA AMP-activated protein kinase, super metabolic regulator Biochem Soc Trans 2003 31 162 168 12546677
Chen HC Bandyopadhyay G Sajan MP Kanoh Y Standaert M Farese RVJ Farese RV Activation of the ERK pathway and atypical protein kinase C isoforms in exercise- and aminoimidazole-4-carboxamide-1-beta-D-riboside (AICAR)-stimulated glucose transport J Biol Chem 2002 277 23554 23562 11978788 10.1074/jbc.M201152200
Halse R Fryer LG McCormack JG Carling D Yeaman SJ Regulation of glycogen synthase by glucose and glycogen: a possible role for AMP-activated protein kinase Diabetes 2003 52 9 15 12502487
Iglesias MA Ye JM Frangioudakis G Saha AK Tomas E Ruderman NB Cooney GJ Kraegen EW AICAR administration causes an apparent enhancement of muscle and liver insulin action in insulin-resistant high-fat-fed rats Diabetes 2002 51 2886 2894 12351423
Bolster DR Crozier SJ Kimball SR Jefferson LS AMP-activated protein kinase suppresses protein synthesis in rat skeletal muscle through down-regulated mammalian target of rapamycin (mTOR) signaling J Biol Chem 2002 277 23977 23980 11997383 10.1074/jbc.C200171200
Barthel A Schmoll D Kruger KD Roth RA Joost HG Regulation of the forkhead transcription factor FKHR (FOXO1a) by glucose starvation and AICAR, an activator of AMP-activated protein kinase Endocrinology 2002 143 3183 3186 12130586 10.1210/en.143.8.3183
Henin N Vincent MF Gruber HE Van den Berghe G Inhibition of fatty acid and cholesterol synthesis by stimulation of AMP-activated protein kinase Faseb J 1995 9 541 546 7737463
Wu G Fang YZ Yang S Lupton JR Turner ND Glutathione metabolism and its implications for health J Nutr 2004 134 489 492 14988435
Floyd RA Antioxidants, oxidative stress, and degenerative neurological disorders Proc Soc Exp Biol Med 1999 222 236 245 10601882 10.1046/j.1525-1373.1999.d01-140.x
Sugaya K Uz T Kumar V Manev H New anti-inflammatory treatment strategy in Alzheimer's disease Jpn J Pharmacol 2000 82 85 94 10877525 10.1254/jjp.82.85
Weldon DT Maggio JE Mantyh PW New insights into the neuropathology and cell biology of Alzheimer's disease Geriatrics 1997 52 Suppl 2 S13 6 9307579
Luth HJ Munch G Arendt T Aberrant expression of NOS isoforms in Alzheimer's disease is structurally related to nitrotyrosine formation Brain Res 2002 953 135 143 12384247 10.1016/S0006-8993(02)03280-8
Pasinetti GM Aisen PS Cyclooxygenase-2 expression is increased in frontal cortex of Alzheimer's disease brain Neuroscience 1998 87 319 324 9740394 10.1016/S0306-4522(98)00218-8
Aisen PS Evaluation of selective COX-2 inhibitors for the treatment of Alzheimer's disease J Pain Symptom Manage 2002 23 S35 40 11992749 10.1016/S0885-3924(02)00374-3
Futerman AH Hannun YA The complex life of simple sphingolipids EMBO Rep 2004 5 777 782 15289826 10.1038/sj.embor.7400208
Carlson NG Neuroprotection of cultured cortical neurons mediated by the cyclooxygenase-2 inhibitor APHS can be reversed by a prostanoid J Neurosci Res 2003 71 79 88 12478616 10.1002/jnr.10465
Qin W Ho L Pompl PN Peng Y Zhao Z Xiang Z Robakis NK Shioi J Suh J Pasinetti GM Cyclooxygenase (COX)-2 and COX-1 potentiate beta-amyloid peptide generation through mechanisms that involve gamma-secretase activity J Biol Chem 2003 278 50970 50977 14507922 10.1074/jbc.M307699200
Candelario-Jalil E Alvarez D Castaneda JM Al-Dalain SM Martinez-Sanchez G Merino N Leon OS The highly selective cyclooxygenase-2 inhibitor DFU is neuroprotective when given several hours after transient cerebral ischemia in gerbils Brain Res 2002 927 212 215 11821016 10.1016/S0006-8993(01)03358-3
Khayyam N Thavendiranathan P Carmichael FJ Kus B Jay V Burnham WM Neuroprotective effects of acetylsalicylic acid in an animal model of focal brain ischemia Neuroreport 1999 10 371 374 10203337
Hwang DY Chae KR Kang TS Hwang JH Lim CH Kang HK Goo JS Lee MR Lim HJ Min SH Cho JY Hong JT Song CW Paik SG Cho JS Kim YK Alterations in behavior, amyloid beta-42, caspase-3, and Cox-2 in mutant PS2 transgenic mouse model of Alzheimer's disease Faseb J 2002 16 805 813 12039862 10.1096/fj.01-0732com
Bazan NG Lukiw WJ Cyclooxygenase-2 and presenilin-1 gene expression induced by interleukin-1beta and amyloid beta 42 peptide is potentiated by hypoxia in primary human neural cells J Biol Chem 2002 277 30359 30367 12050157 10.1074/jbc.M203201200
Acarin L Peluffo H Gonzalez B Castellano B Expression of inducible nitric oxide synthase and cyclooxygenase-2 after excitotoxic damage to the immature rat brain J Neurosci Res 2002 68 745 754 12111835 10.1002/jnr.10261
|
16174294
|
PMC1262754
|
CC BY
|
2021-01-04 16:38:22
|
no
|
J Neuroinflammation. 2005 Sep 20; 2:21
|
utf-8
|
J Neuroinflammation
| 2,005 |
10.1186/1742-2094-2-21
|
oa_comm
|
==== Front
J Neuroengineering RehabilJournal of NeuroEngineering and Rehabilitation1743-0003BioMed Central London 1743-0003-2-291619754310.1186/1743-0003-2-29ResearchThe development of postural strategies in children: a factorial design study Schmid Maurizio [email protected] Silvia [email protected] Luisa [email protected] Paolo [email protected]'Alessio Tommaso [email protected] Dipartimento di Elettronica Applicata, Università degli Studi "Roma TRE", Italy2 Unità di Neurologia Infantile, Università degli Studi di Roma "Tor Vergata", Italy3 Dipartimento di Psicologia, Università degli Studi di Roma "La Sapienza", Italy2005 30 9 2005 2 29 29 17 12 2004 30 9 2005 Copyright © 2005 Schmid et al; licensee BioMed Central Ltd.2005Schmid et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The present study investigates balance control mechanisms, their variations with the absence of visual input, and their development in children from 7 to 11 years old, in order to provide insights on the development of balance control in the pediatric population.
Methods
Posturographic data were recorded during 60 s trials administered on a sample population of 148 primary school children while stepping and then quietly standing on a force plate in two different vision conditions: eyes closed and eyes open. The extraction of posturographic parameters on the quiet standing phase of the experiment was preceded by the implementation of an algorithm to identify the settling time after stepping on the force plate. The effect of different conditions on posturographic parameters was tested with a two-way ANOVA (Age × Vision), and the corresponding eyes-closed/eyes-open (Romberg) Ratios underwent a one-way ANOVA.
Results
Several posturographic measures were found to be sensitive to testing condition (eyes closed vs. eyes open) and some of them to age and anthropometric parameters. The latter relationship did not explain all the data variability with age. An evident modification of postural strategy was observed between 7 and 11 years old children.
Conclusion
Simple measures extracted from posturographic signals resulted sensitive to vision and age: data acquired from force plate made it possible to confirm the hypothesis of the development of postural strategies in children as a more mature selection and re-weighting of proprioceptive inputs to postural control in absence of visual input.
Postural ControlDevelopmentChildren
==== Body
Background
Postural control has been studied throughout a century and a half [1], and the development of balance characteristics associated with the emergence and refinement of motor control has been investigated for three decades [2]. Central Nervous System (CNS) responses and developmental changes occurring in the first years of life have been deeply studied by Assaiante [3], and Woollacott and Shumway-Cook [4]. The quantitative analysis of human movement and posture has been generally exploited on children population to study biomechanical effects on gross motor skills driven by the presence of diverse pathologies, such as Cerebral Palsy [5-8], Spinal Cord Injury [9], and Muscular Dystrophies [10,11]. Starting from the work of Williams et al [12], in more recent years researchers extended the application of quantitative posturography to fine cognitive or learning disabilities [13], autism [14,15], Developmental Coordination Disorder (DCD) [16], Attention Deficit Hyperactivity Disorder (ADHD) [17], and dyslexia [18].
Quantitative posturography can thus be applied to obtain functional markers on fine competencies and their development. For instance, a perturbation in posture with challenges such as a compliant surface [19], or a concurrent cognitive task [20], can help to enlighten possible adjustment strategies or deficiencies, or to monitor balance control variations with age [21]. However, findings obtained from other researchers show some contradictions with the above: as an example, the study of simple orthostatic posture with eyes open has been proven unsuccessful in differentiating controls from autistic patients [15], and children with DCD from controls [16]. Thus, this application field, though promising, needs to be more deeply investigated.
The quantitative analysis of postural control is generally based on data acquired by a force plate that allows one to determine the instantaneous position of the Ground Reaction Force application point, which is referred to as Centre of Pressure (CoP). Several parameters in the time and/or frequency domain [22] are then extracted from these data, or from surrogate functions derived from them [23]. Even if this technique does not allow direct detection of body oscillations, which can be estimated through the use of ad hoc motion analysis systems, the relative simplicity of the set up has encouraged researchers to consider the CoP oscillations as an indirect measure of postural sway [24].
When dealing with posturographic measures, the detection of the stabilization time after stepping on the force plate is crucial: the majority of the parameters used to define the postural ability are summary measures, and their application is based on the assumption of stationarity, in that the statistical properties of the underlying data do not significantly change over time. In presence of a transitory response to an event, such as standing up from a chair or stepping on the force plate, this assumption cannot be considered as valid. Thus the transitory response should be excluded from the analysis. By analysing the first and second order moment of the CoP trajectory, Carroll and Freedman [25] estimated this non-stationary interval to be about 20 seconds long. This assumption can be however challenged by considering that the transitory phase due to a similarly demanding perturbation, such as the Sit to Stand task, has been estimated in about 3 seconds [26]. Carpenter et al. [27] showed that the first order moment of the CoP Power Spectral Density could give insights on the duration of the transitory response.
A significant age dependence of the postural measures has been demonstrated [28,29]: from a longitudinal study, Kirschenbaum et al. [30] showed that the control strategy to maintain balance does not follow a simple linear relationship with age, but a step-like transition at the age of 6 to 8 years occurs. This hypothesis can be linked to a clear rise in normalized stability limits to adult levels at age 7, as calculated by Riach and Starkes [31] by asking children to lean as far as they could in the four directions (forward, backward, left, and right) while standing. These results suggest that, at that age, the exploratory behaviour is reached, and thus the child has to work with a new strategy, which takes into account both open loop and closed loop components of balance control. By analysing postural responses to unpredicted translations of the base of support, Sundermier et al. [32] hypothesized that the development of postural control follows the maturation of fine competencies in muscle coordination.
A variety of posturographic parameters have been shown to depend on biomechanical and anthropometric factors, such as height or weight [33], and when extracting the CoP mean amplitude on a sample population ranging from 7 to 80 years, Peterka showed no changes with age if normalization with height was performed [34].
Thus, the question remains as to whether there is any reliable marker extracted from posturographic data that can give insights on the development of balance control, and whether age significantly affects posturographic data or changes as simply the result of anthropometric factors. Aim of the present study is to investigate mechanisms involved in the development of postural stability by attempting to answer these questions.
Methods
Participants
148 children were selected from classes of three different grades in one primary school, after obtaining proper informed consent from parents and teachers to participate in the study. None of the children had educational needs or certified disabilities. After the collection of height and weight, they were screened with a three-sided testing procedure: Quantitative Posturography, Physical Examination for Neurological Subtle Signs (PANESS), and Teachers' Rating. For the present study, PANESS Assessment [35] and Teachers' Rating were used for inclusion criteria for the sample population, and by excluding subjects outside 10th-90th percentile, the resulting sample size for data analysis on Quantitative Posturography was reduced to 107 children, divided into three age groups (n = 41 for Seven Years' Group, Y7, n = 38 for Nine Years' Group, Y9, and n = 28 for Eleven Years' Group, Y11). Table 1 summarizes data on participants, and Table 2 provides information on PANESS and Teachers' Rating.
Table 1 Population anthropometric data
Age Group Y7 Y9 Y11
N 41 38 28
Age (yrs) 7.0 ± 0.3 (Range 6.5–7.5) 9.0 ± 0.3 (Range 8.0–9.8) 11.0 ± 0.3 (Range 10.5–12.0)
Height (m) 1.22 ± 0.06 1.34 ± 0.07 1.46 ± 0.06
Weight (kg) 25.3 ± 4.7 32.5 ± 7.1 43.1 ± 8.7
BMI (kg/m2) 17.0 ± 2.1 18.0 ± 2.8 20.0 ± 3.1
Table 2 Teachers' Rating and PANESS Assessment
Teachers' Rating
Cluster Definition Score
Read and Write reading: speed and correctness writing: tract quality and correctness oral language production (vocabulary richness and fluency and structure) Scoring 0–3
0 is best score
Arithmetics Arithmetics text: reading and placing numbers
Arithmetcs logic: operations
Sequences: understands and repeats sequences days, months, alphabets and multiplication tables Scoring 0–3
0 is best score
Attention and Movement Motor activity in the gym/garden: follow instructions without confusing left-right, in/out
Motor activity in class: from being able to sit still, to fine movements to gross movements he cannot avoid Attention: attention span Scoring 0–3
0 is best score
Behavior: creativity: having many interests
Social behavior: being integrated in class group and having friends
Team working: following group rules
Autonomy: not needing continuous instructions Scoring 0–3
0 is best score
PANESS*
Cluster Definition Score
Errors errors on tip-toe walking
errors on heel walking
errors on nose-finger (right)
errors on nose-finger (left) scoring 0–3, depending on total number of errors (oscillations or falls during walking, misses or wrong fingers during other tests)
Precision Index-little tapping on thumb (right)
Index-little tapping on thumb (left)
Tandem walking sequence of movements is correct from index to little with no repetitions or misses independently of rhythm Scoring 0–3.
Rhythm Index-little tapping on thumb (right)
Index-little tapping on thumb (left)
Tandem walking the self chosen rhythm is kept during task independently of misses of repetitions.
Scoring 0–3.
*Adapted from Denckla [35].
Total scores for PANESS and Teachers' Rating were obtained by summing each cluster value. Subjects were excluded if at least one total score was outside [10–90] percentile.
Procedure
A posturographic test was performed, which consisted of 2 tests of upright stance (lasting 60 seconds each) corresponding to two different conditions: standing with eyes open (EO), and standing with eyes closed (EC). Between tests an interval of 2 minutes was allowed.
Participants were asked to select a comfortable side-by-side feet position, with their arms relaxed, then make a step forward and position themselves in the middle of the force plate, as indicated by stickers, maintaining a quiet stance. Data acquisition started immediately prior to the subject stepping on the force plate. Illumination and noise were kept under control: diffuse artificial illumination of approximately 40 lux, no remarkable fixed sound sources, experiment performed during lesson time.
Relevant force and torque components were low-pass filtered (corner frequency 20 Hz, 8th order elliptical filter, stopband attenuation 80 dB at 30 Hz, attenuation slope 135 dB/octave) and fed to an AD converter (100 samples/s, DAQCard™-AI-16E-4, by National Instruments Corporation), and then processed to obtain the Centre of Pressure trajectories in both antero/posterior and medio/lateral directions, CoP = {CoPAP(t), CoPML(t)}. The maximum of the vertical component of the ground reaction force marked the subject's stepping on the force plate.
Feature Extraction
A set of 10 summary measures were extracted from CoP data. All of them are defined and summarized in Table 3, and denoted as Posturographic Parameters (PP).
Table 3 Posturographic Parameters Definition
Posturographic Parameter Acronym Definition
Mean Velocity MV
Mean Amplitude MA
Sway Area SA
Mean Frequency MF
Mean Power Frequency{AP, ML} MPF{AP, ML}
Centroidal Frequency {AP, ML} CF{AP, ML}
Frequency at 95% {AP, ML} F95{AP, ML}
T represents the total time for processing (30 s), and CoP{AP, ML} are considered as purged of their mean value
A sample of processed data is represented in Figure 1. Together with the CoPAP trajectory over time, the time history of the corresponding instantaneous mean frequency has been depicted: Following the rationale exposed in [27], in the present work the instantaneous mean frequency (IMF) of the CoPAP trajectory was considered as a marker for the time needed to stabilize, its value was estimated, for every time instant t, using a complex covariance approach [36]. The settling time Tset was then defined as the time instant when the steepest decrease of IMF occurs. This choice can be justified from experimental evidence, i.e. the behaviour of parameters object of the analysis. Using the Mean Amplitude as an example, Figure 2 shows how, after Tset, the actual value of the parameter does not remarkably vary over time. The same applies for all the parameters object of the analysis.
Figure 1 Acquired data. A sample of time histories for the Centre of Pressure trajectory in Antero-Posterior direction (CoPAP, light gray), and instantaneous mean frequency extracted from CoPAP. The settling time Tset is also shown (black dotted line). All the Posturographic Parameters were calculated over the time period [Tset, Tset +30].
Figure 2 Instantaneous Mean Frequency. A sample of time history for the Instantaneous Mean Frequency for the Centre of Pressure Antero-Posterior (upper panel), and the Mean Amplitude value, as calculated by using 30 s starting from the corresponding time instant (lower panel). The settling time Tset used for the actual parameter estimation is also shown (black vertical line).
All PPs were calculated by retaining the first 30 seconds after Tset. Four of them can be directly extracted from the CoP trajectory, while the remaining six are used to characterize the shape of the Power Spectral Density: in particular, the Mean Power Frequency and the Centroidal Frequency are respectively representative of the barycentre and the dispersion of the Power Distribution in the frequency domain, i.e. the Power Spectral Density. F95% is finally representative of the overall breadth of the Spectrum.
PPs underwent statistical analysis, and, for each of them, the corresponding Romberg Ratio (RR), defined as the EC condition measure divided by the EO measure, was also computed and fed to statistics, as described in the following.
Statistical Analysis
All PPs were analyzed through a two-way ANOVA, with vision (EO vs. EC) and age as factors. Each condition was then separately analyzed for parameters exhibiting age effect, in the following way: Bartlett's test verified homogeneity of variances, and for parameters exhibiting different variances, Welch's ANOVA was run instead of traditional ANOVA; a Post Hoc Test for trend was also applied to different age groups.
For the whole population sample, possible relationships between PPs (dependent variables) and selected subject-specific parameters (predictors) were sought to test if differences were dependent on anthropometric factors, such as body mass (m), height (h), and body mass index (BMI = m/h2). The linear correlation between parameters and predictors was measured through the Pearson product-moment coefficient of correlation (r), and deemed reliable if a two-tailed test of significance applied to this coefficient, had p ≤ 0.05. The percentage of each PP variance that can be explained by each reliable predictor was then calculated, and denoted as σexp2.
Then, to test changes for significant interaction between age and vision, the Romberg Ratios (RR) for each parameter underwent a one-way ANOVA, with age as factor.
Results
Figure 3 summarizes sample population mean values and standard deviations for all PPs. Mean Values in EO conditions for Mean Velocity, Mean Amplitude and Sway Area were all fairly higher than those obtained on a healthy population of young adults [37]. The same did not apply to all the frequency features: Mean Power Frequency in antero-posterior (AP) direction was higher in children than in adults whereas the corresponding Centroidal Frequency was almost equal: thus, in children the CoP travelled faster, farther, and with substantially different spectral features than in adults.
Figure 3 Posturographic parameters. Mean values and standard errors in each age group, divided by vision condition. Underneath each column pair, the corresponding Romberg Ratio mean values and standard deviation is shown.
As far as the differential analysis is concerned, most of the PPs were affected by vision, partly as a function of age: the effect of vision was statistically significant in MV, SA, MA, and in all the spectral parameters. This effect was more evident in amplitude parameters, thus confirming that, regardless of age, CoP displacement and velocity increased without visual input.
As reported in Table 4, age affected MA, i.e. the lower the age, the greater the CoP displacement. Moreover, two frequency parameters in AP direction, F95AP, and CFAP, were significantly affected by vision: the spectrum of CoP in AP direction was fairly broadened, even if MPFAP did not significantly increase. Moreover, F95AP was also dependent on the interaction, i.e. its variations with respect to vision were significantly different depending on age.
Table 4 Two-Way ANOVA p-values for posturographic parameters
PP Age Vision Interaction
MV - (0.44) ** (p < 0.001) - (0.99)
SA - (0.15) ** (p < 0.001) - (0.50)
MA * (0.014) ** (p < 0.001) - (0.31)
MF - (0.18) - (0.15) - (0.40)
MPFML - (0.82) - (0.13) - (0.23)
CFML - (0.89) * (0.022) - (0.46)
F95ML - (0.42) * (0.036) - (0.28)
MPFAP - (0.18) * (0.046) - (0.14)
CFAP * (0.034) * (0.013) - (0.24)
F95AP * (0.030) * (0.009) * (0.032)
-: Not Significant
*: p < 0.05
**: p < 0.005
Table 5 shows one-way ANOVA results for the effect of age on MA, CFAP, and F95AP in both vision conditions: Mean Amplitude did not significantly vary in EO, whereas a significant (p < 0.005) and non-random (Test for Trend p < 0.05) effect of age was revealed in EC; CoP mean deviation from its mean position actually decreased with age in no-vision condition (EC), and from Bartlett's Test it can also be speculated that the decrease in variance could be a sign of more homogeneous behaviour. The broadening of the spectrum enlightened by the previous results was principally due to the significant increase of F95AP with age in EC condition (Test of Trend p < 0.005), with a significant change in F95AP variability.
Table 5 Effect of age on Posturographic Parameters
PP Age Bartlett's Test Test for Trend
MA (EO) - (0.22) * (0.046) - (0.21)
MA (EC) ** (0.0037) ** (0.0003) * (0.01)
CFAP (EO) - (0.27) * (0.044) - (0.38)
CFAP (EC) - (0.10) - (0.417) * (0.035)
F95AP (EO) - (0.51) - (0.929) - (0.70)
F95AP (EC) ** (0.005) - (0.704) ** (0.002)
- : Not Significant
* : p < 0.05
** : p < 0.005
One-way ANOVA with post hoc tests for PPs resulting in a significant effect of age, separated for vision condition: Welch ANOVA test was applied for unequal variances resulting from Bartlett's Test (i.e. on first three rows).
The correlation with anthropometric and biomechanical factors yielded the following results: only frequency parameters in EC condition, namely CFAP and F95AP, were found to be slightly dependent on mass and height, but none of them could be satisfactorily predicted by these factors (see Table 6), as the percentage of explained variance did not exceed 10% in any of them. MPFAP was slightly dependent on height, though the percentage of explained variance was only 4%. Thus, the confounding effect driven by the chosen anthropometric factors can be disregarded in this study.
Table 6 Anthropometric effect on posturographic parameters
PP Mass Height BMI
p σexp2 p σexp2 p σexp2
MPFAP (EC) - (0.061) - * (0.040) 4.0% - (0.40) -
CFAP (EC) * (0.013) 5.8% * (0.009) 6.3% - (0.18) -
F95AP (EC) * (0.0055) 7.1% * (0.0058) 7.0% - (0.0945) -
- : Not Significant
* : p < 0.05
** : p < 0.005
Regression Analysis on PP resulting in a dependence with at least one anthropometric factor. p-value, and percentage of the explained variance with the corresponding anthropometric predictor, if significant.
As a final point, the Romberg Ratios (EC/EO) revealed mean values greater than 1 for all the parameters (see Figure 3): in particular, a significant effect of age on MPFAP and F95AP was revealed, which could be the result of a significant broadening of the Power Spectral Density in EC condition in Y11. Welch's test revealed significant differences on RR variances for MPFAP and F95AP (see Table 7).
Table 7 Romberg Ratios: effect of age
RR Age Welch's Test Test for Trend
SA - (0.35) - (0.30) - (0.49)
MA - (0.14) - (0.053) - (0.051)
MPFAP * (0.025) * (0.045) * (0.020)
CFAP - (0.13) -(0.24) - ()
F95AP ** (0.0012) * (0.015) ** (0.0014)
- : Not Significant
* : p < 0.05
** : p < 0.005
One-way ANOVA p-values for Romberg Ratios, with age as factor: significance, Welch's Test for variances, and post hoc test of trend.
Discussion
A large number of posturographic measures were sensitive to the testing condition (i.e. eyes open vs. eyes closed). If the trajectory of the CoP can be considered as an indirect measure of postural sway, and thus a marker for the control of stance, the presented results confirm the well-known thesis that visual input contribution plays a relevant role in postural stabilization. From the results on MV, SA, and MA, it is indeed possible to state that, with eyes closed, the CoP displacement and velocity increased relative to eyes open. It is known that also young adults can improve postural performance by using visual targets [38], and that closing eyes affects postural measures [22]. Ratios between EC and EO in the present study, however, were rather different from those obtained by Prieto [22] on young adults: restricting the analysis to time domain measures, thus including MF which is a surrogate parameter for time domain measures, similar ratios resulted for MV, SA, and MF. On the other hand, MA ratios tended to young adults' figures only at 11 years, while remaining higher for the other ages. For the frequency domain measures, all RR on both CF and F95 revealed higher values than young adults [22], while no comparison was possible for MPF, which is by definition different from the Median Frequency computed by Prieto. Moreover, Prieto removed very low frequency (f < 0.15 Hz) shares to spectral measures, and thus a comparison could be affected by this choice.
A graphical schema of changes in postural sway is represented in Figure 4. A non monotonous trend with age was present: the control of balance, though not to be considered complete at the last stage (Y11), was rather different from the early stages (Y7 and Y9), and confirmed the hypothesis of a nonlinear development of postural control, consistent with [30,31]. To be more specific, if the overall postural performance could be summarized through the MA measure, a clear transition occurred between 9 and 11 years. At 7 and 9 years, the possible presence of a change of strategy in EC condition did not compensate for the absence of vision, thus resulting in an overall increase of MA. At 11 years, a change on the efficacy of strategy occurred, as confirmed by the significant variations on the spectral features of the CoP trajectory, both in antero-posterior and in medio-lateral directions, which determined a significant decrease of MA RR in Y11 with respect to Y9 and Y7. The invariance of both MV and its corresponding Romberg Ratio may conceal two diverse behaviours: at 7 and 9 years, the line integral increased with occluded vision mostly due to the increase of the oscillation amplitude, while at 11 it rises because of an increase in frequency of self-sustained oscillations. Basically, when the child is younger, up to 9 years, her/his postural control with eyes closed relies on major adjustments, characterized by more ample oscillations, and the child probably needs to move to different spots and remain on those until the next adjustment. After that age, data of the present work would suggest that the child can apply minor adjustments that happen over a smaller trajectory, but with higher frequency components, as shown by the substantial increase of F95%AP, and there is no need for big excursions, although overall the path remains constant. The substantial increase of data variability in Romberg Ratios for F95%AP in Y9 with respect to Y7 and Y11 confirms the hypothesis of a change in strategy around that age. This evidence is in accordance with the hypothesis of a more mature selection and re-weighting of proprioceptive inputs to postural control: a major role of this kind of afferents could result in an increase of the high frequency contributions to postural sway [39], and thus in a broadening of the spectrum. The presented results are in accordance with the presence of a non linearity in balance control processes, as evidenced by Hay and Redon [40], who justify this step-like behaviour through the refinement of on-line control, once the feedforward mode has been efficiently developed, and by Baumberger et al. [41], who showed that the age of 10 is a critical point in the development of the visual control of stability.
Figure 4 Postural development schema. A schematical representation of three parameters extracted from each population: the minor axis is proportional to the Sway Area, whereas the major axis is proportional to Mean Amplitude. Code luminance is proportional to F95AP (0.75 Hz corresponds to white, and 1.5 Hz to black). For each age group, inner ellipses turned out for Eyes Open condition, and outer ellipses for Eyes Closed. * Young adults' values are taken from Prieto et al. [22]
Conclusion
The obtained results are in favour of a non monotonic development of postural strategies in children, slightly dependent on anthropometric factors: the role of vision clearly varies within the studied age range, and probably the maturation of balance control is not yet complete, even at the age of 11. Finally, another question is to be unveiled: is the maturation of balance control paralleled by a corresponding change in cognitive processes? The application of dual tasks, such as a concurrent cognitive one, in the execution of quiet stance trials could help in providing information on this issue.
Acknowledgements
The authors are indebted to Prof. Aurelio Cappozzo, who provided the force plate for the experiments, to PsyD Annalisa Conte, for her help in data collection, and to the anonymous reviewers for their constructive feedbacks and comments. The help of the class teachers of the "Istituto Comprensivo Indro Montanelli" is greatly acknowledged. Work partially supported by MIUR.
==== Refs
Romberg MH A Manual of the Nervous Diseases of Man 1853 London, England: Sydenham Society
Kohen-Raz R Developmental patterns of static balance ability and their relation to cognitive school readiness Pediatrics 1970 46 276 285 5432158
Assaiante C Development of locomotor balance control in healthy children Neurosci Biobehav Rev 1998 22 527 532 9595565 10.1016/S0149-7634(97)00040-7
Woollacott MH Shumway-Cook A Changes in posture control across the life span–a systems approach Phys Ther 1990 70 799 807 2236223
Damiano DL Abel MF Relation of gait analysis to gross motor function in cerebral palsy Dev Med Child Neurol 1996 38 389 396 8698147
Nashner LB Shumway-Cook A Marin O Stance posture control in selected group of children with cerebral palsy: Deficits in sensory organization and muscular coordination Exp Brain Res 1983 49 393 409 6641837 10.1007/BF00238781
Ferdjallah M Harris GF Smith P Wertsch JJ Analysis of postural control synergies during quiet standing in healthy children and children with cerebral palsy Clin Biomech 2002 17 203 210 10.1016/S0268-0033(01)00121-8
Harris GF Riedel SA Matesi D Smith P Standing postural stability assessment and signal stationarity in children with cerebral palsy IEEE Trans Biomed Eng 1993 1 35 42
Smith PA Hassani S Reiners K Vogel LC Harris GF Gait analysis in children and adolescents with spinal cord injuries J Spinal Cord Med 2004 27 S44 S49 15503702
Ribot-Ciscar E Trefouret S Aimonetti JM Attarian S Pouget J Roll JP Is muscle spindle proprioceptive function spared in muscular dystrophies? A muscle tendon vibration study Muscle Nerve 2004 29 861 866 15170619 10.1002/mus.20044
Barrett R Hyde SA Scott OM Dubowitz V Changes in center of gravity in boys with Duchenne muscular dystrophy Muscle Nerve 1988 11 1157 1163 3226433 10.1002/mus.880111109
Williams H Fisher J Tritschler K Descriptive analysis of static postural control in 4, 6 and 8 year old normal and motorically awkward children American Journal of Physical Medicine 1983 62 12 26 6824069
Poblano A Ishiwara K de Lourdes Arias M Garcia-Pedroza F Marin H Trujillo M Motor control alteration in posturography in learning-disabled children Arch Med Res 2002 33 485 488 12459321 10.1016/S0188-4409(02)00397-1
Kohen-Raz R Volkmar FR Cohen DJ Postural control in children with autism J Autism Dev Disord 1992 22 419 432 1383190 10.1007/BF01048244
Molloy CA Dietrich KN Bhattacharya A Postural stability in children with autism spectrum disorder J Autism Dev Disord 2003 33 643 652 14714933 10.1023/B:JADD.0000006001.00667.4c
Geuze RH Static balance and developmental coordination disorder Hum Mov Sci 2003 22 527 548 14624832 10.1016/j.humov.2003.09.008
Jucaite A Fernell E Forssberg H Hadders-Algra M Deficient coordination of associated postural adjustments during a lifting task in children with neurodevelopmental disorders Dev Med Child Neurol 2003 45 731 742 14580128 10.1017/S0012162203001373
Moe-Nilssen R Helbostad JL Talcott JB Toennessen FE Balance and gait in children with dyslexia Exp Brain Res 2003 150 237 244 12682807
Ferdjallah M Harris GF Wertsch JJ Instantaneous postural stability characterization using time-frequency analysis Gait Posture 1999 10 129 134 10502646 10.1016/S0966-6362(99)00023-5
Fawcett AJ Nicolson RI Automatisation deficits in balance for dyslexic children Percept Mot Skills 1992 75 507 529 1408614
Forssberg H Nashner LM Ontogenetic development of postural control in man: Adaptation to altered support and visual conditions during stance J Neurosci 1982 2 545 552 7077364
Prieto TE Myklebust JB Hoffmann RG Lovett EG Myklebust BM Measures of postural steadiness: differences between healthy young and elderly adults IEEE Trans Biomed Eng 1996 45 956 965 9214811 10.1109/10.532130
Laughton CA Slavin M Katdare K Nolan L Bean JF Kerrigan DC Phillips E Lipsitz LA Collins JJ Aging, muscle activity, and balance control: physiologic changes associated with balance impairment Gait Posture 2003 18 101 108 14654213 10.1016/S0966-6362(02)00200-X
Foudriat BA Di Fabio RP Anderson JH Sensory organization of balance responses in children 3–6 years of age: a normative study with diagnostic implications Int J Pediatr Otorhinolaryngol 1993 27 255 271 8270364 10.1016/0165-5876(93)90231-Q
Carroll JP Freedman W Nonstationary properties of postural sway J Biomech 1993 26 409 416 8478345 10.1016/0021-9290(93)90004-X
Hennington G Johnson J Penrose J Barr K McMulkin ML Vander Linden DW Effect of bench height on sit-to-stand in children without disabilities and children with cerebral palsy Arch Phys Med Rehabil 2004 85 70 76 14970971 10.1016/S0003-9993(03)00407-6
Carpenter MG Frank JS Winter DA Peysar GW Sampling duration effects on centre of pressure summary measures Gait Posture 2001 13 35 40 11166552 10.1016/S0966-6362(00)00093-X
Shumway-Cook A Woollacott MH The growth of stability: postural control from a development perspective J Mot Behav 1985 17 131 147 15140688
Usui N Maekawa K Hirasawa Y Development of the upright postural sway of children Dev Med Child Neurol 1995 37 985 996 8566459
Kirshenbaum N Riach CL Starkes JL Non-linear development of postural control and strategy use in young children: a longitudinal study Exp Brain Res 2001 140 420 431 11685395 10.1007/s002210100835
Riach CL Starkes JL Stability limits of quiet standing postural control in children and adults Gait Posture 1993 1 105 111 10.1016/0966-6362(93)90021-R
Sundermier L Woollacott M Roncesvalles N Jensen J The development of balance control in children: comparisons of EMG and kinetic variables and chronological and developmental groupings Exp Brain Res 2001 136 340 350 11243476 10.1007/s002210000579
Chiari L Rocchi L Cappello A Stabilometric parameters are affected by anthropometry and foot placement Clin Biomech 2002 17 666 677 10.1016/S0268-0033(02)00107-9
Peterka RJ Black FO Age-related changes in human posture control: motor coordination tests J Vestib Res 1990 1 87 96 1670140
Denckla MB Revised Neurological Examination for Subtle Signs Psychopharmacol Bull 1985 21 773 800 4089106
Conforto S D'Alessio T Real time monitoring of muscular fatigue from dynamic surface myoelectric signals using a complex covariance approach Med Eng Phys 1999 21 225 234 10514040 10.1016/S1350-4533(99)00049-1
Schmid M Conforto S Camomilla V Cappozzo A D'Alessio T The sensitivity of posturographic parameters to acquisition settings Med Eng Phys 2002 24 623 631 12376049 10.1016/S1350-4533(02)00046-2
Lee DN Lishman JR Visual proprioceptive control of stance J Hum Mov Stud 1975 1 87 95
Giacomini PG Alessandrini M Evangelista M Napolitano B Lanciani R Camaioni D Impaired postural control in patients affected by tension-type headache Eur J Pain 2004 8 579 583 15531226 10.1016/j.ejpain.2004.02.004
Hay L Redon C Feedforward versus feedback control in children and adults subjected to a postural disturbance Exp Brain Res 1999 125 153 162 10204768 10.1007/s002210050670
Baumberger B Isableau B Fluckiger M The visual control of stability in children and adults: postural readjustments in a ground optical flow Exp Brain Res 2004 159 33 46 15372130
|
16197543
|
PMC1262755
|
CC BY
|
2021-01-04 16:37:40
|
no
|
J Neuroengineering Rehabil. 2005 Sep 30; 2:29
|
utf-8
|
J Neuroeng Rehabil
| 2,005 |
10.1186/1743-0003-2-29
|
oa_comm
|
==== Front
J Neuroengineering RehabilJournal of NeuroEngineering and Rehabilitation1743-0003BioMed Central London 1743-0003-2-301620213610.1186/1743-0003-2-30ResearchA prototype power assist wheelchair that provides for obstacle detection and avoidance for those with visual impairments Simpson Richard [email protected] Edmund [email protected] Steve [email protected] Songfeng [email protected] Dan [email protected] William [email protected] Vinod [email protected] Rory [email protected] Department of Rehabilitation Science and Technology; University of Pittsburgh; Pittsburgh, PA, USA2 Human Engineering Research Labs; VA Pittsburgh Healthcare System; Pittsburgh, PA, USA3 Department of Bioengineering; University of Pittsburgh; Pittsburgh, PA, USA4 AT Sciences; Pittsburgh, PA, USA2005 3 10 2005 2 30 30 16 2 2005 3 10 2005 Copyright © 2005 Simpson et al; licensee BioMed Central Ltd.2005Simpson et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Almost 10% of all individuals who are legally blind also have a mobility impairment. The majority of these individuals are dependent on others for mobility. The Smart Power Assistance Module (SPAM) for manual wheelchairs is being developed to provide independent mobility for this population.
Methods
A prototype of the SPAM has been developed using Yamaha JWII power assist hubs, sonar and infrared rangefinders, and a microprocessor. The prototype limits the user to moving straight forward, straight backward, or turning in place, and increases the resistance of the wheels based on the proximity of obstacles. The result is haptic feedback to the user regarding the environment surrounding the wheelchair.
Results
The prototype has been evaluated with four blindfolded able-bodied users and one individual who is blind but not mobility impaired. For all individuals, the prototype reduced the number of collisions on a simple navigation task.
Conclusion
The prototype demonstrates the feasibility of providing navigation assistance to manual wheelchair users, but several shortcomings of the system were identified to be addressed in a second generation prototype.
==== Body
Background
Introduction
The concept of power assistance for a manual wheelchair is relatively new, and represents a viable alternative for individuals who are unable to generate sufficient propulsion force to use a manual wheelchair, but do not wish to use a traditional powered mobility device [1-3]. In a power assisted manual wheelchair, the traditional rear wheel hubs are replaced with motorized hubs that serve to magnify or reduce (i.e., brake) the propulsive force applied to the rear wheel push rims by the user. Power assistance is being used as the basis for a Smart Power Assistance Module (SPAM) that provides independent power assistance to the right and left rear wheels of a manual wheelchair. The SPAM (shown in Figure 1 and Figure 2) is able to sense the propulsion forces applied by the wheelchair user and provide a smooth ride by compensating for differences in force applied to each wheel. The SPAM is also able to detect obstacles near the wheelchair, and further modify the forces applied to each wheel to avoid obstacles.
Figure 1 The Smart Power Assistance Module for Manual Wheelchairs (front view).
Figure 2 The Smart Power Assistance Module for Manual Wheelchairs (back view).
The user population for the SPAM consists of individuals with both a visual impairment and a mobility impairment that makes it difficult or impossible to ambulate independently using a white cane, guide dog, or other traditional mobility aid for the visually impaired. The American Federation for the Blind (AFB) has estimated that 9.61% of all individuals who are legally blind also use a wheelchair or scooter, and an additional 5.25% of individuals who have serious difficulties seeing (but are not legally blind) also use a wheelchair or scooter (see Appendix). A large number of potential users of the SPAM are expected to be elderly, since visual and physical impairments often accompany the natural aging process. In 2000, approximately 13% of the total US population, or an estimated 35 million people, were age 65 or older; with about 2% at least age 85. By 2030, the older population is projected to double, expanding to 70 million. People age 85 and older are the fastest growing segment of the American population and the US Census Bureau estimates that there are now 65,000 centenarians [4].
Relevant Research
Currently, the majority of non-ambulatory visually-impaired individuals are seated in a manual wheelchair and pushed by another person [5]. Depending on the extent of useful vision remaining, individuals with low-vision can operate an unmodified manual wheelchair, powered wheelchair or scooter, but the risk of an accident obviously increases with increased visual impairment. There are reports of individuals using a white cane [6] or guide dog [7] along with a wheelchair, but this is not common practice.
Despite a long history of research in smart power wheelchairs, there are very few smart wheelchairs currently on the market. Two North American companies, Applied AI and ActivMedia, sell smart power wheelchair prototypes for use by researchers, but neither system is intended for use outside of a research lab. The CALL Center of the University of Edinburgh, Scotland, has developed a wheelchair with bump sensors, a single sonar sensor, and the ability to follow tape tracks on the floor for use within a wheeled-mobility training program [8]. The CALL Center smart power wheelchair is sold in the United Kingdom (UK) and Europe by Smile Rehab, Ltd. (Berkshire, UK) as the "Smart Wheelchair." The "Smart Box," which is also sold by Smile Rehab in the UK and Europe, is compatible with wheelchairs using either Penny and Giles or Dynamics control electronics and includes bump sensors (but not sonar sensors) and the ability to follow tape tracks on the floor.
One common feature of all of these smart wheelchairs is that they are based on power wheelchairs. Power wheelchairs are a convenient platform for researchers, but have several disadvantages when compared with manual wheelchairs. In general, manual wheelchairs are lighter and more maneuverable than power wheelchairs, and can be transported in a car. Manual wheelchairs that make use of power assist hubs are heavier than traditional manual wheelchairs, and can be more difficult to disassemble for transport depending on how the hubs are attached to the frames, but still provide many of the advantages of traditional manual wheelchairs.
In a search of the literature, only one other smart wheelchair was identified that was based on a manual wheelchair. The Collaborative Wheelchair Assistant [9] controls the direction of a manual wheelchair with small motorized wheels that are placed in contact with the wheelchair's rear tires to transfer torque. Unlike the SPAM, however, the Collaborative Wheelchair Assistant restricts the wheelchair's travel to software-defined "paths."
One of the few products that is commercially-available and accommodates a manual wheelchair is the Wheelchair Pathfinder [10], a commercial product sold by Nurion Industries that can be attached to a manual or power wheelchair. The Wheelchair Pathfinder uses sonar sensors to identify obstacles to the right, left or front of the wheelchair and a laser range finder to detect drop-offs in front of the wheelchair. Feedback is provided to the user through vibrations or differently-pitched beeps. The Wheelchair Pathfinder differs from the SPAM in that the Wheelchair Pathfinder has limited sensor coverage and cannot alter the speed or direction of travel of the wheelchair to avoid obstacles.
Methods
The right side of Figure 3 shows the design of the SPAM prototype, which has been implemented "on top of" a pair of Yamaha JWII power-assist pushrim hubs (sold in the United States as the Quickie Xtender). The SPAM is able to sense (1) the propulsive force applied to each rear wheel of the wheelchair, (2) the magnitude and velocity of rotation of each rear wheel, and (3) the location of obstacles relative to the wheelchair. Information from all sensors is collected by a microprocessor which integrates information about the user's input and the surrounding environment, and passes command signals to the JWII system's microprocessor.
Figure 3 Schematic for unmodified JWII system (left) and SPAM (right).
Several types of sensors have been integrated into the SPAM. These sensors are used for (1) tracking the state of the wheelchair (e.g., wheel velocity, torque applied to each rear wheel by the user) and (2) locating obstacles in the wheelchair's environment. Obstacles are identified using infrared rangefinders, sonar sensors and bump sensors. The sonar sensors have a maximum range of 3.05 m and a minimum range of 2.54 cm. The advantages of a smaller range are that (1) the frequency of sonar readings is increased and (2) the sonar system is able to detect obstacles that are extremely close to the wheelchair, which is important for passing through doorways. Infrared range finders provide a focused, highly modulated infrared beam, providing absolute ranging based on simple triangulation. The result is an accurate range value between 0.1 and 1.0 meters in a variety of circumstances. The infrared signal functions at extremely steep angles, even exceeding sixty degrees, and does so both indoors and outdoors, even in bright sunlight. The infrared rangefinders and sonar sensors are housed in. 09 m × .06 m × .04 m boxes (shown in Figure 4), which are referred to as "sensor modules." Seven sensor modules are mounted on the current prototype. Bump sensors are attached to both footrests and the "anti-tippers" of the manual wheelchair, and are implemented using simple contact switches placed behind mechanical levers. Figure 5 shows how the sensor modules were positioned on the SPAM.
Figure 4 Sensor Module.
Figure 5 Position of Sensors on SPAM.
The SPAM's control software shares control of the wheelchair with the wheelchair operator. The wheelchair operator is responsible for choosing when – and in which direction – the wheelchair moves, while the SPAM modifies the speed of the wheelchair based on the proximity of obstacles in the wheelchair's current direction of travel. The algorithm currently employed by the SPAM forces the rear wheels to turn either at exactly the same speed and direction (moving the wheelchair straight forward or straight backward) or at the same speed and opposite directions (rotating the wheelchair in place). This greatly simplifies the task of avoiding obstacles but limits the wheelchair user's flexibility in choosing paths of travel.
The navigation assistance software was written in C and runs on a TattleTale™ (manufactured by Onset Technologies) 8-bit microprocessor. User input (either forward, backward or turn in place) and sensor data are combined into "cases" that are used to make obstacle avoidance decisions. The specific cases that are in use at any one time varies depending on the specific behavior that is desired from the SPAM (e.g., passing through a narrow doorway versus driving quickly through a room with few obstacles). No single case can cause the software to prevent both forward/backward movement and turning, but multiple cases can be triggered at once and result in a situation in which the wheelchair will not move in any direction. The motorized hubs can be turned off in these situations, at which point the SPAM behaves like a normal (but heavy) manual wheelchair.
Results
Four able-bodied members of the investigative team and an individual who is blind, but does not have a mobility impairment, took part in an evaluation of the SPAM prototype. Approval for this research was obtained from the University of Pittsburgh's Institutional Review Board. All participants used the SPAM to complete the two obstacle courses shown in Figure 6 and Figure 7. Able-bodied participants were asked to complete each course three times blindfolded with navigation assistance from the SPAM. The participant who is blind completed each course nine times, in alternating sets of three trials. The sets of three trials alternated between the SPAM providing navigation assistance (condition woa) and the SPAM acting as a normal manual wheelchair (i.e., the hubs were powered but the SPAM was not acting to avoid collisions; condition noa). All subjects completed trials with Course 1 first.
Figure 6 Obstacle Course 1.
Figure 7 Obstacle Course 2.
As shown in Figure 8, the SPAM did not completely eliminate collisions for able-bodied subjects. However, three of four subjects had no collisions after the first trial on Course 1, and only one of the four subjects had a collision in any trial on Course 2. As shown in Figure 9, able-bodied subjects generally completed both navigation tasks more quickly by the third trial.
Figure 8 Collisions for able-bodied participants, in courses 1 and 2.
Figure 9 Time to complete the navigation task for able-bodied participants, in courses 1 and 2.
As shown in Figure 10, the subject who was visually-impaired had no collisions in the first three trials on Course 1 (with obstacle avoidance active) but did have collisions on Course 1 when obstacle avoidance was removed. On Course 2, where obstacle avoidance was not active during the first three trials, the visually-impaired subject had collisions in the first three trials but did not have collisions once obstacle avoidance was introduced. As shown in Figure 11, there was not a consistent effect of experimental condition on time in Course 1. In Course 2, time to complete the task was extremely consistent despite experimental condition.
Figure 10 Collisions for the visually-impaired participant, in courses 1 and 2.
Figure 11 Time to complete the navigation task for the visually-impaired participant, in courses 1 and 2.
Discussion
One clear observation from our preliminary evaluations of the SPAM is the distinct difference between able-bodied, but blindfolded, individuals and individuals who are completely blind. The participant who is blind was much better at localizing the sound target and keeping track of his location in the course than any of the able-bodied participants. The blind participant also found it much easier to learn the layout of the course. One possible implication of these results is that the SPAM may be more useful for individuals who are newly visually impaired. Another possible implication is that the SPAM may be very useful in novel or frequently-changing environments, but not particularly useful in well-known, static environments.
Our preliminary evaluation of the SPAM demonstrates that the SPAM can increase the safety of visually-impaired manual wheelchair users. Of course, there is a large difference between a constrained laboratory environment and real-world environments, and much additional development and testing remains to be done. Our evaluation also identified several shortcomings. In particular, navigation assistance increased the time required to complete the navigation task. This was the result of an overly conservative obstacle avoidance algorithm, which slowed the SPAM more than necessary.
Our ability to control the SPAM was limited by our decision to retain the original electronics of the JWII hubs in place. This greatly simplified the development process, and allowed us to quickly produce a prototype that could be tested. The trade-off, however, was that our microprocessor and control software were not communicating directly with the motors within the hubs but were, instead, communicating with the JWII microprocessor and control software which controlled the motors. The control algorithms built into the JWII acted as a filter that made small adjustments in the speed and direction of the wheelchair difficult. This is why the motion of the SPAM was limited to straight forward, straight backward, and turning in place.
One unanticipated benefit of using power assist hubs which emerged during development was the ability to provide "haptic feedback" to the wheelchair user. As the SPAM approaches an obstacle, the hubs provide greater resistance. This allows the user to get an impression of the environment around the wheelchair through a series of forward pushes and rotations in place. In addition to individuals with visual impairments, this haptic feedback may also prove helpful for people with traumatic brain injuries.
Conclusion
The lessons learned from the first SPAM prototype are being incorporated into a second generation SPAM prototype (currently under development). Most importantly, the microprocessor used by the JWII hubs is being replaced with a new (programmable) microprocessor, which will allow the SPAM to provide much smoother and more nuanced control of the wheelchair. New enclosures have also been designed for the sensors that provide increased mounting flexibility, and have increased the number of modules. The additional sensor modules have forced us to abandon the case-based approach to obstacle avoidance, and alternative algorithms are being pursued.
Declaration of competing interests
AT Sciences has applied for a patent for the SPAM. AT Sciences will be submitting a Phase II SBIR proposal to the National Eye Institute based, in part, on the results contained in this manuscript. Dr. Simpson is not employed by, nor does he hold any stocks or shares in, AT Sciences. Dr. Simpson participated in this research through a subcontract negotiated between AT Sciences and the University of Pittsburgh.
Authors' contributions
RS, EL and RC conceived of the project and participated in its design and coordination. RS implemented the obstacle avoidance software, conducted the user trials and drafted the manuscript. SG, DD, SH and WA implemented the hardware for the SPAM, and interfaced the TattleTale microprocessor with the JWII electronics. VS reconstructed the SPAM to complete additional user testing. All authors read and approved the final manuscript.
Appendix
The American Federation for the Blind performed the following analysis using data from the 1994 and 1995 National Health Interview Survey on Disability, Phase I. Analysis used variables for being legally blind (location 422) and having serious difficulty seeing (location 401), using a manual wheelchair (location 524), an electric wheelchair (location 526), or a scooter (location 528), and age recode 2 (location 30). Data were extracted from phase I person files for each year, the design variables were recoded, the 2 years of data were combined, and the final weights were adjusted (wfta/2) in order to compute crosstabs in SUDAAN. Among all persons who are legally blind(1), 9.61% (se = 1.10) use a wheelchair. The 95% CI for this statistic is 7.41% to 11.81%. Among persons who are legally blind, (see Table 2) and 7.79% of individuals under the age of 65 (see Table 3), use a wheelchair.
Table 1 Use of Mobility Aids – All Ages
Legally Blind Serious Difficulty Seeing but not legally blind US Population
Totals 1057389.5 5315541 259994178
Uses Any Kind of Wheelchair (Manual, Electric or Scooter) 101565 279070.5 1668244.5
Percent 9.61 5.25 0.64
Table 2 Use of Mobility Aids – Ages 65 and Over
Legally Blind Serious Difficulty Seeing but not legally blind US Population
Totals 482935 2541294 31156585
Uses Any Kind of Wheelchair (Manual, Electric or Scooter) 56789 181005 924301
Percent 11.76 7.12 2.97
Table 3 Use of Mobility Aids – Under Age 65
Legally Blind Serious Difficulty Seeing but not legally blind US Population
Totals 574454 2774247 228837592
Uses Any Kind of Wheelchair (Manual, Electric or Scooter) 44776 98065 743943
Percent 7.79 3.53 0.33
Acknowledgements
This research is funded by a Phase I Small Business Innovation Research grant from the National Eye Institute (#1R43EY014490-01). The pushrims used in this research were donated to the University of Pittsburgh by Yamaha. Roland Frisch and Andrew Martin designed and fabricated the sensor bar and bump sensors for the footrests and rear of the wheelchair.
==== Refs
Cooper R Corfman T Fitzgerald S Boninger M Spaeth D Ammer W Arva J Performance Assessment of a Pushrim Activated Power Assisted Wheelchair IEEE Trans Cntrl Sys Tech 2001 10 121 126 10.1109/87.974345
Cooper R Fitzgerald S Boninger M Prins K Rentschler A Arva J O'Connor T Evaluation of a Pushrim Activated Power Assisted Wheelchair Archives of Physical Medicine and Rehabilitation 2001 82 702 708 11346854 10.1053/apmr.2001.20836
Levy CE Chow JW Tillman MD Hanson C Donohue T Mann WC Variable Ratio Power Assist Wheelchair Eases Wheeling Over a Variety of Terrains for Elders Archives of Physical Medicine and Rehabilitation 2004 85 104 112 14970977 10.1016/S0003-9993(03)00426-X
Woodbury R The Declining Disability of Older Americans Research Highlights in the Demography and Economics of Aging 1999 National Institute on Aging
Anonymous Guiding Blind People Who are Wheelchair Users 2002 London, England , Royal National Institute for the Blind
Pranghofer M Wheels and White Canes: Tips for Helping Blind Wheelchair Users Braille Monitor 1996
Greenbaum M Fernandes S Wainapel S User of a Motorized Wheelchair in Conjunction with a Guide Dog for the Legally Blind and Physically Disabled Archives of Physical Medicine and Rehabilitation 1998 79 216 217 9474006 10.1016/S0003-9993(98)90302-1
Nisbet P Craig J Odor P Aitken S `Smart' Wheelchairs for Mobility Training Technology and Disability 1995 5
Boy ES Teo CL Burdet E Collaborative wheelchair assistant: 30 September - 4 October; Lausanne, Switzerland. 2002 2 IEEE 1511 11516
Kelly D The enhancement of mobility for individuals who are both physically and visually disabled: Long Beach, CA. 1999 RESNA Press 227 229
|
16202136
|
PMC1262756
|
CC BY
|
2021-01-04 16:37:40
|
no
|
J Neuroengineering Rehabil. 2005 Oct 3; 2:30
|
utf-8
|
J Neuroeng Rehabil
| 2,005 |
10.1186/1743-0003-2-30
|
oa_comm
|
==== Front
Lipids Health DisLipids in Health and Disease1476-511XBioMed Central London 1476-511X-4-201619427510.1186/1476-511X-4-20ResearchCentile values for serum lipids and blood pressure for Asian Indian adolescents Madhavan Malini [email protected] Ravindra M [email protected] Anoop [email protected] Naval K [email protected] Vibha [email protected] Kalpana [email protected] Jasjeet S [email protected] Department of Medicine, All India Institute of Medical Sciences New Delhi-110029, India2 Department of Biostatistics, All India Institute of Medical Sciences New Delhi-110029, India3 Department of Biochemistry, All India Institute of Medical Sciences New Delhi-110029, India2005 29 9 2005 4 20 20 1 9 2005 29 9 2005 Copyright © 2005 Madhavan et al; licensee BioMed Central Ltd.2005Madhavan et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Reference data for plasma lipids and blood pressure are not available for Asian Indian adolescents. This study aimed to develop representative age- and sex- specific percentile reference data for serum lipids [total cholesterol (TC), triglycerides (TG), low density lipoprotein cholesterol (LDL-C), high density lipoprotein cholesterol (HDL-C), non-HDL cholesterol] and blood pressure for urban Asian Indian adolescents aged 14–18 years. The sample consisted of 680 boys and 521 girls aged 14–18 years from the cross-sectional population survey, Epidemiological Study of Adolescents and Young Adults (ESAY) for whom the data for serum lipid levels and blood pressure were recorded. Smoothed age- and sex- specific 5th, 10th, 25th, 50th, 75th, 85th, 90th and 95th percentiles where derived using LMS regression.
Results
Percentile-based reference data for serum lipids and blood pressure are presented for adolescent Asian Indian boys and girls for the first time. Asian Indian adolescents had lower levels of serum TC, LDL-C and HDL-C and higher TG than their counterparts in the USA. Interesting trends in TC and HDL-C levels where observed, which might reflect changes in dietary pattern and physical activity in this age group in India.
Conclusion
These reference data could be used to identify adolescents with an elevated risk of developing dyslipidemia, hypertension and cardiovascular disorders, to plan and implement preventive policies, and to study temporal trends.
Asian Indiansadolescentscentileslipidsblood pressure.
==== Body
Introduction
Atherosclerosis begins in childhood and progresses to coronary heart disease (CHD) in adults [1]. Aortic fatty streaks and fibrous plaques occur in children and adolescents [2-4]. The risk factors for atherosclerosis, such as dyslipidemia, hypertension, and insulin resistance may arise in children and adolescents, and contribute significantly to the acceleration of atherosclerosis [1].
Tracking of total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-C) levels from childhood to adults is well known. The National Cholesterol Education Program, Adult Treatment Panel III (NCEP, ATP III) identified LDL-C as the primary target for cholesterol-lowering therapy [5]. Elevated triglycerides (TG) and low levels of high-density lipoprotein cholesterol (HDL-C) are markers of atherogenic dyslipidemia [6-8]. Specifically, low levels of HDL-C is an independent risk factor for CHD [9]. Non-HDL cholesterol also shows strong correlation with CHD mortality [10] and has been recommended as a secondary target of therapy [5].
CHD is increasingly becoming a leading cause of death in India and other developing Asian countries. South Asians have premature and severe CHD as compared to white Caucasians [11]. Important contributory factors for CHD in South Asians are insulin resistance and resultant dyslipidemia [12]. We have recently reported that insulin resistance and low levels of HDL-C are common in Asian Indian adolescents, portending high risk for development of CHD in adults [13,14].
No representative percentiles reference data for plasma lipids and blood pressure in Asian Indian adolescents are currently available. These data are required for proper diagnosis and prevention of dyslipidemia, hypertension, and CHD. The purpose of this study was to develop representative age- and sex- specific percentile reference data for serum lipids (TC, LDL-C, HDL-C, TG, non-HDL cholesterol), and blood pressure [systolic blood pressure (SBP) and diastolic blood pressure (DBP)] for urban Asian Indian adolescents aged 14–18 years (y).
Methods
Subjects
The data were taken from an epidemiological study involving adolescents and young adults (aged 14–25 y) from schools and colleges located in southwest area of New Delhi. Multistage cluster sampling based on modified World Health Organization Expanded Program of Immunization Sampling Plan was used to collect a representative sample of adolescents and young adults [15]. First, two separate lists, one containing the names of schools and the other containing the names of colleges located in the defined area were prepared. A 'cluster' was defined as a school or a college. A total of 40 clusters were randomly selected from the two lists. The number of schools and colleges was determined based on the proportional allocation to ensure the representativeness of the sample with respect to clusters and socioeconomic strata. For both schools and colleges, a 'section' was considered as the primary sampling unit at the second stage of sampling. Subsequently from each of the schools/colleges, two to four sections were selected depending upon the number of the students in the section. All the students in the selected section were included in the study. Informed consent was obtained from the parents of the selected children. The institutional ethics committee approved the study. The data set consists of 680 boys and 521 girls aged 14 to 18 y. The sample characteristics are presented in table 1.
Table 1 Mean (SD) of lipid levels and blood pressure in urban Asian Indian adolescents 14–18 y of age
Age (y) n TC (mg/dl) LDL-C(mg/dl) TG (mg/dl) HDL-C (mg/dl) SBP (mmHg) DBP(mmHg)
Boys Girls Boys Girls Boys Girls Boys Girls Boys Girls Boys Girls Boys Girls
14 58 65 148.9 (21.0) 148.4 (26.0) 83.8 (20.3) 80.7 (26.5) 92.4 (37.8) 95.8 (28.5) 46.5 (8.5) 48.2 (10.3) 113.9 (10.3) 111.6 (9.6) 72.6 (7.6) 72.0 (7.0)
15 134 109 148.9 (27.3) 155.4 (23.6) 84.2 (27.3) 87.8 (23.9) 86.7 (30.5) 92.0 (28.0) 47.4 (7.9) 48.3 (8.3) 112.7 (9.8) 111.7 (8.9) 73.4 (7.4) 73.2 (7.0)
16 218 155 149.3 (24.0) 153.8 (21.4) 82.1 (25.8) 86.4 (23.2) 89.0 (29.7) 96.7 (28.0) 47.9 (6.1) 48.0 (8.5) 114.1 (9.5) 111.6 (9.3) 73.9 (7.0) 72.8 (6.6)
17 182 84 142.3 (27.9) 152.3 (21.4) 77.8 (27.9) 83.2 (24.7) 86.3 (32.1) 92.7 (29.0) 48.2 (6.9) 49.5 (7.8) 115.6 (9.2) 109.7 (9.9) 75.6 (6.4) 71.3 (7.3)
18 88 108 136.9 (24.1) 158.1 (23.8) 71.7 (22.1) 84.3 (24.7) 89.1 (34.0) 92.3 (28.9) 47.7 (7.0) 55.6 (7.6) 116.0 (9.0) 109.8 (8.2) 75.5 (6.7) 72.5 (7.2)
TC, Total cholesterol; LDL-C, Low-density lipoprotein cholesterol; TG, Triglycerides; HDL-C, High-density lipoprotein cholesterol; SBP, Systolic blood pressure; DBP, Diastolic blood pressure.
For conversion to mmol/L for TC, LDL-C and HDL-C, multiply by 0.02586
For conversion to mmol/L for TG, multiply by 0.01129
Lipoproteins
For estimation of lipoproteins, blood samples were obtained after an overnight fast of 10 hours. Levels of TC, TG, and HDL-C were estimated using commercially available kits (Randox Laboratory, San Francisco, CA, USA) on a semi-automated analyzer (das srl, palombara, Sabina, Italy) as described previously [13,16]. Value of LDL-C was calculated according to the Friedewald's equation [17]. The estimation of all lipids was rigorously quality controlled by a consultant biochemist (KL), and frequently checked with values of another reference laboratory. Inter-assay and intra-assay variability of estimations were kept at less than 5%.
Blood Pressure
Blood pressure was measured by a standard mercury sphygmomanometer (Industrial Electronic and Allied Products, Pune, India), after the subject had rested for 5 min in the sitting position, using the appropriate cuff size. Phase 5 Korotkoff sounds were taken for diastolic blood pressure categorization. In case of an abnormal blood pressure recording, another reading was obtained after 5 min rest and the mean of the two values was taken for the final record. The same physician measured the blood pressure using the same instrument for all the subjects. The mercury sphygmomanometer was periodically validated against a Hawksley Random Zero Sphygmomanometer (Hawksley, Lancing, Sussex, UK).
Statistical methods
The data were first examined for outliers. The LMS method was used to obtain smoothed centile curves for each of the anthropometric variables. The need for centile curve arises when the measurement is strongly dependent on some covariate, often age, so that the reference range changes with the covariate. The LMS method uses Box-Cox power transformation, which deals with the skewness present in the distribution of the anthropometric measurement and provides a way to normalize the measurement. The final centile curves are the result of smoothing three-age specific curves called L (lambda), M (mu) and S (sigma). The M and S curves correspond to the median and coefficient of variation of the measurement at each age whereas the L curve allows for the substantial age dependent skewness in the distribution of the measurement. The points on each centile curves are defined by the following formula:
M(1+LSz)1/L,
where L, M and S are the values of the fitted curves at each age, and z denotes the z score, i.e. the standard score with mean 0 and a standard deviation of 1, for the required centile, for example z = 1.645 for the 95th centile. The main assumption underlying the LMS method is that after Box-Cox power transformation the data at each age are normally distributed.
Descriptive statistics were computed using STATA 8.0 intercooled version (STATA Corporation, College Station Road, Houston, Texas) and the LMS regressions were performed using LMS Pro software (The Institute of Child Health, London).
Results
Smoothed age- and sex- specific cut-offs of serum TC, LDL-C, HDL-C, TG, non-HDL-C, SBP and DBP at the 5th, 10th, 25th, 50th, 75th, 85th, 90th, and 95th percentiles are presented in tables 2, 3, 4, 5, 6, 7, 8. Figures 1, 2, 3, 4, 5, 6, 7 present the smoothed percentile curves graphically for boys and figures 8, 9, 10, 11, 12, 13, 14 present the smoothed percentile curves graphically for girls.
Table 2 Smoothed age- and sex- specific serum total cholesterol (mmol/l) percentile values for urban Asian Indian adolescents 14–18 y of age.
Age 5th 10th 25th 50th 75th 85th 90th 95th
Boys
14 2.98 3.15 3.45 3.82 4.22 4.45 4.62 4.88
15 2.87 3.06 3.40 3.81 4.25 4.51 4.69 4.97
16 2.84 3.04 3.39 3.82 4.29 4.56 4.75 5.04
17 2.68 2.86 3.19 3.61 4.08 4.36 4.56 4.88
18 2.60 2.77 3.08 3.47 3.94 4.22 4.43 4.76
Girls
14 2.93 3.09 3.38 3.76 4.20 4.47 4.67 4.99
15 3.09 3.27 3.60 3.99 4.40 4.64 4.80 5.05
16 3.07 3.26 3.59 3.97 4.35 4.56 4.70 4.91
17 3.05 3.23 3.56 3.93 4.31 4.51 4.66 4.87
18 3.16 3.35 3.67 4.05 4.47 4.71 4.88 5.13
Table 3 Smoothed age- and sex- specific percentile values for low-density lipoprotein cholesterol (mmol/l) for urban Asian Indian adolescents 14–18 y of age.
Age 5th 10th 25th 50th 75th 85th 90th 95th
Boys
14 1.24 1.43 1.76 2.14 2.52 2.73 2.88 3.09
15 1.15 1.37 1.74 2.18 2.63 2.88 3.06 3.31
16 1.10 1.29 1.65 2.09 2.57 2.84 3.03 3.32
17 1.04 1.21 1.53 1.94 2.42 2.70 2.91 3.23
18 0.98 1.13 1.41 1.78 2.22 2.49 2.69 3.00
Girls
14 1.12 1.30 1.64 2.05 2.51 2.76 2.94 3.22
15 1.24 1.46 1.82 2.24 2.68 2.92 3.09 3.34
16 1.23 1.45 1.81 2.22 2.64 2.86 3.01 3.24
17 1.18 1.39 1.75 2.16 2.57 2.80 2.95 3.19
18 1.22 1.40 1.73 2.14 2.58 2.84 3.01 3.29
Table 4 Smoothed age- and sex- specific percentile values for high-density lipoprotein cholesterol (mmol/l) for urban Asian Indian adolescents 14–18 y of age.
Age 5th 10th 25th 50th 75th 85th 90th 95th
Boys
14 0.86 0.93 1.05 1.20 1.35 1.43 1.48 1.57
15 0.91 0.98 1.10 1.23 1.36 1.43 1.48 1.55
16 0.96 1.02 1.12 1.24 1.35 1.41 1.46 1.52
17 0.98 1.04 1.13 1.24 1.36 1.42 1.47 1.54
18 0.98 1.03 1.11 1.21 1.33 1.41 1.47 1.56
Girls
14 0.86 0.93 1.07 1.24 1.41 1.51 1.58 1.68
15 0.88 0.96 1.09 1.24 1.40 1.48 1.54 1.63
16 0.90 0.97 1.10 1.24 1.38 1.46 1.52 1.60
17 0.96 1.03 1.14 1.28 1.41 1.49 1.54 1.62
18 1.12 1.19 1.30 1.43 1.57 1.64 1.70 1.77
Table 5 Smoothed age- and sex- specific percentile values for serum triglycerides (mmol/l) for urban Asian Indian adolescents 14–18 y of age.
Age(y) 5th 10th 25th 50th 75th 85th 90th 95th
Boys
14 0.55 0.62 0.75 0.95 1.22 1.40 1.54 1.79
15 0.53 0.60 0.73 0.92 1.16 1.32 1.44 1.63
16 0.54 0.61 0.76 0.95 1.20 1.35 1.46 1.65
17 0.51 0.59 0.73 0.92 1.16 1.31 1.43 1.61
18 0.54 0.60 0.74 0.94 1.19 1.36 1.49 1.71
Girls
14 0.62 0.70 0.85 1.04 1.27 1.41 1.51 1.66
15 0.60 0.68 0.82 1.00 1.22 1.35 1.45 1.60
16 0.64 0.71 0.86 1.05 1.28 1.42 1.52 1.68
17 0.60 0.67 0.82 1.01 1.23 1.37 1.47 1.62
18 0.58 0.66 0.81 1.00 1.23 1.37 1.47 1.64
Table 6 Smoothed age- and sex- specific percentile values for non-HDL cholesterol (mmol/l) for urban Asian Indian adolescents 14–18 y of age.
Age(y) 5th 10th 25th 50th 75th 85th 90th 95th
Boys
14 1.71 1.91 2.24 2.62 3.01 3.23 3.38 3.60
15 1.60 1.81 2.19 2.63 3.08 3.34 3.52 3.78
16 1.56 1.76 2.13 2.57 3.05 3.33 3.52 3.82
17 1.46 1.63 1.95 2.36 2.83 3.12 3.32 3.64
18 1.42 1.56 1.85 2.22 2.68 2.97 3.19 3.53
Girls
14 1.66 1.84 2.17 2.56 2.98 3.23 3.40 3.65
15 1.79 1.99 2.35 2.75 3.17 3.40 3.56 3.80
16 1.77 1.98 2.33 2.73 3.13 3.35 3.50 3.72
17 1.70 1.90 2.25 2.64 3.06 3.28 3.44 3.67
18 1.67 1.86 2.20 2.61 3.06 3.32 3.50 3.78
Table 7 Smoothed age- and sex- specific percentile values for systolic blood pressure (mm Hg) for urban Asian Indian adolescents 14–18 y of age.
Age(y) 5th 10th 25th 50th 75th 85th 90th 95th
Boys
14 99 103 108 115 121 125 127 131
15 98 101 107 113 119 123 125 128
16 98 102 108 114 120 123 125 129
17 100 103 109 115 121 125 127 130
18 102 105 110 116 122 125 128 131
Girls
14 96 100 105 112 118 121 123 127
15 97 100 105 111 118 121 123 127
16 97 100 105 111 118 121 124 127
17 95 98 103 109 116 119 121 125
18 96 99 104 110 116 119 121 124
Table 8 Smoothed age- and sex- specific percentile values for diastolic blood pressure (mm Hg) for urban Asian Indian adolescents 14–18 y of age.
Age(y) 5th 10th 25th 50th 75th 85th 90th 95th
Boys
14 61 64 68 73 78 81 82 85
15 62 64 69 74 79 81 83 85
16 62 65 69 74 78 81 82 85
17 64 67 71 75 80 82 84 86
18 64 67 71 76 80 82 84 86
Girls
14 60 63 68 72 77 79 81 83
15 62 64 69 73 78 80 82 84
16 62 64 68 73 77 80 81 84
17 60 63 67 71 76 78 80 83
18 61 63 68 72 77 80 82 85
Figure 1 Smoothed percentile curves for the 5th, 10th, 25th, 50th, 75th, 85th, 90th and 95th percentiles of total cholesterol for urban boys 14–18 years of age.
Figure 2 Smoothed percentile curves for the 5th, 10th, 25th, 50th, 75th, 85th, 90th and 95th percentiles of serum triglycerides for urban boys 14–18 years of age.
Figure 3 Smoothed percentile curves for the 5th, 10th, 25th, 50th, 75th, 85th, 90th and 95th percentiles of low-density lipoprotein cholesterol for urban boys 14–18 years of age.
Figure 4 Smoothed percentile curves for the 5th, 10th, 25th, 50th, 75th, 85th, 90th and 95th percentiles of high-density lipoprotein cholesterol for urban boys 14–18 years of age.
Figure 5 Smoothed percentile curves for the 5th, 10th, 25th, 50th, 75th, 85th, 90th and 95th percentiles of non high-density lipoprotein cholesterol for urban boys 14–18 years of age.
Figure 6 Smoothed percentile curves for the 5th, 10th, 25th, 50th, 75th, 85th, 90th and 95th percentiles of systolic blood pressure for urban boys 14–18 years of age.
Figure 7 Smoothed percentile curves for the 5th, 10th, 25th, 50th, 75th, 85th, 90th and 95th percentiles of diastolic blood pressure for urban boys 14–18 years of age.
Figure 8 Smoothed percentile curves for the 5th, 10th, 25th, 50th, 75th, 85th, 90th and 95th percentiles of total cholesterol for urban girls 14–18 years of age.
Figure 9 Smoothed percentile curves for the 5th, 10th, 25th, 50th, 75th, 85th, 90th and 95th percentiles of serum triglycerides for urban girls 14–18 years of age.
Figure 10 Smoothed percentile curves for the 5th, 10th, 25th, 50th, 75th, 85th, 90th and 95th percentiles of low-density lipoprotein cholesterol for urban girls 14–18 years of age.
Figure 11 Smoothed percentile curves for the 5th, 10th, 25th, 50th, 75th, 85th, 90th and 95th percentiles of high-density lipoprotein cholesterol for urban girls 14–18 years of age.
Figure 12 Smoothed percentile curves for the 5th, 10th, 25th, 50th, 75th, 85th, 90th and 95th percentiles of non high-density lipoprotein cholesterol for urban girls 14–18 years of age.
Figure 13 Smoothed percentile curves for the 5th, 10th, 25th, 50th, 75th, 85th, 90th and 95th percentiles of systolic blood pressure for urban girls 14–18 years of age.
Figure 14 Smoothed percentile curves for the 5th, 10th, 25th, 50th, 75th, 85th, 90th and 95th percentiles of diastolic blood pressure for urban girls 14–18 years of age.
Boys aged 14 to 16 y showed a minimal increase in mean serum TC followed by a sharp decline at ages 17 and 18 y. Serum TC in adolescent girls, on the other hand, showed a relative minimum at 16 and17 y. Overall girls had higher mean TC than boys. The 95th percentile for TC in girls aged 18 y was the highest estimate for all ages and both sexes. Mean LDL-C in both boys and girls aged 14 to18 y was highest at 15 y and decreased with age beyond 15 y. Girls had higher mean LDL-C than boys at all ages. The serum non-HDL-C levels followed a similar trend. The 95th and 75th percentile of TC and LDL-C (which define high and borderline elevated cholesterol respectively in children and adolescents) are of special note [18]. Mean HDL-C was relatively constant among 14 to 18 y old boys and girls except that in girls mean HDL-C levels increased 1.5 mmol/l from 17 to 18 y. The 5th percentile, which would define low HDL-C in these adolescents, is also represented in table 4. Mean serum TG levels did not vary significantly among 15 to 18 year olds except for a small increase at 16 years for both sexes. The mean TG was generally higher in girls at all ages.
The mean SBP and DBP in both sexes did not vary significantly between the different age groups. Boys aged 16 to 18 y had a greater systolic and diastolic BP than girls.
Discussion
Representative age- and sex- specific reference data on the distribution of lipids and blood pressure in urban Asian Indian adolescents aged 14 to 18 y have been provided for the first time. This data can be used as a reference for the urban Indian population for comparison to other studies, to measure the progress in health of Indian adolescents in the future, and to plan and implement intervention programs for the prevention of cardiovascular disease.
There are interesting trends in the TC and HDL-C levels in urban Indian adolescent population, which may be significantly accounted for by the marked diet and lifestyle changes in the 15 to 18 y age group. 15 to 17 y old adolescents in India are subjected to strenuous and important examinations leading to adverse diet changes and inactivity during this period [19], which may have resulted in high TC and lower HDL-C levels in these adolescents. An average urban Indian adolescent enters college at the age of 17 to 18 y, when physical activity and diet are likely to improve which could lead to lower TC and higher HDL-C levels observed in this age group.
The serum lipid levels were compared with Lipid Research Clinics (LRC) and NHANES III data from the USA [20-22]. TC and LDL-C levels were consistently lower in Asian Indians than in the USA in both sexes and across the age groups studied. Secular data in TC levels in adolescents in the USA, showed a downward trend from 1966–1970 (NHES III) to 1988–1994 (NHANES III) [22]. However, TC levels in Asian Indian adolescents were lower than those observed in the USA in the more recent NHANES III data.
The data show that TG levels in Asian Indian adolescents were lower than that observed in the NHANES III study at percentiles higher than the 75th, despite being higher at the lower percentiles [22]. Asian Indian adolescents had higher HDL-C levels at lower percentiles than their counterparts in the USA [22] and lower HDL-C levels at percentiles greater than the 50th. Although the TC and LDL-C levels are lower in urban Asian Indian adolescents than in other populations, however, the fasting hyperinsulinemia (fasting serum insulin level >20 microunits/ml)[23], indicative of insulin resistance, was seen in 50% of girls and 19% of boys, in addition to the presence of other cardiovascular risk factors such as high C-reactive protein (CRP) levels [13,14,16]. These data indicate that although lipids may be important in pathogenesis of CAD in adult Asian Indians, it may not be always possible to ascertain the risk based on lipid levels alone during childhood. The cardiovascular risk in young Asian Indians could be better estimated by assessing several other biochemical factors (insulin, CRP, non-esterified fatty acids) in addition to lipids.
Conflict of Interest
The author(s) declare that they have no competing interests.
Acknowledgements
We thank Dr. Tim J Cole of The Institute of Child Health, London, for providing the LMS program and for useful suggestions. The study was funded by a grant from the Department of Science and Technology, Ministry of Science and Technology, Government of India, New Delhi. The authors are thankful to Ministry of Education, Government of New Delhi for their assistance in conducting the study. The cooperation of the children who took part in the study, and the help extended by the principals, teachers, and staff of the various schools and colleges where the study was conducted is greatly appreciated.
==== Refs
Misra A Risk factors for atherosclerosis in young individuals J Cardiovasc Risk 2000 7 215 229 11006892
Strong JP The natural history of atherosclerosis in childhood Ann N Y Acad Sci 1991 623 9 15 2042878
McGill HCJ McMahan CA Malcom GT Oalmann MC Strong JP Effects of serum lipoproteins and smoking on atherosclerosis in young men and women. The PDAY Research Group. Pathobiological Determinants of Atherosclerosis in Youth Arterioscler Thromb Vasc Biol 1997 17 95 106 9012643
McGill HCJ McMahan CA Determinants of atherosclerosis in the young. Pathobiological Determinants of Atherosclerosis in Youth (PDAY) Research Group Am J Cardiol 1998 82 30T 36T 9860371 10.1016/S0002-9149(98)00720-6
Executive Summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA 2001 285 2486 2497 11368702 10.1001/jama.285.19.2486
Austin MA Hokanson JE Edwards KL Hypertriglyceridemia as a cardiovascular risk factor Am J Cardiol 1998 81 7B 12B 9462597 10.1016/S0002-9149(98)00031-9
Assmann G Schulte H Funke H von Eckardstein A The emergence of triglycerides as a significant independent risk factor in coronary artery disease Eur Heart J 1998 19 Suppl M M8 14 9821011
Grundy SM Hypertriglyceridemia, atherogenic dyslipidemia, and the metabolic syndrome Am J Cardiol 1998 81 18B 25B 9526809 10.1016/S0002-9149(98)00033-2
Gordon DJ Probstfield JL Garrison RJ Neaton JD Castelli WP Knoke JD Jacobs DRJ Bangdiwala S Tyroler HA High-density lipoprotein cholesterol and cardiovascular disease. Four prospective American studies Circulation 1989 79 8 15 2642759
Cui Y Blumenthal RS Flaws JA Whiteman MK Langenberg P Bachorik PS Bush TL Non-high-density lipoprotein cholesterol level as a predictor of cardiovascular disease mortality Arch Intern Med 2001 161 1413 1419 11386890 10.1001/archinte.161.11.1413
Reddy KS Yusuf S Emerging epidemic of cardiovascular disease in developing countries Circulation 1998 97 596 601 9494031
Misra A Vikram NK Insulin resistance syndrome (metabolic syndrome) and obesity in Asian Indians: evidence and implications Nutrition 2004 20 482 491 15105039 10.1016/j.nut.2004.01.020
Misra A Vikram NK Arya S Pandey RM Dhingra V Chatterjee A Dwivedi M Sharma R Luthra K Guleria R Talwar KK High prevalence of insulin resistance in postpubertal Asian Indian children is associated with adverse truncal body fat patterning, abdominal adiposity and excess body fat Int J Obes Relat Metab Disord 2004 28 1217 1226 15314636 10.1038/sj.ijo.0802704
Vikram NK Misra A Pandey RM Dwivedi M Luthra K Adiponectin levels in postpubertal asian Indian adolescents: Relationships with insulin resistance and C-reactive protein Metabolism 2004 53 1336 1341 15375791 10.1016/j.metabol.2004.05.010
Lemeshow S Stroh G Sampling techniques for evaluating health parameters in developing countries National Academy Press, Washington, DC 1988
Vikram NK Misra A Dwivedi M Sharma R Pandey RM Luthra K Chatterjee A Dhingra V Jailkhani BL Talwar KK Guleria R Correlations of C-reactive protein levels with anthropometric profile, percentage of body fat and lipids in healthy adolescents and young adults in urban North India Atherosclerosis 2003 168 305 313 12801614 10.1016/S0021-9150(03)00096-0
Friedewald WT Levy RI Fredrikson DS Estimation of the concentration of low-density lipoprotein cholesterol in plasma without use of the preparative ultracentrifuge. Clin Chem 1972 18 499 502 4337382
National Cholesterol Education Program (NCEP): highlights of the report of the Expert Panel on Blood Cholesterol Levels in Children and Adolescents Pediatrics 1992 89 495 501 1741227
Dhingra V Chatterjee A Guleria R Sharma R Pandey RM Talwar KK Misra A Adverse physical activity pattern in urban adolescents J Assoc Physicians India 2002 50 1521
American Academy of Pediatrics. National Cholesterol Education Program: Report of the Expert Panel on Blood Cholesterol Levels in Children and Adolescents Pediatrics 1992 89 525 584 1538956
Plasma lipid distributions in selected North American populations: the Lipid Research Clinics Program Prevalence Study. The Lipid Research Clinics Program Epidemiology Committee Circulation 1979 60 427 439 312704
Hickman TB Briefel RR Carroll MD Rifkind BM Cleeman JI Maurer KR Johnson CL Distributions and trends of serum lipid levels among United States children and adolescents ages 4-19 years: data from the Third National Health and Nutrition Examination Survey Prev Med 1998 27 879 890 9922071 10.1006/pmed.1998.0376
Alberti KG Zimmet PZ Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus provisional report of a WHO consultation Diabet Med 1998 15 539 553 9686693 10.1002/(SICI)1096-9136(199807)15:7<539::AID-DIA668>3.0.CO;2-S
|
16194275
|
PMC1262757
|
CC BY
|
2021-01-04 16:39:19
|
no
|
Lipids Health Dis. 2005 Sep 29; 4:20
|
utf-8
|
Lipids Health Dis
| 2,005 |
10.1186/1476-511X-4-20
|
oa_comm
|
==== Front
Malar JMalaria Journal1475-2875BioMed Central London 1475-2875-4-451617908510.1186/1475-2875-4-45ResearchSafety of the methylene blue plus chloroquine combination in the treatment of uncomplicated falciparum malaria in young children of Burkina Faso [ISRCTN27290841] Meissner Peter E [email protected] Germain [email protected] Steffen [email protected] Boubacar [email protected] Ulrich [email protected] Jens [email protected] Wolfgang [email protected] Albrecht [email protected] Mamadou [email protected] Théophile [email protected] Ingeborg [email protected] Gerd [email protected] Jürgen [email protected] Klaus-Dieter [email protected] Heiner [email protected]é Bocar [email protected]üller Olaf [email protected] Department of Tropical Hygiene and Public Health, Ruprecht-Karls-University, INF 324, 69120 Heidelberg,, Germany2 Department of Paediatrics IV Neonatology, Ruprecht-Karls-University, Heidelberg, Germany3 Centre de Recherche en Santé de Nouna, Burkina Faso, POB 02, Nouna, Burkina Faso4 Institute of Medical Biometrics and Informatics, Ruprecht-Karls-University, Heidelberg, Germany5 Institute of Medical Information, Biometry, and Epidemiology, Ludwig-Maximilians-University, Munich, Germany6 Department of Internal Medicine VI, Clinical Pharmacology and Pharmacoepidemiology, Ruprecht-Karls University, Heidelberg, Germany7 DSM Fine Chemicals, Linz, Austria8 Biochemistry Center, Ruprecht-Karls-University, Heidelberg, Germany2005 22 9 2005 4 45 45 7 7 2005 22 9 2005 Copyright © 2005 Meissner et al; licensee BioMed Central Ltd.2005Meissner et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Safe, effective and affordable drug combinations against falciparum malaria are urgently needed for the poor populations in malaria endemic countries. Methylene blue (MB) combined with chloroquine (CQ) has been considered as one promising new regimen.
Objectives
The primary objective of this study was to evaluate the safety of CQ-MB in African children with uncomplicated falciparum malaria. Secondary objectives were to assess the efficacy and the acceptance of CQ-MB in a rural population of West Africa.
Methods
In this hospital-based randomized controlled trial, 226 children (6–59 months) with uncomplicated falciparum malaria were treated in Burkina Faso. The children were 4:1 randomized to CQ-MB (n = 181; 25 mg/kg CQ and 12 mg/kg MB over three days) or CQ (n = 45; 25 mg/kg over three days) respectively. The primary outcome was the incidence of severe haemolysis or other serious adverse events (SAEs). Efficacy outcomes were defined according to the WHO 2003 classification system. Patients were hospitalized for four days and followed up until day 14.
Results
No differences in the incidence of SAEs and other adverse events were observed between children treated with CQ-MB (including 24 cases of G6PD deficiency) compared to children treated with CQ. There was no case of severe haemolysis and also no significant difference in mean haemoglobin between study groups. Treatment failure rates were 53.7% (95% CI [37.4%; 69.3%]) in the CQ group compared to 44.0% (95% CI [36.3%; 51.9%]) in the CQ-MB group.
Conclusion
MB is safe for the treatment of uncomplicated falciparum malaria, even in G6PD deficient African children. However, the efficacy of the CQ-MB combination has not been sufficient at the MB dose used in this study. Future studies need to assess the efficacy of MB at higher doses and in combination with appropriate partner drugs.
==== Body
Introduction
Malaria is globally responsible for 1.5 – 2.7 million deaths per year and remains a leading cause of mortality in children less than five years of age, especially in African countries [38,5]. Only a few safe and effective antimalarial drugs are presently affordable for the great majority of sub-Saharan African (SSA) populations [4,40]. The increasing resistance of Plasmodium falciparum to chloroquine and other well-known antimalarials like sulphadoxine/pyrimethamine demonstrates the urgency of developing new drugs against malaria [36]. The combination of the drugs chloroquine (CQ) and methylene blue (MB) appears to be a promising new regimen in regions where CQ sensitivity prevails [32].
MB is a registered drug for the treatment of methaemoglobinemia and is also applied for cancer treatment, mainly at i.v. dosages of 1–2 mg/kg [18]. MB was already successfully used over 100 years ago for the treatment of malaria, even in children [14,9,11]. It was forgotten when other drugs (e.g. chloroquine) were introduced to the market. In recent years in vitro experiments have confirmed the high antimalarial potency of MB [1]. MB specifically inhibits the glutathione reductase of the malarial parasite and it prevents the polymerization of haem into haemozoin similar to 4-amino-quinoline antimalarials [31,32]. It has, thus, the potential to reverse CQ resistance [10,22,13].
There is some concern that the application of MB could be followed by haemolysis in glucose-6-phosphate dehydrogenase (G6PD) deficient individuals in malaria-endemic regions. This is likely to be not clinically relevant in West Africa where the milder form (class III) of G6PD deficiency dominates [12]. Moreover, the safety of CQ-MB is supported by the results of two studies in adult populations in 2003. The combination was well tolerated in healthy G6PD sufficient adult males and females from Germany and in healthy G6PD deficient adult males from Burkina Faso [30,20].
The primary objective of the present trial was to study the safety of CQ-MB in young West African children with uncomplicated falciparum malaria. Secondary objectives were to assess efficacy and acceptance of the combination.
Materials and methods
Study area and patients
The study was conducted in the district hospital of Nouna in the province of Kossi in north-western Burkina Faso. Nouna town has about 20,000 inhabitants of different ethnic groups. Formal health services in the Kossi province are restricted to the district hospital and to a small number of village-based health centres. Therefore, access to health services is limited, particularly during the rainy season [26]. The area is highly endemic for P. falciparum malaria [25]. Most malaria transmission and disease takes place during or shortly after the rainy season, which lasts from July-October. CQ is the first-line treatment for uncomplicated malaria in Burkina Faso. Day 14 CQ clinical failure rates in the villages surrounding Nouna town were shown to be around 10% in recent years [27].
The study was explained to the population of Nouna town during a series of community meetings prior to selection of study children. In October and November 2003, consecutive children aged 6–59 months presenting with uncomplicated falciparum malaria (axillary temperature ≥ 37.5°C and ≥ 2,000 P. falciparum asexual parasites per μl blood) at the outpatient department of Nouna district hospital and whose parents or caretaker had given informed consent were enrolled in the study. Most study children were from the town of Nouna, only a few children were from villages close to Nouna. Children with complicated or severe malaria (repeated vomiting, seizures or other neurological impairment), anaemia (haemoglobin <8 g/dl or haematocrit <24%) or any other apparent significant disease (e.g. pneumonia, meningitis, hepatitis, severe diarrhoea, measles, severe malnutrition) were excluded from the trial.
Study design and procedures
The study was designed as a sex-stratified (male:female, 4:1) parallel group randomized controlled trial (RCT). In each sex stratum, children were block-randomized by envelopes to receive either treatment with CQ or CQ-MB by computer-generated, randomly permutated codes. The randomization ratio of CQ-MB : CQ was 4:1. The unbalanced stratification and randomization scheme was justified by the need to achieve a high number of G6PD deficient children treated with CQ-MB. The prevalence of G6PD deficiency in boys under five years is 18%, while the prevalence in girls is below 4% [7]. The study was open label, with blinding only for the laboratory technicians responsible for parasite determination in blood smears.
All study children received a standard total dose of 25 mg/kg of CQ syrup (10 mg/ml) over a period of three days (first and second day: 10 mg/kg, third day: 5 mg/kg). Chloroquine was taken from the quality-controlled essential drug stock of the Ministry of Health. The CQ-MB group additionally received orally a 0.5% MB solution (4 mg/kg/day) for three days, divided into morning and evening doses (produced by Mayrhofer Pharmazeutika / Linz, Austria).
All study children were hospitalized for four days, and study medications were given according to weight-based tables by a study nurse, directly supervised by a study physician. Axillary temperatures were taken every six hours during the time of hospitalization. At least once during the time of hospitalization, the colour of the urine was checked to assess MB compliance through visual observation. In case of vomiting within half an hour after the study medication, the medication was repeated. All drugs were allowed as concomitant treatments, except dapsone and other sulfones, acetanilide and phenacetin, nalidixic acid, niridazole, nitrofurantoins and sulphonamides (themselves known to cause haemolysis in certain types of inherited G6PD deficiency) [12]. All children having fever ≥ 38.5°C received standard doses of paracetamol (10 mg/kg every six hours) until symptoms subsided.
Study participants were examined by the study physicians twice daily until discharge from hospital and again after 14 days. Treatment failures were treated according to national guidelines with standard doses of sulfadoxine-pyrimethamine (currently still the second-line treatment in Burkina Faso) or with quinine i.m. or i.v. if indicated. Acceptance of MB was evaluated at day 14 by asking the caretakers standardized questions about adverse effects.
At inclusion, every day in hospital and on day 14 a blood sample was taken and malaria parasitaemia (thin and thick blood smears) as well as haemoglobin (in case of venous blood sampling) and haematocrit values were measured. After Giemsa-staining, all blood smears were examined by two experienced laboratory technicians blinded to the group allocation. In case of disagreement, a third laboratory technician examined the slides. Thick and thin blood films were analysed for the species-specific asexual parasite density per μl by counting against 200 white blood cells and multiplying by 50. Slides were declared negative if no parasites were seen in 400 fields on the thick film. Re-examination of a 10% random sample of blood films at the laboratory of the Heidelberg School of Tropical Medicine showed a 95% concordance of malaria diagnoses. Furthermore, serum creatinine levels were monitored with a spectrophotometer (Ultrospec 1000®) during the time of hospitalization.
The G6PD status was determined at inclusion using the NADPH fluorescence test of Beutler and Mitchell [2] in miniaturized form on paper (NFP-test) [6,7]. The G6PD results were later validated by PCR. Sequences flanking the putative mutations were amplified in two separate nested PCR assays [17]. The alleles were distinguished by restriction fragment length polymorphism. Finally, genetic resistance to CQ in the P. falciparum isolates was estimated by analysing the prevalence of the mutation K76T in the Pfcrt gene product according to Djimde et al. [8]. Both, G6PD PCR and CQ resistance marker diagnosis were conducted at the Tropical Institute in Berlin (Germany).
The primary study endpoint was safety, assessed by the frequency and proportion of children with at least one serious adverse event that could be possibly, probably or definitely related to the drug. Secondary endpoints were efficacy outcomes, judged by the incidence of early treatment failures (ETF) on day four and late clinical failures (LCF), clinical failure (CF = ETF+LCF) and late parasitological failures (LPF) on day 14; fever clearance time, parasite clearance time, change in haematocrit after four days compared to baseline, incidence of observed or self-reported adverse events over the 14 days follow-up period and monitoring of concomitant drug intake. In addition, the impact of G6PD status based on the NFP-test and PCR genotyping on safety parameters was investigated. An adverse event was defined according to internationally established principles for Good Clinical Practice (GCP).
Statistical methods
With a study population of 200 children, 160 were randomized into the CQ-MB group and with high probability (80%) at least 22 G6PD deficient children will be exposed to MB. In order to reject the Null hypothesis of a high haemolysis risk (>20%) on a significance level of 5% and with a power of 90% if the true risk is small (<2%), 22 G6PD deficient children were to be included. A total of 225 patients was planned and a 10% drop-out rate was assumed. The WHO definitions for ETF, LCF and LPF were applied [39]. Study days were defined as 24 h intervals after first drug intake. Losses to follow-up and drop-outs due to other reasons were considered as ETF or LCF, depending on the time of drop-out in an intention-to-treat manner. All tests used apart from the primary analysis have exploratory character. The (continuity adjusted) Chi square test (Chi) was used to compare rates, the non-parametric Wilcoxon-Mann-Whitney test (WMW) to compare metric or ordinal data. If possible, estimates and the corresponding 95% confidence interval are given. The statistical calculation was performed with SAS release 8.02 (SAS® Institute Inc, Cary, NC, USA).
Roughly 5% mortality over the first five years of childhood in malaria endemic areas of SSA is directly caused by malaria [34]. This implies an annual malaria specific mortality risk of 1%. From a public health-related risk benefit calculation it follows that a risk for life-threatening haemolysis above 10% is not acceptable for a first-line antimalarial drug, because it could increase the overall childhood mortality. Assuming a G6PD deficiency prevalence of at least 10% in the CRSN study area and the fact that every child in the CRSN study area will be treated at least once per year with a malaria drug which eliminates malaria related mortality, but implies life-threatening haemolysis in less than 10% of G6PD deficient children, the annual population risk would be below 1% (0.1 × 0.1).
Ethical aspects
The trial was conducted in accordance with local law, the internationally established principles for Good Clinical Practice, which had their origin in the Declaration of Helsinki of the World Medical Association, and in accordance with the "Note for Guidance on Clinical Investigation of Medicinal Products in Children". Data collection and analysis followed established quality principles. The protocol was approved by the Ethics Committee of the Medical Faculty of Heidelberg University and the local Ethics Committee in Burkina Faso. The safety of the trial was also monitored by a data safety monitoring board (DSMB).
After having received detailed information from the study physician about all risks and benefits of the study through translation of a detailed research consent form into the local language, caretakers were asked for their written consent. They were clearly informed that they could withdraw from the study at any time and without disadvantage. A standard blood transfusion service was available at the hospital, and study physicians and emergency medications were available 24 hours per day. For the study participants not only malaria treatment but all treatments were free of charge. Furthermore, all children in the specified age group who were presented to the hospital during the study period for other conditions besides malaria also received free treatment.
Results
Study group characteristics
229 children with uncomplicated malaria were enrolled (Figure 1). Three boys in group CQ-MB were excluded on day one, one left for family reasons before application of the first medication, two refused to take the first dose of MB. The following analysis is based on the remaining 226 children. This is the full analysis set (FAS) for the intention to treat analysis, 45 receiving CQ and 181 CQ-MB. In group CQ and CQ-MB respectively 41/45 and 166/181 children had taken the study drugs as per protocol (= PP, Figure 1). There were no statistically significant differences in sex ratio, age (pWMW = 0.801), weight (pWMW = 0.898), creatinine, malaria parasitaemia and the prevalence of the genetic resistance marker pfcrt K76T of the malaria parasites (pChi = 0.793) between the two groups (Table 1). There was no difference in the haemoglobin (pWMW = 0.864) but in haematocrit (pWMW = 0.007) values between the treatment groups at baseline (Table 1). G6PD status was assessed in all 226 children. Overall, 30/226 children (13.3%, 95% CI [9.1%, 18.4%]) were diagnosed as G6PD deficient using the NFP test (six in the CQ group, 24 in the CQ-MB group) (Table 1). There was good agreement between the NFP test and the PCR results; these findings are published separately [23].
Figure 1 Flow chart of study patients.
Table 1 Baseline characteristics by study group
Characteristic CQ (N = 45) CQ-MB (N = 181) Total (N = 226)
Sex
- male 36(80.0%) 145(80.1%) 181(80.1%)
- female 9(20.0%) 36(19.9%) 45(19.9%)
Age [months]
- Mean ± SD 30.0 ± 13.1 29.5 ± 14.9 29.6 ± 14.5
- Median 28.0 27.0 27.5
Weight [kg]
- Mean ± SD 10.5 ± 2.8 10.6 ± 2.8 10.6 ± 2.8
- Median 10.0 10.0 10.0
G6PD deficiency (NFP-test)
- No 39 (86.7%) 157 (86.7%) 196 (86.7%)
- Yes 6 (13.3%) 24 (13.3%) 30 (13.3%)
pfcrt gene
- wildtype 11 (24.4%) 40 (22.6%) 51 (23.0%)
- K76T mutant 34 (75.6%) 137 (77.4%) 171 (77.0%)
- Missing data 0 4 4
Haemoglobin [g/dl]
- Mean ± SD 10.3 ± 1.3 10.4 ± 1.4 10.4 ± 1.4
- Median 10.2 10.2 10.2
Haematocrit [%]
- Mean ± SD 27.9 ± 2.6 29.4 ± 3.3 29.1 ± 3.2
- Median 28.0 30.0 28.0
Creatinine [μmol/l]
- Mean ± SD 75.3 ± 15.2 77.2 ± 14.5 76.8 ± 14.6
- Median 74.4 76.4 75.3
P. falciparum parasites/μl log10 values
- Mean ± SD 4.28 ± 0.54 4.35 ± 0.54 4.33 ± 0.54
- Median 4.32 4.38 4.36
- Min, Max 3.40, 5.29 3.08, 5.71 3.08, 5.71
Compliance with study drugs
The compliance is illustrated in Figure 1. In some cases, particularly in the CQ-MB group, the drug administration needed repeating after being initially rejected by the children (1/45 CQ, 35/181 CQ-MB). Application of the bitter-tasting 0.5% MB solution was more difficult in children <2 years compared to older children (26/61 versus 9/120, pChi < 0.0001).
Safety of study drugs
There was no case of severe haemolysis within the group of 24 G6PD deficient children who received MB (one-sided exact 95% CI [0%, 11.73%]). Combining these results with the pilot study in adult males [20], no haemolysis was observed in 98 G6PD-deficient subjects under CQ-MB. This updates the risk estimate to 3% (upper confidence bound). One patient in the G6PD sufficient CQ-MB group had a serious adverse event (SAE; prolonged hospitalization) which was unrelated to the study medication. The haemoglobin of this 21 months old girl with an initially high parasitaemia of 193,000/μl and haemorrhagic diarrhoea dropped within four days from 11.8 g/dl to 6.6 g/dl. There was no need for blood transfusion. With quinine and antibiotics she quickly recovered within five days after admission to hospital. There was only one febrile child with severe jaundice (group CQ-MB), later diagnosed to have acute hepatitis A. There was also no marked difference in the incidence of adverse events between the two groups (Table 2). Only pruritus was observed with a higher incidence in the CQ group compared to the CQ-MB group (8/45 versus 9/181; pChi = 0.017). Until the end of hospitalization there was no significant increase of serum creatinine in the CQ-MB versus the CQ group.
Table 2 Adverse events according to MedDRA SOC by study group
CQ (N = 45) CQ-MB (N = 181)
Gastrointestinal disorders 23 (38.3%) 93 (41.5%)
Respiratory disorders 16 (26.7%) 78 (34.8%)
Infections and infestations 10 (16.7%) 38 (17.0%)
Skin and subcutaneous tissue disorders 8 (13.3%) a 9 (4.0%) a
Others 3 (5.1%) 6 (2.7%)
a p = 0.017
There were no differences between CQ and CQ-MB study groups in haemoglobin and haematocrit values over time in both the G6PD deficient and sufficient subgroups (Table 3). The absolute median increase in haematocrit from baseline until day 14 was 2% with no differences between the overall study groups and between the G6PD deficient subgroups.
Table 3 Changes of laboratory values between baseline and study day 3 (haemoglobin, Hb), day 4 (haematocrit, HCT) and day 14 (HCT) by study group and G6PD status.
CQ CQ-MB Group Comparison
Baseline – day 3/4: G6PD deficient children
Changes in Hb [g/dl] (N = 26) 0.4 ± 1.2 -0.7 ± 1.5 [-0.4; 2.5], pWMW = 0.126
Changes in HCT [%] (N = 30) 1.3 ± 4.1 -1.4 ± 3.9 [-0.9; 6.4], pWMW = 0.259
G6PD sufficient children
Changes in Hb [g/dl] (N = 176) -0.5 ± 1.3 -0.4 ± 1.6 [-0.6; 0.6], pWMW = 0.906
Changes in HCT [%] (N = 193) -1.2 ± 2.8 -1.1 ± 4.1 [-1.5; 1.3], pWMW = 0.871
Baseline – day 14: G6PD deficient children
Changes in HCT [%] (N = 30) 1.3 ± 3.3 1.4 ± 3.4 [-3.2; 3.0], pWMW = 0.979
G6PD sufficient children
Changes in HCT [%] (N = 187) 1.9 ± 3.8 1.2 ± 5.1 [-1.1; 2.5], pWMW = 0.698
Mean ± SD and 95% CI for the mean difference; no difference in change of Hb, HCT between G6PD deficient and sufficient children in the CQ-MB group for Hb (pWMW = 0.538), HCT (day 4, pWMW = 0.897) and HCT (day 14, pWMW = 0.933)
Efficacy of study drugs
Among the fully compliant patients there were 22/41 (53.7%, 95% CI [37.4%; 69.3%]) treatment failures in group CQ (15 ETF, 7 LCF) and 73/166 (44.0%, 95% CI [36.3%; 51.9%]) in group CQ-MB (48 ETF, 25 LCF). Comparing CQ and CQ-MB no significant differences were found regarding CF (pChi = 0.348, OR = 0.68, 95% CI [0.34; 1.35]), ETF (pChi = 0.444, OR = 0.71, 95% CI [0.34; 1.45]), LCF (pChi = 0.938, OR = 0.86, 95% CI [0.34; 2.16]) and LPF (pChi= 1.0, OR = 1.04, 95% CI [0.37; 2.95]) (Table 4).
Table 4 Efficacy according to study group (per protocol analysis)
CQ (N = 41) CQ-MB (N = 163)
Clinical failure (CF) 22 (53.7%) 73 (44.0%)
Early treatment failure (ETF) 15 (36.6%) 48 (28.9%)
Late clinical failure (LCF) 7 (17.1%) 25 (15.1%)
Late parasitological failure (LPF) 5 (12.2%) 21 (12.9%)
The median fever clearance time in the fully compliant patients for group CQ compared with CQ-MB was 8.0 and 8.8 hours respectively (pWMW = 0.725) and the median parasite clearance time was 91.3 and 86.4 hours respectively (pWMW = 0.140). The group comparisons are similar using the intention-to-treat approach.
Acceptance of study drugs
The acceptance of the MB treatment was good, despite some blue staining of clothes by the children's urine. There was no response from 4/181 caretakers, 94/181 stated that they had no problem with the CQ-MB treatment, while 83/181 mentioned minor difficulties concerning the washing procedure of the stained clothes.
Discussion
MB has been used systematically against malaria in different patient populations during the late 19th and the early 20th century, but the effects of this drug were poorly documented in these old studies [14,11,32]. The present report provides the first data on the safety and efficacy of methylene blue in young children with falciparum malaria in SSA. MB was chosen to be applied in combination with CQ for reasons of expected synergy and potential reversal of CQ resistance and because combination therapy is the new paradigm in malaria therapy [32,37,28]. Moreover, CQ has been the first-line treatment for malaria in Burkina Faso at the time of the study.
G6PD deficiency was diagnosed through a phenotypic method and this method was shown to have a good concordance with results from genotyping [23]. During treatment of 181 children, including 24 G6PD-deficient children with uncomplicated falciparum malaria in Burkina Faso, no drug related SAEs and, particularly, no cases of severe haemolysis were observed. This does not mean that SAEs can be excluded totally, but the likelihood is certainly smaller than the risk of young SSA children dying from malaria [34]. Moreover, no other adverse events likely to be related to the study drugs were noted. Thus, the findings from this study for the first time demonstrate the safety of a methylene blue-based combination in the treatment of malaria in young children of SSA. These results support previous findings on the safety of CQ-MB in G6PD sufficient adults in Europe and G6PD deficient adults in Burkina Faso [30,20]. Although there were no safety problems with the oral combination of CQ-MB, the administration of the bitter-tasting MB solution was sometimes difficult, especially in younger children. As a consequence, a paediatric formulation for taste-masking of MB is currently under development.
The findings from this study show that MB appears to be safe at an oral dose of up to 4 mg/kg/day over three days in SSA populations with dominating class III G6PD deficiency, despite its being on the list of drugs reported to potentially cause severe haemolysis in G6PD-deficient populations [12]. However, this listing at least in part, may have its origin in falsely attributing haemolysis caused by the underlying infectious disease to the drug used for treatment [3]. Nevertheless, as MB is a redox-cycling oxidant and G6PD has an important role in the elimination of reactive oxygen species in the erythrocyte, the safety of MB may be influenced by the prevailing type of G6PD deficiency [15]. Further studies are needed in populations where G6PD deficiency class II occurs [12].
The combination of dapsone and chloroproguanil (Lapdap) has recently been registered for malaria therapy in SSA [19]. With regard to the potential of haemolysis development in G6PD-deficient populations, MB belongs to the same risk category as dapsone [12]. Thus, our findings may also be reassuring regarding the safety of dapsone. Lapdap is currently undergoing phase IV studies in different SSA populations.
Compared to CQ resistance data from the surrounding villages, the observed rate of clinical failures during CQ treatment was surprisingly high in the urban/semi-urban population of this study (details will be published separately) [27]. In most of Burkina Faso, the level of resistance to CQ has remained remarkably stable during recent years [29,35,24]. However, much higher CQ failure rates were already documented from the capital town Ouagadougou [33]. As a consequence, a policy change regarding first-line treatment of uncomplicated malaria in Burkina Faso was decided upon in early 2005 and it is planned to use artemisinin-based combination therapy (ACT) starting in 2006.
There were no differences in baseline characteristics between the two study groups, with the exception of a higher haematocrit in the group CQ-MB compared to the CQ group, a finding likely to be explained by chance. The clinical failure rate was higher in the CQ arm (54%) compared to the CQ-MB arm (44%), but this difference was not significant. This result could possibly be explained by too low a dose of MB chosen in this study. Such an assumption is supported by much higher MB doses reported for treatment of malaria patients some 100 years ago [12,14,21,16]. In one of these studies, a good safety and efficacy of MB at oral doses of 20–50 mg/kg/day MB over several days to weeks was demonstrated during treatment of 40 young children with malaria in Brasil [11]. However, in these times treatment was usually uncontrolled, and efficacy outcomes in these old publications often were poorly documented. As a consequence of the lack of efficacy of the CQ-MB combination in this study, a MB dose finding study has been conducted in 2004 in the same study area. Findings confirm the efficacy of MB at higher doses and will be published separately.
In the present investigation, methylene blue was again well accepted by the study population [20]. Staining of clothes was not considered as a major problem by mothers. Only in a few cases, two to three traditional washes were needed before stains were totally removed (Sanon, unpublished data). This confirms results from a former experimental study on the reversibility of MB stains in local clothes and supports findings from an anthropological study – on community perceptions of blue urine and blue clothes – conducted during the rainy season of the year 2003 in the rural Nouna study area (Sanon, unpublished data).
In conclusion, MB has been shown to be safe for the treatment of uncomplicated falciparum malaria in young West African children with a high prevalence of G6PD deficiency. However, the efficacy of the CQ-MB combination has not been sufficient at the MB dose used in this study. Future studies need to assess the efficacy of MB at higher doses and in combination with appropriate partner drugs.
Authors' contributions
P Meissner and G Mandi contributed equally to the study. P Meissner, G Mandi, S Witte, U Mansmann, A Jahn, I Walter-Sack, H Schirmer, B Kouyaté and O Müller designed the study. P Meissner, G Mandi, B Coulibaly, M Sanon, J Rengelshausen, W Schiek, G Mikus, J Burhenne and KD Riedel conducted the laboratory and clinical work. S Witte, U Mansmann and T Tapsoba did the statistical analysis. All authors contributed to the writing of the paper. O Müller was the principal investigator.
Acknowledgements
The study was funded by an award from DSM Fine Chemicals, Linz, Austria and by the Deutsche Forschungsgemeinschaft (SFB 544 "Control of Tropical Infectious Diseases") at Ruprecht-Karls-University Heidelberg.
==== Refs
Amaral L Viveiros M Kristiansen JE Phenothiazines: potential alternatives for the management of antibiotic resistant infections of tuberculosis and malaria in developing countries Trop Med Int Health 2001 6 1016 1022 11737839 10.1046/j.1365-3156.2001.00804.x
Beutler E Mitchell M Special modifications of the fluorescent screening method for glucose-6-phosphate dehydrogenase deficiency Blood 1968 32 816 818 4386875
Beutler E G6PD deficiency Blood 1994 84 3613 3636 7949118
Bloland PB Ettling M Meek S Combination therapy for malaria in Africa: hype or hope? Bull World Health Organ 2000 78 1378 1388 11196485
Breman JG Alilio MS Mills A Conquering the intolerable burden of malaria: what's new, what's needed: a summary Am J Trop Med Hyg 2004 71 1 15 15331814
Coulibaly B Malaria-related studies on enzymopathies, methaemoglobin, and methylene blue Dissertation 2004 Heidelberg University
Coulibaly B Eubel J Gromer S Schirmer RH Becher H, Kouyaté B Biochemistry-based health care research. Methylene blue, malaria, methaemoglobin, and malnutrition Health research in developing countries 2005 Springer Verlag Berlin Heidelberg 285 292
Djimdé A Doumbo OK Cortese JF Kayentao K Doumbo S Diourte Y Dicko A Su XZ Nomura T Fidock DA Wellems TE Plowe CV Coulibaly D A molecular marker for chloroquine-resistant falciparum malaria N Engl J Med 2001 344 257 263 11172152 10.1056/NEJM200101253440403
Ehrlich P Chemotherapeutics: scientific principles, methods, and results Lancet 1913 182 445 451 10.1016/S0140-6736(01)38705-6
Färber PM Arscott LD Williams CH Becker K Schirmer RH Recombinant Plasmodium falciparum glutathione reductase is inhibited by the antimalarial dye methylene blue FEBS Letters 1998 422 311 314 9498806 10.1016/S0014-5793(98)00031-3
Ferreira C Sur l'emploi du bleu de méthylène dans la malaria infantile Rev Ther Med Chir 1893 488 525
Fleming AF de Silva PS Cook GC, Zumla A G6PD deficiency Manson's Tropical Diseases 2003 Elsevier Science Ltd, Edinburgh, UK 208 212
Ginsburg Golenser J Glutathione is involved in the antimalarial action of chloroquine and its modulation affects drug sensitivity of human and murine species of Plasmodium Redox Rep 2003 8 276 279 14962364 10.1179/135100003225002907
Guttmann P Ehrlich P Über die Wirkung des Methylenblau bei Malaria Berliner Klinische Wochenschrift 1891 28 953 956
Janssen WJ Dhaliwal G Collard HR Saint S Clinical problem-solving. Why "why" matters N Engl J Med 2004 351 2429 2434 15575060 10.1056/NEJMcps040669
Klemperer G Grundriss der klinischen Therapie innerer Erkrankungen, 2 1924 Auflage, Urban und Schwarzenberg, Berlin und Wien 34 35
Kotea R Kaeda JS Yan SLK Sem Fa N Beesoon S Jankee S Ramasawmy R Vulliamy T Bradnock RW Bautista J Luzzatto L Krishnamoorthy R Mason PJ Three major G6PD-deficient polymorphic variants identified among the Mauritian population Br J Haematol 1999 104 849 854 10192449 10.1046/j.1365-2141.1999.01230.x
Küpfer A Aeschlimann C Wermuth B Cerny T Prophylaxis and reversal of ifosfamide encephalopathy with methylene blue Lancet 1994 343 763 764 7510815 10.1016/S0140-6736(94)91839-2
Lang T Greenwood B The development of Lapdap, an affordable new treatment for malaria Lancet Infect Dis 2003 3 162 168 12614733 10.1016/S1473-3099(03)00547-4
Mandi G Witte S Meissner P Coulibaly B Mansmann U Rengelshausen J Schiek W Jahn A Sanon M Wust K Walter-Sack I Mikus G Burhenne J Riedel KD Schirmer H Kouyate B Muller O Safety of the combination of chloroquine and methylene blue in healthy adult men with G6PD deficiency from rural Burkina Faso Trop Med Int Health 2005 10 32 38 15655011 10.1111/j.1365-3156.2004.01356.x
Mayer M Über die Wirkung von Methylenblau bei Malaria quartana Dtsch Med Wochenschrift 1919 45 1052 1053
Meierjohann S Walter RD Müller S Regulation of intracellular glutathione levels in erythrocytes infected with chloroquine-sensitive and chloroquine-resistant Plasmodium falciparum Biochem J 2002 368 761 768 12225291 10.1042/BJ20020962
Meissner P Coulibaly B Mandi G Mansmann U Witte S Schiek W Müller O Schirmer RH Mockenhaupt FP Bienzle U Diagnosis of red cell G6PD deficiency in rural Burkina Faso. Comparison of a rapid fluorescent enzyme test on filter paper with PCR based genotyping Br J Haematol 2005
Modiano D Luoni G Sirima BS Lanfrancotti A Petrarca V Cruciani F Simpore J Ciminelli BM Foglietta E Grisanti P Bianco I Modiano G Coluzzi M The lower susceptibility to Plasmodium falciparum malaria of Fulani of Burkina Faso (West Africa) is associated with low frequencies of classic malaria-resistance genes Trans R Soc Trop Med Hyg 2001 95 149 152 11355545 10.1016/S0035-9203(01)90141-5
Müller O Becher H Baltussen A Ye Y Diallo DA Konate AT Gbangou A Kouyate B Garenne M Effect of zinc supplementation on malaria morbidity among West African children: a randomized double-blind placebo-controlled trial BMJ 2001 322 1567 1572 11431296 10.1136/bmj.322.7302.1567
Müller O Traoré C Kouyaté B Becher H Malaria morbidity, treatment seeking behaviour, and mortality in a cohort of young children in rural Burkina Faso Trop Med Int Health 2003 8 290 296 12667146 10.1046/j.1365-3156.2003.01030.x
Müller O Traoré C Kouyaté B Clinical efficacy of chloroquine in young children with uncomplicated malaria – a community based study in rural Burkina Faso Trop Med Int Health 2003 8 202 203 12631308 10.1046/j.1365-3156.2003.01018.x
Nosten F Brasseur P Combination therapy for malaria Drugs 2002 62 1315 1329 12076181
Ouedraogo JB Dutheil Y Tinto H Traore B Zampa H Tall F Coulibaly SO Guiguemde TR In vitro sensitivity of Plasmodium falciparum to halofantrine compared with chloroquine, quinine and mefloquine in the region of Bobo-Dioulasso, Burkina Faso (West Africa) Trop Med Int Health 1998 3 381 384 9623943
Rengelshausen J Burhenne J Fröhlich M Tayrouz Y Singh SK Riedel KD Muller O Hoppe-Tichy T Haefeli WE Mikus G Walter-Sack I Pharmacokinetic interaction of chloroquine and methylene blue combination against malaria Eur J Clin Pharm 2004 60 709 715 10.1007/s00228-004-0818-0
Sarma GN Savvides SN Becker K Schirmer M Schirmer RH Karplus PA Glutathione reductase of the malarial parasite Plasmodium falciparum: Crystal structure and inhibitor development J Mol Biol 2003 328 893 907 12729762 10.1016/S0022-2836(03)00347-4
Schirmer RH Coulibaly B Stich A Scheiwein M Merkle H Eubel J Becker K Becher H Muller O Zich T Schiek W Kouyate B Methylene blue as an antimalarial agent Redox Rep 2003 8 272 276 14962363 10.1179/135100003225002899
Sirima SB Tiono AB Konate A Diarra A Castelli F Pinoges L Mugittu K Taylor WR Olliaros PL Efficacy of artesunate plus chloroquine for the treatment of uncomplicated malaria in children in Burkina Faso: a double-blind, randomized, controlled trial Trans R Soc Trop Med Hyg 2003 97 345 349 15228257 10.1016/S0035-9203(03)90166-0
Snow RW Craig M Deichmann U Marsh K Estimating mortality, morbidity and disability due to malaria among Africa's non-pregnant population Bull World Health Organ 1999 77 624 640 10516785
Tinto H Zoungrana EB Coulibaly SO Ouedraogo JB Traore M Guiguemde TR Van Marck E D'Alessandro U Chloroquine and sulphadoxine-pyrimethamine efficacy for uncomplicated malaria treatment and haematological recovery in children in Bobo-Dioulasso, Burkina Faso during a 3-year period 1998–2000 Trop Med Int Health 2002 7 925 930 12390597 10.1046/j.1365-3156.2002.00952.x
Trape J The public health impact of chloroquine resistance in Africa Am J Trop Med Hyg 2001 64 12 17 11425173
White NJ Nosten F Looareesuwan S Watkins WM Marsh K Snow RW Kokwaro G Ouma J Hien TT Molyneux ME Taylor TE Newbold CI Ruebush TK 2ndDanis M Greenwood BM Anderson RM Olliaro P Averting a malaria disaster Lancet 1999 353 1965 1967 10371589 10.1016/S0140-6736(98)07367-X
WHO World malaria situation 1994 Weekly Epid Rec 1997 72 269 292
WHO Assessment and monitoring of antimalarial drug efficacy for the treatment of uncomplicated falciparum malaria 2003 World Health Organization, Geneva WHO/HTM/RBM/2003.50
Winstanley P Modern chemotherapeutic options for malaria Lancet Infect Dis 2001 1 242 250 11871511 10.1016/S1473-3099(01)00119-0
|
16179085
|
PMC1262758
|
CC BY
|
2021-01-04 16:24:13
|
no
|
Malar J. 2005 Sep 22; 4:45
|
utf-8
|
Malar J
| 2,005 |
10.1186/1475-2875-4-45
|
oa_comm
|
==== Front
Malar JMalaria Journal1475-2875BioMed Central London 1475-2875-4-471618802310.1186/1475-2875-4-47ResearchCommunity concepts of malaria-related illness with and without convulsions in southern Ghana Ahorlu Collins K [email protected] Kwadwo A [email protected] Cynthia [email protected] Savigny Don [email protected] Mitchell G [email protected] Noguchi Memorial Institute for Medical Research, University of Ghana, Box LG581, Legon, Ghana2 Swiss Tropical Institute, Socinstrasse 57, CH-4002, Basel, Switzerland3 Department of Social Work, University of Ghana, Legon, Ghana2005 27 9 2005 4 47 47 8 6 2005 27 9 2005 Copyright © 2005 Ahorlu et al; licensee BioMed Central Ltd.2005Ahorlu et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Malaria, both with or without convulsions, is a serious hardship for people living in endemic areas, especially in sub-Saharan Africa. Community references to malaria, however, may encompass other conditions, which was collectively designated malaria-related illness (MRI). Inasmuch as the presence or absence of convulsions reportedly affects timely help-seeking for malaria, a local comparison of these conditions is needed to inform malaria control.
Methods
Vignette-based EMIC interviews (insider-perspective interviews) for MRI with convulsions (convulsion positive, MRI-CP) and without convulsions (convulsion negative, MRI-CN) were developed to study relevant features of MRI-related experience, meaning and behaviour in two rural communities in Ghana. These semi-structured interviews elicited both qualitative narrative and categorical codes for quantitative analysis. Interviews with 201 respondents were conducted.
Results
The conditions depicted in the vignettes were well recognized by respondents and named with various local terms. Both presentations were considered serious, but MRI-CP was more frequently regarded potentially fatal than MRI-CN. More than 90.0% of respondents in both groups acknowledged the need to seek outside help. However, significantly more respondents advised appropriate help-seeking within 24 (p = 0.01) and 48 (p = 0.01) hours for MRI-CP. Over 50.0% of respondents responding to questions about MRI-CP identified MRI-CN as a cause of convulsions.
Conclusion
Local comparison of MRI-CP and MRI-CN based on vignettes found a similar profile of reported categories of perceived causes, patterns of distress, help-seeking and preventive measures for both presentations. This differs from previous findings in sub-Saharan Africa, which assert communities regard the two conditions to be unrelated. The perceived relationships should be acknowledged in formulating strategies to control malaria through timely help-seeking and treatment to reduce childhood mortality.
==== Body
Background
According to the WHO about 1.3 million people died of malaria in 2003, and about 90% of these deaths occurred in sub-Saharan Africa [1-3]. Convulsions are features of many of these fatal illnesses, and caretakers easily recognise convulsions as a health problem in children. Though biomedical reasons for convulsions with malaria remain unclear, studies have shown important practical differences in local actions for managing childhood malaria-related illnesses (MRI) with convulsions (i.e. convulsion positive, MRI-CP) and without convulsions (i.e. convulsion negative, MRI-CN) [4,5]. Many studies argue that convulsions, often considered unrelated to MRI, lead to a significant alteration in the meanings of the illness experience and behaviour, indicating more frequent reliance on traditional healers as the primary source of treatment or in combination with biomedicine [4,6-12].
Practical questions remain about how this cultural epidemiology of MRI-CP relates to timely, appropriate health-seeking by caretakers. Many studies have reported that MRI-CN is generally treated first at home – either with modern pharmaceuticals (mainly analgesics and inadequate antimalarials), or herbal medications, or both – and caretakers seek outside help only when the illness persists or when they observe a high fever [8,10,13-16]. A recent report by de Savigny et al. [14], on the other hand, found that for about 80% of malaria-related deaths in Tanzania, modern help was used first. Most control programmes focus on various clinical features of malaria apart from convulsions, and few studies have compared community views of MRI-CP and MRI-CN. Strategies for malaria control assume that appropriate interventions at the febrile stage of the disease will prevent progression to more serious, life- threatening illness.
To explain locally perceived relationships between MRI-CP and MRI-CN, and to consider their practical implications for local management and timely, appropriate help-seeking, this study compares local illness experiences, meaning and behaviour associated with the two conditions. It aims to clarify their common and distinctive feature, and to consider implications for local management and timely, appropriate help-seeking. Based on the framework for local classification, the study examines the assertion that they are locally perceived to be two distinct conditions, as reported in the literature [6-9].
Study Area
This study was conducted from October, 2002 to April, 2004 in two malaria-endemic villages, Galo-Sota in the Keta District and Obosomase in the Akuapim North District of Ghana. Keta District is located in the coastal savannah vegetation zone of the Volta region, where about a third of the total surface area is covered with lakes and ponds. The district has a population of 137,751 (Government of Ghana national population census, 2000), consisting mainly of the Anlo patrilineal descent group (98.8%). The Anlo language is one of the closely related dialects spoken among the Ewe-speaking people of Ghana (Keta District Annual Report, 2001). The Anlo people are predominantly subsistence food-crop farmers, but they also cultivate shallot (a tropical spice grown in commercial quantities). Some are fishermen and petty traders. Galo-Sota is a rural village with a population of about 6,000 to 7,000. A health post is situated at the centre of the village, staffed by a nurse/midwife, two community nurses and two auxiliary workers. Malaria was the most common health problem treated at the community health post in 2002. A tributary of the Volta River passes through the village, demarcating a boundary between Galo and Sota, which together constitute Galo-Sota.
The Akuapim North District in the eastern region of Ghana is situated in the forest zone. The district population is 113,915, according to the last census (National population census, 2000). The Akuapim-Twi-speaking people, members of the Akan matrilineal descent group, predominate in the district. These Akuapim people are also mainly food-crop farmers and petty traders. Oil palm, a cash crop, is cultivated on a limited scale. The district is currently being prepared as a designated site for a malaria vaccine trial. Among health science activities in the area is a Centre for Scientific Research into Plant Medicine (Akuapim North District Annual Report, 2001). Obosomase (population 7,000 to 8,000) is the rural study village in the Akuapim North District. It has a community clinic staffed by a nurse/midwife, a community health nurse and one auxiliary staff. As in Galo-Sota, malaria was the most common health problem treated at the community clinic in 2002.
Study Methods
Two EMIC interview tools were developed locally to study the relationship between sociocultural factors and appropriate treatment-seeking for children up to five years of age. EMIC interviews are locally adapted instruments for assessing representations of illness or specified health problems from the perspective of affected persons, their family or community members. They elaborate the distribution of local insiders' illness-related experience, meaning and behaviour integrating qualitative and quantitative approaches [17,18]. The designs of these semi-structured vignette-based interviews were informed by baseline ethnographic data which generated illness narratives that specified locally relevant MRI-related categories of distress, perceived causes and help-seeking behaviours. The vignettes for MRI-CP and MRI-CN are included in an appendix "a" and "b". To account for gender differences, two vignettes were presented with men responding to male vignettes and women responding to female vignettes.
Respondents aged 20 years or more were selected randomly from community registers to participate in the interviews. However, younger respondents between the ages of 20 and 25 years of age were unwilling to participate in the interview using the vignette depicting MRI with convulsion, which introduced some age difference between the two groups of respondents. The reason for refusal was, they claimed, that they did not have any experience to talk about. Informal discussions, however, revealed that some people believed that if they talked about convulsions, this might bring that condition to their children. This attitude was observed in both communities. One hundred interviews were conducted using vignettes depicting MRI-CN and 101 using vignettes depicting MRI-CP. Respondents were evenly distributed by sex and community. Few of those originally selected eventually refused to participate in either of the interviews – two men and one woman at Obosomase and three men at Galo-Sota. The dissenting individuals were replaced from the registers.
Interviews were conducted by the first author in the local languages, and data were recorded by a research assistant with a degree in sociology, who was trained to record both qualitative and quantitative interview data. The EMIC instrument was pre-tested to refine it and gain experience. The pretesting showed it was unnecessary to tape record the interviews, inasmuch as the data were recorded very well by the research assistant.
Data from the two communities were pooled for this analysis. Item-specific qualitative narrative data were entered into a word processor (Microsoft Word) and imported in a structured format for automatic coding, referencing text segments of EMIC interview items in MAXqda, a programme for textual analysis [19]. These data were analysed to complement and elaborate the quantitative accounts and to clarify relevant aspects of illness-related experience, meaning and behaviour. Variables of interest in the quantitative data-base were imported into MAXqda as selection variables. This integrated approach to quantitative and qualitative analysis enabled us to perform a phenomenological analysis of relevant coded segments from selected respondents to categories that were analysed with quantitative methods.
Quantitative data were double entered in DOS EpiInfo version 6.04 and subsequently analysed with the EpiInfo Windows version 3.3 [20]. The frequencies of spontaneously reported and probed responses were examined to specified categories of cultural epidemiological variables for patterns of distress (PD), perceived causes (PC), self-help at home (SH), outside help-seeking (HS) and preventive measure (PC). The prominence of reported variables (2 = spontaneously reported, 1 = probed and 0 = not reported) was compared in the MRI-CP and MRI-CN groups using the Wilcoxon test. The percentage reporting each category and the fraction reporting the category spontaneously are presented in tables. Respondents were also asked to identify the most troubling symptom (PD) and the most important perceived cause (PC). These variables were compared in the two groups with Fisher's exact test. The variables, as initially coded, were also grouped and analysed under relevant subgroups, based on investigators' judgement of shared meanings under a broader heading.
Reported duration from illness onset to observe, wait or treat at home before seeking-help from outside the home for the children presented in the two vignettes was analysed to compute anticipated timely, appropriate help-seeking for MRI-CP and MRI-CN within 24 and 48 hours. The reported timeliness was then compared for the two vignette groups.
Results
The socio-demographic characteristics of respondents are presented in Table 1. Households were headed mainly by men, 75.2% and 67.0% in MRI-CP and MRI-CN groups respectively. Many respondents could not specify their household income, which was about the same in the two groups of respondents.
Table 1 Demographic characteristics of respondents
Demographic characteristics of respondents Vignettes presented
With convulsion (N = 101) Without convulsion (N = 100)
Age of respondents*
Mean age 49.0 38.0
Std. Dev. 15.8 12.5
Mode 35 25
Female: male ratio 1:1 1:1
Education
Mean years of education 6.1 6.6
No education (%) 22.9 19.0
Highest education (years) 20 14.0
Marital status (%)
Married 66.3 58.0
Never married 3.0 20.0
Divorced/separated 16.8 16.0
Widowed 13.9 6.0
Religion (%)
Christianity 55.4 64.0
Traditional religion 40.6 35.0
Islamic 3.0 1.0
Others 1.0 0.0
Household income (%)
Regular & dependable 23.8 23
Uncertain/not sure 52.5 48
Not regular & dependable 23.7 29
* P < .01 (T-test)
The children presented in the two vignettes were the focus of interview questions. Local names or terms used to describe their illnesses and their approximate English translations or meanings are presented in Table 2. The two conditions presented were identified by different names and terms. All the terms had been identified during ethnographic study as local terms for MRI-CP and MRI-CN.
Table 2 Local terms and their approximate English equivalents
Vignette with convulsions
Obosomase English translation Galo-Sota English translation
Sroakyereno Attacked by the sky Xeivitsoe Taken/attacked by a bird
Adiatorniso Obsessed or possessed Adukpodzidor dzedzi Attacked by garbage dump illness
Atridii barima akyereno Attacked by male malaria Hehedor dzedzi Attacked by stretching illness
Dordzagla/dorsese dzedzi Attacked by a strong illness
Dzifotorwotsoe Taken by the people of the sky
Miatorwotsoe Taken by our friends
Vignette without convulsions1
Atridii Hot body, yellowish urine, yellowish eyes, Vomiting, cold, and shivering, bodily pains, weakness, refusal of food, easily startled, paleness, weight loss, etc Asra Hot body, yellowish urine, yellowish eyes, Vomiting, cold and shivering, bodily pains, weakness, refusal of food, easily startled, paleness, weight loss, etc
Ebun Fever
Feve Nudza
Malaria Malaria
1Local terms and names for MRI without convulsions have no single equivalent in English, and were used interchangeably to represent similar conditions. "Malaria" and "fever" have also been incorporated in local usage as terms and names.
Respondents in the MRI-CP group appeared somewhat more likely than MRI-CN respondents to consider the condition "usually fatal" (39.6% against 27.0%). Narrative elaborated concerns about mortality: "This kind of illness kills children easily, especially if parents do not respond quickly to treating the child."
"As for this problem, it is a very serious one. If it does not kill the child then it could destroy the child's ability to reason properly."
For MRI-CN children, mortality was likely to be associated with some vulnerability or neglected treatment: "This is a very serious problem; so long as the child is still very small she may die if not treated properly and early."
"Fever can kill children because people do not consider the disease serious when it is starting, so before they realize it, it has already become worse for the child. What I mean is that it is not a condition that should usually kill, but if it is not seen and treated early, it does kill."
In both situations, a majority of respondents (98.0% for MRI-CP and 96.0% for MRI-CN), said it was necessary that someone stays at home to care for the sick child. The mother of the child was the obvious choice for 97.0% of MRI-CP and for 95.0% of the respondents for vignette depicting MRI-CN.
Patterns of distress for childhood MRI with and without convulsions
Reported symptoms of both conditions are presented in Table 3. Some of these were distinctive for each presentation, and others were reported for both presentations, but with some differences. Convulsion-related symptoms were more prominent in the MRI-CP group, and more frequently reported as most troubling. On the other hand, MRI-CN respondents reported significantly more fever-related symptom and identified them more frequently as most troubling. Non specific symptoms, except for breathlessness, were also most prominent in responses of the MRI-CN group.
Table 3 Reported symptoms and single most troubling symptom of MRI with and without convulsions
Categories of distress reported1 Reported spontaneously and probed2 Most troubling
With Convulsion (n = 101) Without convulsion (n = 100) With Convulsion (n = 101) Without convulsion (n = 100)
Total (%) Fraction Spont. Total (%) Fraction Spont.
Convulsions related symptoms 98.3 0.72 64.0 0.05** 67.3 6.0**
Unconscious 86.2 0.39 18.0 0.06** 26.7 2.0**
Stiffness x x 0.0 0.00 21.8 0.0**
Easily startled/frightened 76.3 0.32 59.0 0.03** 7.9 3.0
Rolling the eye balls x x 0.0 0.00 5.9 0.0**
Biting the lips 42.5 0.17 0.0 0.00** 2.0 0.0
Foaming mouth 55.4 0.31 0.0 0.00** 2.0 0.0
Folded arms 52.5 0.64 0.0 0.00** 1.0 0.0
Shaking x x 0.0 0.00 0.0 1.0
Fever related symptoms 37.6 0.55 98.0 0.79** 9.9 34.0**
Hot bodies x x x x 6.9 11.0
Sweating 14.0 0.14 55.0 0.06 1.0 0.0
Yellowish eyes 28.8 0.65 97.0 0.78** 1.0 18.0**
Yellowish urine x x x X 1.0 4.0
Chills and Rigors x x x X 0.0 1.0
Non specific symptoms 93.1 0.52 99.0 0.66** 20.8 56.0**
Weakness 72.3 0.52 86.0 0.51* 11.9 15.0
Breathlessness 51.4 0.14 29.0 0.07** 5.9 1.0
Refusal of food x x x X 1.0 16.0**
Weight loss 35.5 0.14 63.0 0.25** 1.0 12.0
Diarrhoea 0.0 0.00 33.0 0.06** 1.0 0.0
Vomiting x x x X 0.0 7.0**
Paleness 44.5 0.20 74.0 0.23** 0.0 4.0
Sleepiness 0.0 0.00 12.0 0.25** 0.0 1.0
Bitterness in the mouth 0.0 0.00 33.0 0.61** 0.0 0.0
Joint and bodily pains 16.0 0.25 36.0 0.22** 0.0 0.0
Crying x x x x 0.0 0.0
1 Symptoms analysed as groups (in bold) based on reported categories that follow.
*p ≤ 0.10, ** p ≤ 0.05. Wilcoxon test for comparison of prominence of reported categories (2 = spontaneous, 1 = probed response, 0 = not reported), and Fisher's exact test for most troubling symptoms.
2Column values indicate frequency of reported categories and the fraction of these reported spontaneously. Column values marked by "x" indicate categories specified in the vignettes.
The following narratives indicate how people explain the symptoms presented in the vignettes. A woman speaking about convulsive illness characterized typical features: "In most cases, the child's jaws are locked, and the child becomes very stiff. Also, foaming fluid comes from the mouth. In some cases the child becomes unconscious."
Another respondent explained, "Among most typical symptoms, the child begins to shiver and all of a sudden becomes very stiff and hot, and rolls the eye balls."
A man explained typical symptoms for MRI-CN: "The first sign is that the child feels very cold and then at certain times feels hot. It is also possible that the child loses weight and sweats so much. The child's eyes also turn yellow because of the fever. The colour of his urine will look yellowish, and he may become very weak."
A woman explained, "The child's body becomes hot and this makes her cry a lot. She may complain of headache, and feels cold and shivers. She may also find it difficult to eat; her eyes become yellow, and she looks pale and weak."
Most troubling symptoms reported more frequently by MRI-CP respondents included unconsciousness and stiffness. For MRI-CN respondents, most troubling symptoms more frequently reported were yellowish eyes, refusal of food, and weight loss. Weakness was identified by some respondents in both groups as most troubling.
Among features of distress, apart from somatic symptoms of malaria, a number of problems were reported to affect the families of the children in the vignettes with and without convulsions (as specified by percent reported/ fraction spontaneous). Nearly all respondents reported loss of income for families of the affected children (99.0% reporting/0. 42 fraction spontaneous for MRI-CP, and 100% reporting/0.46 fraction spontaneous for MRI-CN), concern about the course of illness (100.0%/0.90 MRI-CP, and 99.0%/0.90 MRI-CN) and sadness, anxiety or worry (99.0%/0.94 MRI-CP, and 100%/0.93 MRI-CN). However, financial concerns (unavailability of funds for treatment and inability to work for money), which leads to anxiety, were frequently reported as the most troubling categories of distress for the family, (71.3% for MRI-CP, and 76.0% for MRI-CN). Concern about the course of illness was also considered one of the most troubling categories of distress for the family (26.7% for MRI-CP, and 21.0%% for MRI-CN). These were common features of both groups without statistically significant difference between them.
The following representative narratives explain the importance of income loss to the family of a child with MRI. An MRI-CP respondent elaborated: "Generally, if a child becomes ill, the parents are worried because they do not know what could happen to their child, and in addition to that, it could lead to financial problems for the family." Another respondent lamented: "The child's condition can adversely affect her in the future, so the parents would become so bothered and worried about the child, especially if they do not have money. The child could become deaf and mute or have a mental problem. I have seen one like that before."
Referring to MRI-CN, a man said: "Fever kills so the family would be worried. Also money issues could be a problem for them."
Another respondent explained: "It is very normal that when your child is sick, you become worried, especially if you do not have money."
Perceived causes
The prominence of many perceived causes differed for the two groups (Table 4). The most frequently reported perceived causes of convulsions reported by MRI-CP respondents were spirits, phlegm, worm infections and atridii/asra/malaria; this last category refers to local terms for MRI as a cause of convulsions. Most frequently reported perceived causes of MRI by MRI-CN respondents were mosquito bites, eating too much fatty or oily food and heat from the sun.
Table 4 Reported perceived causes and the single most important cause for MRI with and without convulsions
Perceived causes reported1 How reported2 Most important
With convulsion (n = 101) Without convulsion (n = 100) With convulsion (n = 101) Without convulsion (n = 100)
Total (%) Fraction Spont. Total (%) Fraction Spont.
Insect bites
Mosquito bite 44.4 0.24 94.0 0.48** 6.9 35.0**
Infections & malaria-related 95.1 0.64 73.0 0.07** 43.6 1.0**
Phlegm 77.1 0.41 24.0 0.04** 15.8 0.0**
Atridii/asra/malaria 57.4 0.54 0.0 0.00** 13.9 0.0**
Worm infections 58.2 0.17 68.0 0.06 13.9 1.0**
Exposure 39.6 0.67 89.0 0.63** 9.9 33.0**
Heat from the sun or fire 27.8 0.75 81.0 0.68** 8.9 31.0**
Airborne/exposure 16.9 0.41 22.0 0.14 1.0 2.0
Sanitation & hygiene 58.4 0.63 90.0 0.51** 10.0 11.0
Personal hygiene 36.8 0.78 51.0 0.65* 5.0 4.0
Sanitation 25.9 0.73 30.0 0.73 4.0 4.0
Playing on the ground 21.9 0.46 43.0 0.37** 1.0 0.0
Houseflies 34.7 0.14 78.0 0.13** 0.0 3.0*
Supernatural 84.2 0.35 79.0 0.04** 19.8 1.0**
Spirits 78.3 0.38 63.0 0.02** 19.8 1.0**
Evil eyes or sorcery 50.4 0.24 41.0 0.00** 0.0 0.0
Food and drink 36.7 0.38 96.0 0.41** 2.0 18.0**
Eating unbalanced diet 0.0 0.00 17.0 0.12** 2.0 3.0
Fatty/oily food 26.8 0.15 84.0 0.12** 0.0 8.0**
Eating unripe fruits 12.9 0.00 72.0 0.08 0.0 6.0**
Impure water 11.8 0.83 31.0 0.39** 0.0 1.0
1 Perceived causes analysed as groups (in bold) based on reported categories that follow.
*p ≤ 0.10 and ** p ≤ 0.05. Wilcoxon test for comparison of prominence of reported categories (2 = spontaneous, 1 = probed response, 0 = not reported), and Fisher's exact test for most important perceived causes.
2Column values indicate frequency of reported categories and the fraction of these reported spontaneously.
Overlapping meanings were also reported in respondents' accounts of perceived causes. For example, a woman in the MRI-CP group explained: "Some people claim that if children play on refuse dumps they easily get convulsion, but I also think that a child can get it through mosquito bites, because these give fever, which can lead to this condition. Worms can also cause a child to get this condition, because worms release some substances into the child's stomach, which in turn gives the child phlegm. And this can cause a convulsion. Evil spirits can also cause children to have this illness."
A man said: "Malaria is the major cause, but in some cases, spiritual forces can also cause a convulsion. It also depends on the kinds of food that children take, like unbalanced diet."
Explaining MRI-CN, a woman commented: "This condition could have been caused by worm infestations or houseflies that perch on food and contaminate it before it is eaten. Sometimes too mosquito bites can cause it."
Another said: "Maybe she wasn't eating good food. Bad food like fatty/oily foods can lead to this problem. I know that mosquito bites or living in a dirty environment could also cause it. And exposure to the heat from the sun can also cause it."
Self-help at home
Respondents reported various self-help options for both conditions. The most prominent among reported sources of self-help for MRI of both types was herbal-based remedies for drinking (Table 5). Among pharmaceutical medicines, purchasing drugs from the drug or chemical shops was most prominent in responses of both groups, but significantly more so for the MRI-CN group. Other traditional remedies, such as scarification were reported more for the MRI-CP vignette.
Table 5 Self-help at home for MRI with and without convulsions
Self help at home reported1 How reported2
Vignette with convulsion (n = 101) Vignette without convulsion (n = 100)
Total Fraction Spont. Total Fraction Spont.
Herbal & other local actions 96.0 0.92 99.0 0.76**
Home-prepared herbal medications for drinking 72.3 0.75 83.0 0.60
Bathing with ordinary cold water (tepid sponging) 67.3 0.34 70.0 0.34
Home-prepared herbal medications for enema 52.4 0.62 44.0 0.46*
Home-prepared herbal medications for bathing 40.6 0.73 33.0 0.30*
Other actions (scarification etc) 22.9 0.96 5.0 1.00**
Home-prepared herbal medications for other uses 8.0 0.88 8.0 0.50
Modern pharmaceuticals 52.5 0.34 92.0 0.60**
Drug/chemical shops to purchase drugs 47.6 0.31 89.0 0.57**
Other leftover drugs at home 7.7 0.26 9.0 0.33
Leftover antimalarials at home 1.0 1.00 7.0 0.86**
1 Self-help at home analysed as groups (in bold) based on reported categories that follow.
*p ≤ 0.10 and ** p ≤ 0.05. Wilcoxon test for comparison of prominence of reported categories (2 = spontaneous, 1 = probed response, 0 = not reported).
2Column values indicate frequency of reported categories and the fraction of these reported spontaneously.
A woman emphasized the value of various traditional remedies for MRI-CP: "In some cases they say that water kept in a 'banku pot' (utensil for preparing a local maize meal) overnight, mixed with urine, can be used to bathe the child for relief. Herbal preparations can also be used to bathe the child. Some people prefer to prepare some herbs for the child to drink. Some also give honey, while some people may buy drugs from the chemical sellers."
Commenting on MRI-CP vignette, a man indicated how various interventions might all be appropriate: "Some of the possible actions may be to buy drugs from a chemical seller for the child to drink or prepare herbs for drinking or enema."
A respondent in the MRI-CN vignette group indicated the value of diverse treatment for that condition as well: "Some medicine could be given at home as a measure to reduce the severity of the illness. The medicine could be bought from the drug stores if there are no leftovers. Some herbal medications can also be given to the child."
Another said: "The family may decide to buy drugs for the child; they could also decide to prepare herbs for the child to drink."
Outside help-seeking
The need to seek help outside the home was reported by nearly everyone for both conditions (Table 6). More MRI-CP respondents, however, were concerned about getting treatment right away. In the MRI-CP group, 29.0% said outside help should be sought within 24 hours, compared with 9.0% for the MRI-CN group (p < 0.01). The same relative priority was indicated by responses advising treatment within 48 hours (53.0% MRI-CP and 34.0% MRI-CN, p = 0.01).
Table 6 Help-seeking from outside the home for MRI with and without convulsions
Outside help seeking reported1 How reported2
Vignette with convulsion (n = 101) Vignette without convulsion (n = 100)
Total Fraction Spont. Total Fraction Spont.
Clinic & hospital 99.0 0.96 100.0 1.00
Government/community clinic 95.1 0.96 99.0 0.97
Government hospital 95.0 0.96 98.0 1.00
Local healers & religious leaders 84.2 0.40 66.0 0.26**
Local healers 69.4 0.46 57.0 0.26**
Religious leaders (pastors/Imams) 44.3 0.11 35.0 0.06
Other source of modern pharmaceuticals 13.8 0.43 19.0 0.21
Health worker in the community for advice 12.8 0.38 17.0 0.12
Drug/chemical shops for advice 3.0 0.33 8.0 0.25
1 Outside help-seeking analysed as groups (in bold) based on reported categories that follow.
*p ≤ 0.10 and ** p ≤ 0.05. Wilcoxon test for comparison of prominence of reported categories (2 = spontaneous, 1 = probed response, 0 = not reported).
2Column values indicate frequency of reported categories and the fraction of these reported spontaneously.
Similar outside sources of help were identified by respondents in both groups (Table 6). Although more MRI-CP respondents recommended traditional healers, other health care providers were suggested by similar percentages from both groups.
Commenting on outside help-seeking for MRI-CP, a woman explained that home remedies at some point were not enough: "If the herbs used at home do not work other people who know more herbs, like traditional healers, could be consulted. The child could also be taken to the clinic or hospital, but this costs money."
Another observed: "As soon as it is clear that what is done at home does not work, the child would have to be sent to a clinic or hospital. The hospital or clinic is the best place, and the family must go, but they can also see a traditional healer for treatment."
Some MRI-CN respondents like this woman, compared modern and traditional health care providers favourably: "A clinic or a hospital should be the best places to go but some traditional healers also know about herbs that work."
A man emphasized the importance of not waiting too long before getting help from a doctor: "As soon as the child's condition does not get better after the home treatment, the family should consult a doctor."
Prevention
Most respondents in both groups (74.3% MRI-CP and 84.0% MRI-CN) said the conditions in the vignettes could have been prevented, there was, however, a borderline significant difference (p = 0.06). Categories of preventive measures suggested for both presentations were similar (Table 7). Preventive measures frequently reported included preventing mosquito bites, staying less in the sun, maintaining personal hygiene, environmental cleanliness, drinking clean water, avoiding fatty or oily foods, and reducing strenuous or hard work/play. These were reported more frequently by respondents in MRI-CN group. Preventive measures based on magico-religious ideas were mentioned more frequently in the MRI-CP group. Measures frequently reported by similar percentages in both groups included taking medications regularly (herbal or biomedicine) and deworming children regularly.
Table 7 Preventive measures for MRI with and without convulsions
Preventive measures reported1 How reported2
Vignette with convulsion (n = 101) Vignette without convulsion (n = 100)
Total Fraction Spont. Total Fraction Spont.
Insect bites
Prevent mosquitoes' bite3 58.4 0.27 94.0 0.31**
Regular medications 88.1 0.24 87.0 0.27
Taking herbal or biomedicine regularly 75.4 0.45 73.0 0.44
Deworming regularly 50.4 0.18 49.0 0.16
Sun & strenuous play 33.6 0.21 80.0 0.29**
Stay less in the sun or near fire 31.9 0.22 79.0 0.28**
Reduction in strenuous play 8.0 0.13 35.0 0.20**
Sanitation & hygiene 55.4 0.66 71.0 0.66**
Cleaning the environment 29.7 0.63 49.0 0.74**
Keeping personal hygiene 51.5 0.65 66.0 0.67**
Food and drink 59.4 0.47 86.0 0.38**
Eating balanced diet 33.6 0.59 40.0 0.63
Avoid fatty/oily foods 29.8 0.03 63.0 0.14**
Drinking clean water 18.8 0.37 32.0 0.41**
Eating less starchy food 17.8 0.11 28.0 0.11*
Eating on time (not going hungry for long) 10.9 0.00 16.0 0.13
Drink a lot of vegetable soup 10.9 0.28 18.0 0.28
Magico-religious 47.5 0.19 37.0 0.00**
Avoid offending evil spirits like the witches 23.8 0.25 15.0 0.00*
Attend to ancestral spirits and family gods 40.7 0.20 32.0 0.00
Don't know/cannot tell 22.8 1.00 14.0 1.00
1 Preventive measures analysed as groups (in bold) based on reported categories that follow.
*p ≤ 0.10 and ** p ≤ 0.05. Wilcoxon test for comparison of prominence of reported categories (2 = spontaneous, 1 = probed response, 0 = not reported).
2Column values indicate frequency of reported categories and the fraction of these reported spontaneously.
Discussion
This study identifies similarities and differences between local concepts, meanings, self help at home, help-seeking from outside the home and recommended preventive measures for MRI with and without convulsions. As reported in many studies across sub-Saharan Africa, spiritual forces dominate perceived causes for MRI-CP, compared to MRI-CN [4,6,8,9,14,22-24] However, Findings reported here show that priority of timely, appropriate care is not reduced by local traditional perceived causes of convulsive illness. Although more respondents in the convulsions group reported magico-religious causes than respondents in non-convulsions groups, they also more frequently recommended medical treatment within 24 and 48 hours of illness onset.
Although concerns about supernatural causes of convulsions were evident, even affecting our ability to recruit young adult respondents for the MRI-CP sample, most respondents for the two presentations reported that the children in the vignettes should be taken to the clinic or hospital for treatment. This differs from many reports across sub-Saharan Africa, where studies emphases use of traditional healers as the primary source of treatment for convulsions, rather than modern medical care [6-9,21]. Some studies however, have reported the use of both biomedicine and traditional healers [10,11,22]. Findings are consistent with those of de Savigny et al. [14] in Tanzania, reporting that 78.7% of fatal malaria cases received modern treatment as the first resort for their last illness episode.
The 28.0% of MRI-CP respondents recommending appropriate treatment within 24 hours was significantly more than the 9.0% of MRI-CN respondents who did so (p < 0.01), but the rates for both are relatively low, and far below the designated Abuja target. The percentages were higher for 48 hours (53.0% and 34.0%, p = 0.01), but still lower than the 60.0% specified in the Abuja target for 24 hours (15). This could mean that the message of the priority of timely treatment has either not yet reached these communities or it is not compelling enough to motivate action and hence needs reinforcement.
It is notable that despite the distinctiveness in many studies of MRI-CP and MRI-CN illnesses, 57.4% of respondents in the MRI-CP group have reported that malaria is a cause of convulsions, and 13.9% said malaria is the most important cause of convulsions. The emerging local awareness of the link between mosquitoes, malaria and convulsions should be strengthened to reinforce the priority of timely, appropriate treatment for febrile malaria without convulsions. Information, education and communication (IEC) have important roles to play in that regard. More people acknowledge the value of home-based treatment for MRI-CN, consistent with a policy to promote that option for uncomplicated malaria. Reliance on traditional healers, especially for MRI-CP, however, remains a problem, inasmuch as this results in many children not receiving timely antimalarial treatment. The use of a mix of traditional remedies (herbal and rituals) and biomedical treatments are consistent with the literature [5,10-12,22].
For both conditions, findings also suggest approaches to prevention that are related to local perceived causes, mostly involving avoidance of identified causes. Most people, similar in both groups, recommended regular medications as a preventive measure, indicating favourable prospects for implementing intermittent preventive treatment (IPT) for pregnant women. The finding suggests it may be feasible to introduce intermittent preventive treatment for children under five years of age to reduce morbidity and mortality in this vulnerable age group. When evaluating such policy options, the risk of drugs being used inappropriately should be weighed against prospects for reducing mortality.
Widespread recommendations to avoid mosquito bites to prevent both conditions indicate good prospects in these communities and others like them for acceptance of insecticide-treated bednets. The idea that uncomplicated malaria may progress to convulsions may further reinforce such an approach to prevention. These issues would need emphasis when developing IEC as an intervention to reduce MRI-related morbidity and mortality in the study communities.
The study shows that despite the complexity of local experience, meaning and behaviour with respect to malaria-related illnesses, it is possible to identify the distribution of categories and explain local illness behaviours, their sociocultural determinants, and practical implications in endemic local rural communities. Relating results to timely help-seeking and malaria prevention suggests ways of incorporating local relevance ideas into the design and implementation of local programme strategies, especially IEC, an indication of how cultural epidemiology may inform malaria control activities to make them sustainable to reduce MRI-related morbidity and mortality.
Though, this study was carried out in two rural communities in southern Ghana findings may be generally applicable in most part of Ghana, especially the southern half of the country. However, local variations must be considered when interpreting findings for areas outside the study localities.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
CK Ahorlu was involved in the conception and design of the study, fieldwork, data management, analysis, interpretation and writing of this paper.
KA Koram was involved in the conception and design of the study, and data management.
C Ahorlu was involved in the conception of the study, fieldwork, and data management.
D de Savigny was involved in analysis and interpretation of data, and writing of this paper.
MG Weiss was involved in the conception and design of the study, data analysis, interpretation and writing of this paper.
Ethical Review
The study was approved by the institutional review boards of Noguchi Memorial Institute for Medical Research and the Swiss Tropical Institute. It was also reviewed by the WHO/TDR Ethical Review Committee.
Table 8 - Appendix 1a. Appendix 1a. vignette depicting MRI with convulsion positive (MRI-CP)
Introduction to vignette 'I appreciate your agreeing to talk to me about a problem that affects many children in this district. I want to understand how you think about it. Keep in mind that it is your ideas that I am interested in, so please do not feel there is a right or wrong answer to the questions I will ask you. Do not be shy to tell me what you think. So then, let me tell you a story about a child called Kofi/Ama who has this problem.'
MALE 'Kofi is a 2/12 year-old boy who had been feeling fine and playing happily. One day last week, Kofi woke up crying, and his mother found that his body felt very hot. Kofi seemed to be feeling cold and he was shivering. This was on and off for some time and he refused to eat anything. His urine was yellow in colour. He has vomited too. A few hours later, he started rolling his eyes. With his eyes opened wide, he was shaking and became stiff' FEMALE 'Ama is a 2/12 year-old girl who had been feeling fine and playing happily. One day last week, Ama woke up crying, and her mother found that her body felt very hot. Ama seemed to be feeling cold and she was shivering. This was on and off for sometime and she refused to eat anything. Her urine was yellow in colour. She vomited too. A few hours later, she started rolling her eyes. With her eyes opened wide, she was shaking and became stiff'
Appendix 1b. Vignette depicting MRI with convulsion negative (MRI-CN)
Introduction to vignette 'I appreciate your agreeing to talk to me about a problem that affects many people in this district. I want to understand how you think about it. Keep in mind that it is your ideas that I am interested in, so please do not feel there is a right or wrong answer to the questions I will ask you. Do not be shy to tell me what you think. So then, let me tell you a story about a child called Kofi/Ama who has this problem.'
'Kofi is a 2/12 year-old boy who had been feeling fine and playing happily. One day last week, Kofi woke up crying, and his mother found that his body felt very hot. Kofi seemed to be feeling cold and he was shivering. This was on and off for sometime. He refused to eat anything. His urine was a yellow colour, and after a few hours, he vomited.' 'Ama is a 2/12 year-old girl who had been feeling fine and playing happily. One day last week, Ama woke up crying, and her mother found that her body felt very hot. Ama seemed to be feeling cold and she was shivering. This was on and off for sometime. She refused to eat anything. Her urine was a yellow colour, and after a few hours, she vomited.'
Acknowledgements
The authors wish to thank the chiefs, elders and residents of Obosomase and Galo-Sota, especially the respondents, for participating in the study. We also thank Fred Ayifli for his assistance during the field study. Thanks also go to the community assistants Saviour and Koanya at Galo-Sota and Amankwah and Oloso Ayeh at Obosomase for their role during the fieldwork. We sincerely thank Abdallah Abouihia for his statistical support. Final thanks go to the staff of the Epidemiology Department of Noguchi Memorial Institute for Medical Research, University of Ghana, Legon for their support. This investigation received financial support from a TDR Research Training Grant awarded to Collins K. Ahorlu.
==== Refs
WHO The World Health Report 2004 Geneva, Switzerland
Bryce J Boschi-Pinto C Shibuya K Black RE the WHO Child Health Epidemiology Reference Group WHO estimates of the causes of death in children Lancet 2005 365 1147 1152 15794969 10.1016/S0140-6736(05)71877-8
Korenromp EL Williams BG Gouws E Dye C Snow RW Measurement of trends in childhood malaria mortality in Africa: an assessment of progress toward targets based on verbal autopsy Lancet Infect Dis 2003 3 349 358 12781507 10.1016/S1473-3099(03)00657-1
Adongo BP Kirkwood B Kendal C How local community knowledge about malaria affects insecticide-treated net use in northern Ghana Trop Med Int Health 2005 10 366 378 15807801 10.1111/j.1365-3156.2005.01361.x
Williams HA Jones CO A critical review of behavioural issues related to malaria control in sub-Saharan Africa: what contributions have social scientists made? Soc Sci Med 2004 59 501 523 15144761 10.1016/j.socscimed.2003.11.010
Hill Z Kendal C Arthur P Kirkwood B Adjei E Recognizing childhood illnesses and their traditional explanations: exploring options for care-seeking interventions in the context of IMCI strategy in rural Ghana Trop Med Int Health 2003 8 668 676 12828551 10.1046/j.1365-3156.2003.01058.x
Nsungwa-Sabiiti J Kallander K Nsabagasani X Namusisi K Pariyo G Annika J Tomson G Peterson S Local fever illness classifications: implications for home management of malaria strategies Trop Med Int Health 2004 9 191 199 10.1111/j.1365-3156.2004.01319.x
Hausmann-Muela S Muela RJ Illness naming and home treatment practices for malaria An example from Tanzania 1998 Paper presented at the workshop on people and Medicines in East Africa, Mbale, Uganda
Snow RW Mwenesi H Rapuoda B Malaria situation analysis for Kenya 1998 Technical Report prepared for the Department for International Development (DFID)
Ahorlu CK Dunyo SK Afari EA Koram KA Nkrumah FK Malaria-related beliefs and behaviour in southern Ghana: implications for treatment, prevention and control Trop Med Int Health 1997 2 488 499 9217705 10.1046/j.1365-3156.1997.d01-298.x
Adongo P Hudelson P The management of malaria on young children in northern Ghana: A report of a rapid ethnographic study 1995 Unpublish manuscript, Navrongo, Ghana
Molyneux CS Murira G Masha J Snow RW Intra-household relations and treatment decision-making for childhood illness: A Kenya case study J Biosoc Sci 2002 34 109 131 11814209
Jenkins C Review of four ethnographic studies on the home management of malaria in Kenya, Ghana, Ethiopia and Malawi 1998 Technical Report prepared for the WHO/TDR
de Savigny D Mayombana C Mwageni E Masanja H Minhaj A Mkilindi Y Mbuya C Kasale H Reid G Care-seeking patterns for fatal malaria in Tanzania Malar J 2004 3 27 15282029 10.1186/1475-2875-3-27
WHO The Africa Malaria Report 2003 Geneva: WHO/CDS/MAL/2003
Nyamongo I Health care switching behaviour of malaria patients in a Kenyan rural community Soc Sci Med 2002 54 377 386 11824914 10.1016/S0277-9536(01)00036-3
Williams H Kachur P Nalwamba C Hightower A Simoonga C Mphande P A community perspective on the efficacy of malaria treatment options for children in Lundazi District, Zambia Trop Med Int Health 1999 4 641 652 10583897 10.1046/j.1365-3156.1999.00471.x
Agyepong IA Malaria ethnomedical perceptions and practice in an Adangbe farming community and implications for control Soc Sci Med 1992 35 131 137 1509302 10.1016/0277-9536(92)90160-R
Max Qualitative Data Analysis Introduction Verbi Software 2001 Consult Sozialforschung GnbH, Berlin
Weiss MG Cultural epidemiology: an introduction and overview Anthropology and Medicine 2001 8 5 29 10.1080/13648470120070980
Weiss MG The Explanatory Model Interview Catalogue (EMIC). Framework for comparative study of illness Transcult Psychiatry 1997 34 235 263
Ortega L Binka F Clinical parameters associated with the diagnosis of malaria in primary health care settings 1994 Technical Report prepared for the UNDP/world bank/WHO/TDR
Baume C Helitzer-Allen D Kachur P Patterns of care for childhood malaria in Zambia Soc Sci Med 2000 51 1491 1503 11077952 10.1016/S0277-9536(00)00049-6
Alilio M Eversole H Bammek J A KAP study on malaria in Zanziba: Implications for prevention and control. A study conducted for UNICEF sub-office Zanziba Eval Program Plann 1998 21 409 413 10.1016/S0149-7189(98)00030-5
Tarimo DS Urassa DP Msamanga GI Caretaker's perceptions of clinical manifestations of childhood malaria in holo-endemic rural communities in Tanzania East Afr Med J 1998 75 93 96 9640831
|
16188023
|
PMC1262759
|
CC BY
|
2021-01-04 16:24:13
|
no
|
Malar J. 2005 Sep 27; 4:47
|
utf-8
|
Malar J
| 2,005 |
10.1186/1475-2875-4-47
|
oa_comm
|
==== Front
Malar JMalaria Journal1475-2875BioMed Central London 1475-2875-4-481619427410.1186/1475-2875-4-48ResearchCytophilic antibodies to Plasmodium falciparum Glutamate Rich Protein are associated with malaria protection in an area of holoendemic transmission Lusingu John PA [email protected] Lasse S [email protected] Michael [email protected] Bruno P [email protected] Michael [email protected] Andrew Y [email protected] Martha M [email protected] Thor G [email protected] National Institute for Medical Research, Amani Medical Research Centre, Tanga & Headquarters, Dar es Salaam, Tanzania2 Centre for Medical Parasitology, Institute of Medical Microbiology and Immunology, University of Copenhagen and Department of Infectious Diseases, Copenhagen University Hospital (Rigshospitalet), Denmark3 Statens Serum Institut, Copenhagen, Denmark2005 29 9 2005 4 48 48 8 6 2005 29 9 2005 Copyright © 2005 Lusingu et al; licensee BioMed Central Ltd.2005Lusingu et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Several studies conducted in areas of medium or low malaria transmission intensity have found associations between malaria immunity and plasma antibody levels to glutamate rich protein (GLURP). This study was conducted to analyse if a similar relationship could be documented in an area of intense malaria transmission.
Methods
A six month longitudinal study was conducted in an area of holoendemic malaria transmission in north-eastern Tanzania, where the incidence of febrile malaria decreased sharply by the age of three years, and anaemia constituted a significant part of the malaria disease burden. Plasma antibodies to glutamate rich protein (GLURP) were analysed and related with protection against malaria morbidity in models correcting for the effect of age.
Results
The risk of febrile malaria episodes was reduced significantly in children with measurable anti-GLURP IgG1 antibodies at enrolment [adjusted odds ratio: 0.39 (95% CI: 0.15, 0.99); P = 0.047]. Interestingly, there was an inverse relationship between the plasma anti-GLURP IgG1 and IgG3 levels and the levels of parasitaemia at enrolment. However, anti-GLURP IgG2 and IgG4 levels were not associated with reduction in parasite density. Similarly, antibody levels were not associated with haemoglobin levels or anaemia risk.
Conclusion
Cytophilic IgG1 and IgG3 antibodies against R0-GLURP may contribute to the control of parasite multiplication and reduction in febrile malaria incidence in children living in an area of intense malaria transmission.
==== Body
Background
In areas of stable malaria transmission, immunity is acquired during childhood [1,2], and the protection is mainly mediated by antibodies directed against the blood stages of the parasite [3]. The relationship between malaria morbidity and antibody levels to malaria antigens has been analysed in several prospective longitudinal studies performed in different parts of Africa and Asia [4-9]. The Glutamate Rich Protein (GLURP) is a Plasmodium falciparum antigen, which has been studied extensively. It is a 220 kD protein expressed in the hepatic, asexual and sexual stages of the parasite life cycle [10]. The protein can be divided into an N-terminal non-repeat region (R25–500 or R0), a central repeat region (R1) and a C-terminal repeat region (R2) [11]. GLURP is a malaria vaccine candidate, which has undergone phase 1 trials in Europe and trials are planned to take place in Africa in the near future.
Several immuno-epidemiological studies using sera and clinical data from various sites have consistently identified high anti-R0-GLURP immunoglobulin G (IgG) levels as significant predictors of protection against high levels of parasitaemia, and febrile malaria episodes [6,12-16]. The protective antibodies are thought to elicit antibody dependent cytotoxic inhibition (ADCI) [17] through binding to the surfaces of merozoites [18]. Most of these studies have been performed in areas of moderate malaria transmission where protection against malaria fevers is achieved in those aged 5–15 years. In this report, plasma antibody levels to R0-GLURP was measured and related to malaria morbidity in a village subjected to holoendemic transmission and entomological inoculation rates exceeding one infectious bite per night [2]. In this community the incidence of febrile malaria decreases sharply by the age of three years and anaemia constitutes a significant part of the malaria disease burden [19]. Antibody levels to R0-GLURP in two other villages located in areas of moderate and low transmission were measured to compare the age related acquisition of antibodies in individuals living under different malaria transmission intensity.
Materials and methods
Study sites and population
A longitudinal malariometric study was carried out in three villages with different malaria transmission intensity in the Tanga region, Tanzania, as described in detail elsewhere [19]. The villages are situated at varying altitudes, which in north-eastern Tanzania is a proxy for malaria transmission intensity [20]. Malariometric surveys were conducted and blood samples were collected in April, July and September. Haemoglobin levels were measured using a HemoCue® photometer (Ångelholm, Sweden) and thick and thin blood smears for malarial microscopy were prepared. Thereafter, blood was centrifuged to obtain plasma, which was frozen at -20°C. Local village helpers and health workers at nearby health facilities performed passive case detection during the six month study period. The village helpers were provided with first-line antimalarial drug (sulphadoxine-pyrimethamine), paracetamol, microscope slides, blood lancets, treatment charts, febrile case detection forms and storage boxes. Villagers could seek treatment at any time from these helpers. Patients with symptoms of malaria were treated with the first-line antimalarial drug. If they had severe symptoms or did not respond adequately to the first-line treatment, they were referred to a health facility. Prior to treatment, the village helpers collected clinical information and a malaria blood smear. At each nearby health facility, two permanent staff members monitored study participants seeking medical treatment at the facility. If study participants presented at the facility with a history of fever, a form was completed and a blood smear collected. Active febrile case detection was undertaken once per month by the research team. During active case detection, study participants were seen by a trained physician and a blood smear was taken from all study participants who had reported a history of fever within two days and/or had axillary temperature ≥ 37.5°C.
Case definitions and selection of plasma samples for antibody assays
Febrile malaria episodes were defined as an axillary temperature ≥ 37.5°C and/or a history of fever within the previous 48 hours in the presence of asexual P. falciparum parasites ≥ 5000 parasites/μl [19]. Anaemia was defined as haemoglobin < 11.0 g/dl [21].
The incidence of febrile malaria episodes was low in Magamba and Ubiri [19]. In Mgome, 219 of the 254 individuals completed the longitudinal follow-up and from 171 of these individuals, sufficient plasma was available to measure R0-GLURP IgG class and subclass levels. Thus, antibody levels were measured in 9, 24, 33, 52, 35, and 18 individuals belonging to the 0–11 months, 1–2 years, 3–4 years, 5–9 years, 10–14 years, and 15–19 years age groups, respectively. Of the 171 individuals, 54 had febrile malaria episodes and 44 developed anaemia during the follow-up period.
To compare age-specific acquisition of R0-GLURP IgG class antibodies in areas of different endemicity, plasma samples of 40 individuals from Ubiri village (moderate transmission) and Magamba village (low transmission) were also tested. The samples were selected randomly from asymptomatic individuals to represent four age groups (0–4, 5–9, 10–14 and 15–19 years, N = 10 in each group). Since malaria morbidity was low in Ubiri and Magamba, no attempt was made to relate morbidity and anti-R0-GLURP IgG levels in these villages.
Antibody assays
Antibodies to R0-GLURP were measured by enzyme-linked immunosorbent assay (ELISA) based on a protocol developed by Afro Immunoassay. Briefly, microtitre plates (Maxisorp Nunc, Roskilde, Denmark) were coated overnight at 4°C with purified his-tag produced recombinant R0-GLURP (0.5 μg/ml) diluted in phosphate buffered saline (PBS). The plates were blocked with 3% powdered-milk-containing-phosphate buffer for one hour. Plasma from samples diluted 1:200 in dilution buffer (PBS with 1% powdered-milk and 0.1% Tween-20) was added in duplicate. The plates were then incubated at room temperature for one hour, where after peroxidase-conjugated rabbit anti-human IgG or IgM (Dako, Glostrup, Denmark) was added. Plates were washed four times with washing buffer (PBS with 0.1% Tween-20 and 0.5 M NaCl) between steps. Colour was developed using hydrogen peroxide with O-phenylenediamine (Dako, Glostrup, Denmark), and reading of antibody absorbance was done at 492 nm. Samples were retested if the measured differences in absorbance values between duplicate samples were higher than 15%. For determination of IgG subclasses (IgG1, IgG2, IgG3 and IgG4), plasma samples diluted 1:50 were added in duplicate and incubated for one hour at room temperature. The following monoclonal mouse anti-human subclasses were used: clone NL16 for IgG1 (Sky lab), clone NP6002 for IgG2 (Sigma), clone ZG4 for IgG3 (Sky lab) and clone RJ4 for IgG4 (Sky lab). The monoclonal IgG subclasses were diluted 1:2000 for IgG1, 1:3000 for IgG2, IgG3 and IgG4 in dilution buffer and incubated for one hour at room temperature. Goat anti-mouse IgG conjugated to peroxidase (Caltag) diluted 1:3000 in dilution buffer was then added and plates incubated for one hour. Colour development and reading of antibody absorbance was done as described above. Antibody levels were measured relative to the titration of IgG and IgG1-4 standard solutions. Plasma samples from 31 adult healthy Danes without any previous exposure to malaria were used as negative controls to generate cut-off values, which was defined as mean plus two standard deviations for the respective IgG class or subclasses. A positive control containing a pool of plasma from adult Liberian individuals (kindly provided by Dr. S. Jepsen, Statens Serum Institut, Copenhagen, Denmark through Afro Immuno Assay) were tested in parallel to the study samples. The positive plasma pool contained 27.7, 48.7, 35.7, 18.5, and 46.6 arbitrary units (AU)/ml anti-R0GLURP antibodies of IgG, IgG1, IgG2, IgG3, and IgG4, respectively. The cut-off values in arbitrary units were 3.5, 1.4, 1.4, 1.7, and 0 AU/ml for IgG, IgG1, IgG2, IgG3, and IgG4, respectively. The highest OD values in positive samples were 2.7 for IgG and 3.8 for IgG1-4.
Data analysis
Data were analysed using Stata/SE version 8.2 (Stata Corporation, Texas, USA; ). The age dependence of antibody responses was analyzed by use of Spearman's rank-order correlation and the mean matched-pairs test was used to compare R0-GLURP IgG levels between samples collected at the initiation and at the end of the study. The Mann-Whitney test was used to test R0-GLURP IgG class and subclasses levels between females and males. Univariate and multivariate models were fitted to estimate to what extent R0-GLURP IgG class and subclass levels could be attributed to protection against high parasite density, febrile malaria episodes or anaemia in individuals from the high transmission village. Differences were considered statistically significant if the 95% confidence interval was not overlapping or P < 0.05.
Results
Plasma antibody levels to R0-GLURP in areas of different malaria endemicity
The prevalence and mean levels of anti-R0-GLURP IgG were highest in residents living under intense and holoendemic malaria transmission in Mgome, lower in individuals living under moderate and seasonal transmission in Ubiri and lowest in residents living under low and unstable transmission in Magamba (Figure 1). The differences in IgG levels between these villages were highly significant (P(z) < 0.001, Trend test). The prevalence and levels of antibodies increased with donor age in Mgome and Ubiri villages (Figure 1), although there was a tendency of slight decrease in antibody levels after the age of 14 years in Mgome. Antibody levels did not differ significantly between males and females (Mann-Whitney test Mgome P(z) = 0.40; Ubiri P(z) = 0.12.; Magamba P(z) = 0.94). In Mgome, there was no statistical difference in antibody levels between samples collected at the beginning and at the end of the study (mean difference [95% confidence interval (CI)]: 4.0 arbitrary units (AU) [-15.8, 23.9], P(t) = 0.68, paired T-test). The mean antibody levels were slightly higher in samples collected at the end of the study in Ubiri (mean difference [95%CI]: 6.4 AU [-2.1, 15.0], P(t) = 0.13, paired T-test) and Magamba (mean difference [95%CI]: 7.6 AU [6.2, 9.0], P(t) < 0.0001, paired T-test).
Figure 1 Age-specific IgG responses to recombinant R0-GLURP in three villages: Mgome (high malaria transmission), Ubiri (moderate malaria transmission) and Magamba (low malaria transmission). The levels are presented as arbitrary units on a logarithmic scale. Box plots illustrate medians with 25th, 75th and whiskers for 10th and 90th percentiles including outliers (5th/95th percentiles). Lines with filled triangles represent proportion of responders in percentages.
In all three villages, R0-GLURP IgM was detected at very low levels with no distinct patterns observed with respect to age, sex or season (data not shown).
IgG subclass responses to R0-GLURP in Mgome village
The prevalence and mean levels of R0-GLURP IgG antibodies increased with age for all subclasses. For the cytophilic IgG1 and IgG3, there was a marked increase in three to four year old children (Figure 2). For the non-cytophilic IgG2 and IgG4, there was a steady increase in the levels and proportion of responders throughout the age groups (Figure 2). The levels of IgG4 antibodies were generally low.
Figure 2 Age-specific IgG subclass responses to recombinant R0-GLURP in Mgome (high transmission village). Panels A, B, C and D represent IgG1, IgG2, IgG3, and IgG4 subclasses, respectively. The levels are presented as arbitrary units on a logarithmic scale. Box plots illustrate medians with 25th, 75th and whiskers for 10th and 90th percentiles including outliers for each IgG subclass. Lines with filled triangles represent prevalence of responders in percentages.
Relationship between R0-GLURP IgG levels and Plasmodium falciparum density in Mgome (high transmission village)
Multiple linear regression models adjusting for age were generated to determine whether the anti-R0-GLURP IgG levels were associated with P. falciparum density at the initiation of the study. Interestingly, there was a significant association between cytophilic IgG subclass antibodies and reduced parasite density. For IgG1, one log unit increase in antibodies was associated with a reduction in P. falciparum density by 1,401 parasites/μl [95% CI: 231, 2571; P = 0.019]; and for IgG3 level, one log unit increase was associated with a reduction in P. falciparum density at enrolment by 983 parasites/μl [95% CI: 126, 1841; P = 0.025]. Although similar associations were observed in univariate analysis for non-cytophilic IgG subclasses such associations were not significant after adjusting for age (Table 1).
Table 1 Relationship between P. falciparum density (parasites/μl) and anti R0-GLURP IgG levels (log arbitrary unit/ml) at enrolment.
R0-GLURP antibodies Unadjusted coefficients (95% CI)1 P-value Adjusted coefficients (95% CI)2 P-value
IgG -773 (-805, 689) 0.378 -146 (-533, 3457) 0.15
IgG1 -2136 (-3477, -795) 0.002 -1401 (-2571, -231) 0.019
IgG2 -1512 (-2894, -128) 0.032 -719 (-2135, 697) 0.318
IgG3 -1482(-2542, -423) 0.006 -983 (-1841, -126) 0.025
IgG4 58 (-590, 706) 0.86 195 (-436, 826) 0.542
1 Univariate linear regression. 2 Multiple linear regressions adjusting for the effect of age including age as square root of age. Models using other forms of age correction (age, age × age, log age) gave similar results.
Relationship between R0-GLURP IgG and risk of developing febrile malaria in Mgome
The risk of developing a febrile malaria episode during the study decreased with age and was 65.2% (43/66), 11.5% (6/52), 11.4% (4/37) and 5.6% (1/18) for the age groups 0–4, 5–9, 10–14 and 15–19, respectively. Logistic regression models correcting for age indicated that the presence of a measurable R0-GLURP IgG1 was associated with a reduced risk of febrile malaria episodes [adjusted odds ratio (AOR) 0.39 (95% CI: 0.15, 0.99), P = 0.047]. The age adjusted odds ratio for individuals who had a measurable IgG3 response was 0.52, but the 95% confidence interval for this estimate was wide and not significantly different from one (Table 2). The presence of antibodies to R0-GLURP of the IgG2 or IgG4 subclasses was not associated with significant reduction in the risk of febrile malaria episodes.
Table 2 Odds ratio for the risk of febrile malaria episodes
R0-GLURP antibodies Unadjusted odds ratio (95% CI)1 P-value Adjusted odds ratio (95% CI)2 P-value
IgG 0.49 (0.25 – 0.94) 0.031 1.13 (0.50 – 2.53) 0.77
IgG1 0.17 (0.07 – 0.42) <0.001 0.39 (0.15 – 0.99) 0.047
IgG2 0.47 (0.20 – 1.08) 0.075 1.42 (0.58 – 3.46) 0.445
IgG3 0.25 (0.11 – 0.58) 0.001 0.52 (0.20 – 1.40) 0.197
IgG4 0.42 (0.18 – 0.99) 0.046 0.77 (0.30 – 2.03) 0.602
1 Logistic regression indicating risk of febrile malaria episode in individuals with detectable plasma levels of anti R0-GLURP antibodies (for IgG, IgG1, IgG2, IgG3 and IgG4, respectively) relative to those without a measurable antibody response. 2 Adjusted for the effect of age including age as square root of age.
Relationship between R0-GLURP IgG and risk of anaemia in Mgome
In logistic regression models including age and sex, there were no statistically significant associations between the risk of anaemia and having a measurable IgG [AOR: 1.03 (0.45, 2.35); P = 0.95] or IgG subclass (data not shown) R0-GLURP antibody response. Similarly, in linear regression models correcting for age and sex, there were no associations between plasma R0-GLURP IgG or IgG subclass levels and the haemoglobin level at enrolment (data not shown).
Discussion
Studies in which observed clinical protection is linked to the level of malaria antibodies on an individual level have been used to identify vaccine targets [5,6,15,22]. In these types of studies, participants are often divided into susceptible and protected individuals and this obviously requires that a reasonable number of the participants develop clinical symptoms during follow up. To allow comparisons between immunological parameters, it is also preferable that the protected individuals are as closely age matched to the susceptible individuals as possible. Due to problems of obtaining sufficient amounts of plasma from infants and young children, most studies have been conducted in areas of moderate transmission targeting children between 5–15 years or in areas of high transmission targeting children over 5 years and adults. These studies have documented that the presence of R0-GLURP IgG is associated with protection against febrile malaria attacks in areas of moderate and seasonal malaria transmission [6,13-15]. The acquisition of malaria immunity is governed by the transmission intensity [23], and in areas of holoendemic transmission children of three to four years of age have already developed considerable protection against febrile episodes [14,15]. In the current study, individuals with a measurable anti R0-GLURP IgG1 response had a statistically significant reduction in the risk of getting a febrile malaria attack compared with those without such antibodies. The results cannot unravel whether GLURP antibodies directly were responsible for the effect or whether they constitute a marker for other immunological activities. The protection provided by GLURP antibodies is mainly thought to be mediated through IgG1 and IgG3 antibodies which dampen the growth of blood stage parasites by antibody dependent cellular inhibition (ADCI) [13,14,17,24]; although levels of IgG2 GLURP antibodies have also been implicated in protection [13,14]. It is, therefore, of interest that a significant association between the level of anti R0-GLURP IgG1 and IgG3 antibodies and the parasite density at enrolment has been found in this study.
The acquisition of IgG against GLURP was highly dependant on malaria transmission intensity. In the high transmission village antibody levels increased markedly after the age of two years, and in the three to four years old children a very high percentage had a detectable antibodies response. In the village with moderate transmission antibody acquisition was much slower and GLURP antibody response rates over 50% were only seen in the age groups older than 10 years.
In areas of high malaria transmission one of the major disease burdens attributable to P. falciparum infection is anaemia occurring in infants and young children [2,19,21]. In Mgome, all children under two years had haemoglobin levels under 11 g/dl. Thus, it was disappointing that the levels of anti R0-GLURP antibodies were not associated with haemoglobin levels in multiple linear regression models or with anaemia in multiple logistic regression models.
Conclusion
In conclusion, this study conducted in an area of intense malaria transmission detected an association between protection against febrile malaria disease and presence of anti R0-GLURP antibodies and indicated that increasing levels of antibodies of the IgG1 and the IgG3 subclasses are associated with a reduction in P. falciparum parasite densities.
Authors' contributions
JPAL carried out field surveys, performed ELISA, analysed data and drafted the manuscript. LSV carried out field surveys and in collaboration with JPAL, MA and MT contributed to the set up of the ELISA. BPM carried out field surveys and participated in data analysis. MT, AYK, MML and TGT conceived the design of the study. All authors amended and approved the final manuscript.
Acknowledgements
We are grateful to all study participants including their parents/guardians as well as village helpers and health management teams in Tanga region. We acknowledge with thanks Anne Corfitz, Juma Akida, Zacharia Savael, Susanne Pedersen, Jimmy Weng, Fabio-Avit Massawe, John Hiza, William Chambo, Donald Mwanjeluka and Seth Nguhu for excellent technical assistance throughout the study. We are grateful to Drs Lars Hviid, Thomas Scheike and Daniel Dodoo for several suggestions and advice during the analysis. The study was conducted under the auspices of the Joint Malaria Programme, a collaborative research initiative between the Centre for Medical Parasitology at the University of Copenhagen and Copenhagen University Hospital, the Kilimanjaro Christian Medical College, the London School of Hygiene and Tropical Medicine and the Tanzania National Institute for Medical Research. John Lusingu is supported by a PhD scholarship from the Gates Malaria Partnership. The study was also supported by the ENRECA programme of the Danish International Development Agency (DANIDA).
==== Refs
Marsh K Snow RW Host-parasite interaction and morbidity in malaria endemic areas Philos Trans R Soc Lond B Biol Sci 1997 352 1385 1394 9355131 10.1098/rstb.1997.0124
Ellman R Maxwell C Finch R Shayo D Malaria and anaemia at different altitudes in the Muheza district of Tanzania: childhood morbidity in relation to level of exposure to infection Ann Trop Med Parasitol 1998 92 741 753 9924532 10.1080/00034989858989
Sabchareon A Burnouf T Ouattara D Attanath P Bouharoun-Tayoun H Chantavanich P Foucault C Chongsuphajaisiddhi T Druilhe P Parasitologic and clinical human response to immunoglobulin administration in falciparum malaria Am J Trop Med Hyg 1991 45 297 308 1928564
Riley EM Allen SJ Wheeler JG Blackman MJ Bennett S Takacs B Schonfeld HJ Holder AA Greenwood BM Naturally acquired cellular and humoral immune responses to the major merozoite surface antigen (PfMSP1) of Plasmodium falciparum are associated with reduced malaria morbidity Parasite Immunol 1992 14 321 337 1625908
Conway DJ Cavanagh DR Tanabe K Roper C Mikes ZS Sakihama N Bojang KA Oduola AM Kremsner PG Arnot DE Greenwood BM McBride JS A principal target of human immunity to malaria identified by molecular population genetic and immunological analyses Nat Med 2000 6 689 692 10835687 10.1038/76272
Soe S Theisen M Roussilhon C Aye KS Druilhe P Association between protection against clinical malaria and antibodies to merozoite surface antigens in an area of hyperendemicity in Myanmar: complementarity between responses to merozoite surface protein 3 and the 220-kilodalton glutamate-rich protein Infect Immun 2004 72 247 252 14688102 10.1128/IAI.72.1.247-252.2004
Dodoo D Theander TG Kurtzhals JA Koram K Riley E Akanmori BD Nkrumah FK Hviid L Levels of antibody to conserved parts of Plasmodium falciparum merozoite surface protein 1 in Ghanaian children are not associated with protection from clinical malaria Infect Immun 1999 67 2131 2137 10225865
Bull PC Lowe BS Kortok M Molyneux CS Newbold CI Marsh K Parasite antigens on the infected red cell surface are targets for naturally acquired immunity to malaria Nat Med 1998 4 358 360 9500614 10.1038/nm0398-358
Giha HA Staalsoe T Dodoo D Roper C Satti GM Arnot DE Hviid L Theander TG Antibodies to variable Plasmodium falciparum-infected erythrocyte surface antigens are associated with protection from novel malaria infections Immunol Lett 2000 71 117 126 10714439 10.1016/S0165-2478(99)00173-X
Dziegiel M Borre MB Jepsen S Hogh B Petersen E Vuust J Recombinant Plasmodium falciparum glutamate rich protein; purification and use in enzyme-linked immunosorbent assay Am J Trop Med Hyg 1991 44 306 313 2035752
Theisen M Vuust J Gottschau A Jepsen S Hogh B Antigenicity and immunogenicity of recombinant glutamate-rich protein of Plasmodium falciparum expressed in Escherichia coli Clin Diagn Lab Immunol 1995 2 30 34 7719909
Hogh B Petersen E Dziegiel M David K Hanson A Borre M Holm A Vuust J Jepsen S Antibodies to a recombinant glutamate-rich Plasmodium falciparum protein: evidence for protection of individuals living in a holoendemic area of Liberia Am J Trop Med Hyg 1992 46 307 313 1558270
Dodoo D Theisen M Kurtzhals JA Akanmori BD Koram KA Jepsen S Nkrumah FK Theander TG Hviid L Naturally acquired antibodies to the glutamate-rich protein are associated with protection against Plasmodium falciparum malaria J Infect Dis 2000 181 1202 1205 10720556 10.1086/315341
Oeuvray C Theisen M Rogier C Trape JF Jepsen S Druilhe P Cytophilic immunoglobulin responses to Plasmodium falciparum glutamate-rich protein are correlated with protection against clinical malaria in Dielmo, Senegal Infect Immun 2000 68 2617 2620 10768952 10.1128/IAI.68.5.2617-2620.2000
Meraldi V Nebie I Tiono AB Diallo D Sanogo E Theisen M Druilhe P Corradin G Moret R Sirima BS Natural antibody response to Plasmodium falciparum Exp-1, MSP-3 and GLURP long synthetic peptides and association with protection Parasite Immunol 2004 26 265 272 15541030 10.1111/j.0141-9838.2004.00705.x
Dziegiel M Rowe P Bennett S Allen SJ Olerup O Gottschau A Borre M Riley EM Immunoglobulin M and G antibody responses to Plasmodium falciparum glutamate-rich protein: correlation with clinical immunity in Gambian children Infect Immun 1993 61 103 108 8418032
Bouharoun-Tayoun H Oeuvray C Lunel F Druilhe P Mechanisms underlying the monocyte-mediated antibody-dependent killing of Plasmodium falciparum asexual blood stages J Exp Med 1995 182 409 418 7629503 10.1084/jem.182.2.409
Theisen M Soe S Oeuvray C Thomas AW Vuust J Danielsen S Jepsen S Druilhe P The glutamate-rich protein (GLURP) of Plasmodium falciparum is a target for antibody-dependent monocyte-mediated inhibition of parasite growth in vitro Infect Immun 1998 66 11 17 9423833
Lusingu JP Vestergaard LS Mmbando BP Drakeley CJ Jones C Akida J Savaeli ZX Kitua AY Lemnge MM Theander TG Malaria morbidity and immunity among residents of villages with different Plasmodium falciparum transmission intensity in North-Eastern Tanzania Malar J 2004 3 26 15282030 10.1186/1475-2875-3-26
Bodker R Akida J Shayo D Kisinza W Msangeni HA Pedersen EM Lindsay SW Relationship between altitude and intensity of malaria transmission in the Usambara Mountains, Tanzania J Med Entomol 2003 40 706 717 14596287
Schellenberg D Schellenberg JR Mushi A Savigny D Mgalula L Mbuya C Victora CG The silent burden of anaemia in Tanzanian children: a community-based study Bull World Health Organ 2003 81 581 590 14576890
Polley SD Mwangi T Kocken CH Thomas AW Dutta S Lanar DE E R Ross A G W Mwanakasale V G M B. L J.D. C Marsh K Human antibodies to recombinant protein constructs of Plasmodium falciparum Apical Membrane Antigen 1 (AMA 1) and their associations with protection from malaria. Vaccine 2004 23 718 728 15542195 10.1016/j.vaccine.2004.05.031
MacDonald The epidemiology and control of malaria 1957 London, Oxford University Press
Theisen M Soe S Jessing SG Okkels LM Danielsen S Oeuvray C Druilhe P Jepsen S Identification of a major B-cell epitope of the Plasmodium falciparum glutamate-rich protein (GLURP), targeted by human antibodies mediating parasite killing Vaccine 2000 19 204 212 10930674 10.1016/S0264-410X(00)00181-X
|
16194274
|
PMC1262760
|
CC BY
|
2021-01-04 16:24:13
|
no
|
Malar J. 2005 Sep 29; 4:48
|
utf-8
|
Malar J
| 2,005 |
10.1186/1475-2875-4-48
|
oa_comm
|
==== Front
Microb Cell FactMicrobial Cell Factories1475-2859BioMed Central London 1475-2859-4-281620216710.1186/1475-2859-4-28ResearchExpression and cytosolic assembly of the S-layer fusion protein mSbsC-EGFP in eukaryotic cells Blecha Andreas [email protected] Kristof [email protected] Klaas A [email protected] Marten [email protected]ödel Gerhard [email protected] Institut für Genetik, Technische Universität Dresden, D-01062 Dresden, Germany.2 Eukaryotic Microbiology, Groningen Biomolecular Sciences and Biotechnology Institute (GBB), University of Groningen, PO Box 14, NL-9750 AA Haren, The Netherlands.2005 4 10 2005 4 28 28 16 8 2005 4 10 2005 Copyright © 2005 Blecha et al; licensee BioMed Central Ltd.2005Blecha et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Native as well as recombinant bacterial cell surface layer (S-layer) protein of Geobacillus (G.) stearothermophilus ATCC 12980 assembles to supramolecular structures with an oblique symmetry. Upon expression in E. coli, S-layer self assembly products are formed in the cytosol. We tested the expression and assembly of a fusion protein, consisting of the mature part (aa 31–1099) of the S-layer protein and EGFP (enhanced green fluorescent protein), in eukaryotic host cells, the yeast Saccharomyces cerevisiae and human HeLa cells.
Results
Upon expression in E. coli the recombinant mSbsC-EGFP fusion protein was recovered from the insoluble fraction. After denaturation by Guanidine (Gua)-HCl treatment and subsequent dialysis the fusion protein assembled in solution and yielded green fluorescent cylindric structures with regular symmetry comparable to that of the authentic SbsC. For expression in the eukaryotic host Saccharomyces (S.) cerevisiae mSbsC-EGFP was cloned in a multi-copy expression vector bearing the strong constitutive GPD1 (glyceraldehyde-3-phosophate-dehydrogenase) promoter. The respective yeast transfomants were only slightly impaired in growth and exhibited a needle-like green fluorescent pattern. Transmission electron microscopy (TEM) studies revealed the presence of closely packed cylindrical structures in the cytosol with regular symmetry comparable to those obtained after in vitro recrystallization. Similar structures are observed in HeLa cells expressing mSbsC-EGFP from the Cytomegalovirus (CMV IE) promoter.
Conclusion
The mSbsC-EGFP fusion protein is stably expressed both in the yeast, Saccharomyces cerevisiae, and in HeLa cells. Recombinant mSbsC-EGFP combines properties of both fusion partners: it assembles both in vitro and in vivo to cylindrical structures that show an intensive green fluorescence. Fusion of proteins to S-layer proteins may be a useful tool for high level expression in yeast and HeLa cells of otherwise instable proteins in their native conformation. In addition the self assembly properties of the fusion proteins allow their simple purification. Moreover the binding properties of the S-layer part can be used to immobilize the fusion proteins to various surfaces. Arrays of highly ordered and densely structured proteins either immobilized on surfaces or within living cells may be advantageous over the respective soluble variants with respect to stability and their potential interference with cellular metabolism.
==== Body
Background
Bacterial cell surface layers (S-layer) as the outermost cell envelope components are a common feature of many bacteria and archaea species (for review see [1,2]). With few exceptions S-layers consist of a single species of subunits that occasionally is posttranslationally modified by phosphorylation [3], or glycosylation [4,5]. S-layer monomers assemble to two-dimensional highly porous arrays with either oblique, square or hexagonal symmetry. The interactions between the S-layer subunits as well as between the S-layer and the supporting envelope can be disrupted in a reversible manner by cation substitution or high concentrations of chaotropic agents [6]. Upon removal of the denaturing agent the isolated S-layer subunits assemble in vitro into regular arrays exhibiting structural features of the authentic cell surface layer. Contrary to the situation in vivo, the in vitro self-assembly process of S-layer proteins in solution can also result in double layer sheets or in tube-like structures [7].
S-layers have been recognized as important structures for biotechnological applications [8]. However, large-scale preparation of S-layers from the authentic organisms is sometimes limited. For example, some bacterial strains have been reported to loose their ability to produce S-layers under laboratory conditions [9]. Due to alterations in the cultivation conditions expression of truncated forms of the S-layer protein may result in the loss of S-layer sheets [10]. To circumvent such difficulties recombinant S-layer proteins have been heterologously produced in prokaryotic systems, like E. coli, Bacillus subtilis, Lactobacillus casei [9], or Lactococcus lactis [11]. For example, high level expression of the S-layer protein SbsC from Geobacillus stearothermophilus ATCC 12980 has been reported in E. coli [12]. SbsC possesses an N-terminal secretion signal of 30 amino acids (aa) that is cleaved off during secretion, a secondary cell wall polymer (SCWP)-binding domain (aa 31–258), and a central portion responsible for formation of lattice symmetry and self-assembly [12]. Upon expression in E. coli, the mature S-layer protein mSbsC(31–1099) with an apparent molecular mass of 112 kDa forms monolayer cylinders and spirally wound sheets-like structures in the cytosol that can be recovered from the insoluble fraction upon cell lysis. Neither deletion of the C-terminal 179 aa (rSbsC(31–920)) [7] nor fusion of this truncated form with major birch pollen antigen Bet v1 (rSbsC(31–920)-Betv1) interferes with the assembly process and the oblique lattice symmetry. Interestingly, the Bet v1 epitope was accessible to antibodies demonstrating that the fused portion is protruding from the respective assembly structure [13]. Contrary to the situation of SbsC, recombinant S-layer protein SbpA from B. sphaericus CCM 2177 and its derivatives fail to assemble in the E. coli cytosol, but form insoluble inclusion bodies [14].
In the present study we addressed the question whether the cytosol of eukaryotic host cells can provide a suitable environment for the formation of S-layer self assembly products, despite the presence of numerous chaperons and proteases. Assembly of S-layer fusion proteins could provide a simple purification scheme for soluble proteins, especially if the formation of their native structures depends on the conditions of the eukaryotic cytosol. In addition, such fusion proteins could stabilize proteolytically sensitive proteins and thus increase their yield.
We used the yeast Saccharomyces cerevisiae and the HeLa cells as eukaryotic model organisms for expression of a bifunctional S-layer fusion protein, consisting of mature SbsC (aa 31–1099) and the enhanced green fluorescent protein (EGFP).
Results
Expression of mSbsC-EGFP in E. coli
The 3207 bp open reading frame (ORF) encoding mSbsC(31–1099) was fused with the 720 bp ORF of EGFP and cloned into pET17b as described in Methods. Both ORFs are separated by a two-aa-linker (Leu-Glu) that was introduced by the non-template-encoded XhoI site (CTCGAG). E. coli BL21(DE3), transformed with pET17b-mSbsC-EGFP, were cultivated at 30°C and 37°C and analysed by fluorescence microscopy. Expression at 30°C resulted in an enhanced fluorescence emission, indicating a higher yield of correctly folded EGFP [14]. 0, 1, 2, and 4 hours after induction by IPTG cell lysates were tested for the presence of the fusion protein by Western blot analysis with anti-GFP antibody. Already after 1 h the mSbsC-EGFP-fusion protein, with an expected molecular weight of 139 kDa, could be detected (Fig. 1). The protein was recovered from the pellet fraction, indicating the formation of inclusion bodies or of assembly products as recently reported in the case of mature SbsC [7]. In addition to the predominant 139 kDa band a number of low molecular weight protein bands are detected by the anti-GFP antibody, that likely reflect degradation products, although internal start of translation cannot be excluded.
Figure 1 Expression of mSbsC-EGFP in E. coli. Expression of mSbsC-EGFP in E. coli transformants was induced by addition of IPTG as described in Material and methods. Immediately before induction (0 h), 1 h, 2 h and 4 h after induction samples of the cells were lysed and the soluble (S) and insoluble (I) fractions were prepared. 10 μg of each fraction were subjected to Western blot analysis with anti-GFP antibody. Sizes of marker proteins (Roti®-Mark prestained, Carl Roth GmbH; M) are indicated on the left hand side. Arrow indicates the full-size protein of 140 kDa.
Pure preparations of mSbsC-EGFP were obtained by FPLC as described in Methods. REM analysis of dialysed FPLC fractions of the purified full-size mSbsC-EGFP revealed predominantly tube-like structures. According to TEM analysis these tubes exhibit a regular symmetry (Fig. 2). The cylindrical structures had an average length of 10–20 μm and a diameter of approximately 60 nm. This partly resembles the situation of mSbsC which has been described to assemble into sheets and cylindric structures (albeit with a diameter of 70–110 nm) [7].
Figure 2 TEM analysis of recrystallized recombinant mSbsC-EGFP. Recombinant mSbsC-EGFP was isolated from E. coli transformants, denatured with 5 M Gua-HCl, dialysed against dH2O, and the assembled structures were subjected to TEM-analysis. The cylindrical structures with a diameter of approximately 60 nm exhibit regular symmetry. bar = 1μm
When recrystallization was performed in the presence of G. stearothermophilus wild type strain ATCC 12980 devoid of its native S-layer due to Gua-HCl treatment, cells were covered with a green fluorescent layer indicating binding to the SCWP (data not shown). The same result was obtained with other G. stearothermophilus strains (DSM 297, DSM 13240 and DSM 1550), but not with the Gua-HCl treated yeast strain BY4741.
Expression of mSbsC-EGFP in S. cerevisiae
For expression in a eukaryotic system p426-GPD-mSbsC-EGFP was transformed into the S. cerevisiae strain BY4741. As a control BY4741 was transformed with p426-GPD-EGFP. Expression of the S-layer fusion protein has only a very moderate effect on growth of the transformants (Fig. 3). Despite of a slightly prolonged lag phase and a marginally elongated doubling time, cell densities in the stationary phase are identical. Western blot analysis of the 20,000 × g pellet from whole cell lysates with anti-GFP antibody showed a strong signal by a distinct protein band whose size is identical with that of mSbsC-EGFP expressed in E. coli (Fig. 4C). Contrary to the situation in E. coli no further signals were detected, indicating that expression mSbsC-EGFP in yeast is not accompanied by any obvious degradation. Fluorescence microscopy revealed that EGFP-expressing cells show a homogeneous distribution of green fluorescence (Fig. 4A), while mSbsC-EGFP expression results in the formation of green fluorescent needle-like structures in the cytosol (Fig. 4C). Contrary to the situation of EGFP-expressing cells which all exhibit a similar fluorescence intensity, yeast cells expressing the fusion protein vary widely in their green fluorescence from low to extremely strong signals.
Figure 3 Growth curves of EGFP and mSbsC-EGFP expressing yeast transformants. Batch cultures of S. cerevisiae strain BY4741 expressing EGFP (open circles) or mSbsC-EGFP (filled circles) were inoculated in selective minimal medium with an OD600 of 0,02 and grown for 30 h at 30°C. Growth was followed by the increase of OD600.
Figure 4 Fluorescence microscopy of EGFP and mSbsC-EGFP expressing yeast transformants. Yeast cells expressing EGFP (A) or mSbsC-EGFP (C) were cultivated in selective minimal medium to exponential phase and analysed by fluorescence microscopy. Westernblot analysis of whole cell extract using anti-GFP antibodies reveal the presence of either EGFP (B) or mSbsC-EGFP (D). bar = 2 μm
TEM analysis of recombinant mSbsC-EGFP structures in yeast
TEM analysis of ultrathin sections of whole cells reveals tube-like self assembly products, that are exclusively located in the cytosol and neither associated with organelles, the cytoskeleton, or the plasma membrane (Fig. 5A). To verify the identity of these structures as the product of mSbsC-EGFP self assembly, anti-GFP antibody in combination with a secondary colloidal gold-conjugated antibody were used for detection in thin-sectioned protoplasts. Figure 5B shows that the densely packed protein crystals are selectively immunogold-labeled with primary anti-GFP antibody, thus confirming their identity as the recombinant S-layer fusion protein. TEM analysis of gently disrupted protoplasts revealed cylindrical, closely packed structures (Fig. 5C) exhibiting a regular symmetry (insert, for a more detailed image see 5D). These results show that mSbsC-EGFP monomers possess the intrinsic ability to form highly ordered assembly structures comparable to those observed after in vitro crystallization of mSbsC-EGFP (see Fig. 2).
Figure 5 TEM-analysis of ultra-thin section of a yeast transformant expressing mSbsC-EGFP. (A) mSbsC-EGFP expressing yeast spheroplast showing rod like structures (*). (B) Immunocytochemistry of mSbsC-EGFP expressing yeast cell using antibodies against GFP. (C) Negative staining of osmotically shocked yeast protoplasts expressing mSbsC-EGFP showing in vivo formed assembly products with similar structures (*) as in Fig 2, the inset and image (D) show a higher magnification of these structures, a lattice is discernible. Key: M-mitochondrion, N-nucleus, V-vacuole. The bar represents 1 μm, for the inset 100 nm.
Expression of mSbsC-EGFP in HeLa-cells
The ORFs of mSbsC and EGFP were fused in-frame within the vector pEGFP-N1. HeLa cells were liposome-transfected with the resulting plasmid pmSbsC-EGFP-N1. 16 h after transfection adherent HeLa cells exhibit green fluorescent filamentous structures. The fluorescence was exclusively restricted to these structures, other intracellular areas were competely devoid of green fluorescence. Obviously the fluorescent structures result from self assembly of mSbsC-EGFP within the cytosol. The morphology of these structures is very similar to that of the assembly products formed in mSbsC-EGFP producing yeast cells.
Transfected HeLa-cells were mechanically lysed with a dounce homogenizer, separated in a soluble and a pellet fraction by high speed centrifugation (20,000 × g), and the fractions analysed in a Western blot using anti-GFP antibodies. As shown in Figure 6, mSbsC-EGFP can be detected as an abundant protein band in the range of 139 kDa in the pellet fraction. Similar to the situation in yeast transformants we did not observe degradation products, indicating low or absent proteolytic activities.
Figure 6 Fluoresence microscopy and Western blot analysis of HeLa cells expressing mSbsC-EGFP. (A) mSbsC-EGFP expressing HeLa cells 16 h after liposome transfection. Adherent cells show a filament-like green fluorescent pattern that resembles the structures obtained in yeast cells expressing mSbsC-EGFP. (B) Western blot analysis of lysate from mSbsC-EGFP expressing HeLa cells. Lysed cells were centrifuged at 20,000 × g and the resulting supernatant (S) and pellet (P) fractions were analysed by SDS-PAGE and Western blot with anti-GFP antibody. The mSbsC-EGFP fusion protein is mainly present in the insoluble fraction.
Discussion
Despite of their potential for biotechnological application only few S-layers genes have been identified and were used for genetic modification and recombinant production. Usually prokaryotic systems like Lactobacillus casei, B. subtilis, E. coli (for review see: [9]) and Lactococcus lactis [11] were used for heterologous expression.
Our report describes for the first time the expression of a S-layer fusion protein in eukaryotic host cells. Because of the ability of mSbsC to form sheet-like and cylindrical structures in the cytosol of E. coli [7], we decided to use this S-layer protein for fusion with EGFP. Expression was first investigated in the yeast S. cerevisiae, one of the best-studied eukaryotic organism. We observed no alterations of the cell shape and only a slightly reduced vegetative growth, possibly reflecting the high energy demand due to the constitutive high level synthesis of 139 kDa mSbsC-EGFP. Western blot analysis of insoluble material with anti-GFP antibody gave no indication for N-terminal proteolytic degradation of the recombinant protein as observed in E. coli. Possibly the organization of mSbsC-EGFP in assembled structures confers protection from ubiquitin-mediated degradation in yeast. In respect to protein stability S. cerevisiae seems to be a superior expression system. Fluorescence microscopy revealed that yeast transformants exhibited an intense green fluorescence indicating proper folding of the EGFP part within mSbsC-EGFP. In line with the formation of self-assembly structures in the cytosol, the fluorescence pattern differs significantly from that of EGFP-expressing yeast cells. TEM studies revealed the presence of closely packed cylindrical structures with regular symmetry comparable with those obtained after in vitro recrystallization. Our study demonstrates that the assembly process of SbsC-EGFP is not affected by the presence of eukaryote-specific cytosolic chaperons, like members of the CCT family [15].
Recently it was reported that some S-layer monomers assemble on hydrophobic surfaces. S-layer proteins of the strains G. stearothermophilus PV72/p2 and NRS 2004/3a form protein monolayers on substrates like silicon wafers with a native oxide layer [16]. The self-assembly structures of SbsC-fusion proteins in yeast are not associated with the inner surface of the plasma membrane or cellular organelles despite their hydrophobic membranes. Obviously the hydrophobic membrane patches do not act in anchoring assembled S-layer structures.
As in yeast cells, mSbsC-EGFP monomers self assemble into filamentous structures within the cytosol of HeLa cells that are neither associated with subcellular structures nor affect the cell shape. We did not observe necrotic or apoptotic events in the transfected cells that expressed mSbsC-EGFP (data not shown). The detection of a single band of the expected molecular weight in Western blot analysis with anti-GFP antibodies indicates that the recombinant protein is stable and not subjected to N-terminal degradation.
A positively charged N-terminal domain of S-layer proteins directly interacts via electrostatic forces with the negatively charged SCWP whose chemical composition is identical in G. stearothermophilus wild-type strains [17]. In our study we show that fusion of EGFP to the mSbsC, resulting in an extension of the S-layer protein by ~240 aa, does not interfere with the binding.
Conclusion
Our results show that the mSbsC-EGFP fusion protein is efficiently synthesized, and self-assembles under the physiological conditions of the cytosol of eukaryotic cells. Both in yeast and human cells cytosolic accumulation of S-layer self assembly products is not accompanied by proteolytic degradation. The green fluorescence of mSbsC-EGFP suggests that both in vitro and in vivo folding of both protein portions within the fusion protein is independent, and not affected by each other. This observation opens the possibility to fuse other proteins, e.g. enzymes, to the S-layer part in order to obtain functional fusion constructs in a densely packed and stable structure.
Fusion of otherwise instable proteins to S-layer proteins may be a useful tool for their high level expression in eukaryotic cells. The self assembly properties of the fusion proteins can be exploited to purify them by simple centrifugation steps from cell lysate. By engineering a suitable protease cleavage site between the S-layer and the fused protein part, incubation of the assembly products with the respective proteases could allow separation and subsequent purification of the protein of interest. Arrays of highly ordered and densely structured S-layer fusion proteins can efficiently be immobilized on a variety of surfaces, either directly or upon treating with SCWP.
Besides technical applications, an ordered aggregation of S-layer fusion proteins within living cells may offer novel strategies for cell manipulation. For example, if expression of proteins would interfere with cellular functions, e.g. as a result of intracellular transport, fusion to an S-layer protein permits their retention in the cytosol. "In vivo affinity chromatography" may be another potential field for applying S-layer fusion proteins. Enrichment of soluble biomolecules within living cells may be achieved by assembly structures consisting of fusion proteins between the respective receptor and a suitable S-layer protein.
Methods
Strains
Geobacillus (G.) stearothermophilus ATCC 12980
G. stearothermophilus DSM 1550 (German Collection of Microorganisms and Cell Cultures (DSMZ) GmbH, Braunschweig, Germany)
G. stearothermophilus DSM 13240 (DSMZ)
G. stearothermophilus DSM 297 (DSMZ)
Escherichia (E.) coli DH5á (BRL)
E. coli BL21(DE3) (Invitrogen)
Saccharomyces (S.) cerevisiae BY4741 (EUROSCARF)
HeLa cells (kind gift of Frank Pfennig, Institut für Zoologie, Technische Universität Dresden, Germany)
Primers
#1: 5' TATATATACATATGGCAACGGACGTGGCGACGGTC 3'
#2: 5' TATATATACTCGAGCGATGCTGATTTTGTACCAATTTG 3'
#3: 5' TATATATACTCGAGATGGTGAGCAAGGGCGAGGAG 3'
#4: 5' TATATATAGCTCAGCTTACTTGTACAGCTCGTCCATGC 3'
#5: 5' TATATATAGGATCCATGGCAACGGACGTGGCGACGGTC 3'
#6: 5' TATATATAGGTACCTCACTATTACTTGTACAGCTCGTCCATGC 3'
#7: 5' TATATATACTCGAGATGGCAACGGACGTGGCG 3'#8: 5' TATATATACCGCGGCGATGCTGATTTTGTACCAATTTG 3'
Cultivation of G. stearothermophilus strains, isolation and recrystallization of S-layer proteins
Batch cultures of the G. stearothermophilus strains were grown in tryptone-enriched LB-medium at 55°C (strain DSM 13240 at 65°C) to an OD600 of 0,6. Surface proteins attached to the cell wall layer via noncovalent forces were removed by Guanidine-HCl (Gua-HCl) treatment of cells from a 500 ml culture. Cells were washed 3 times in distilled water, suspended in 10 ml of 5 M Gua-HCl and incubated for 30 min on ice with occasional mixing. Cells devoid of their S-layers were recovered by centrifugation at 8,000 × g and washed twice with 5 M Gua-HCl. For preparation of S-layer proteins the supernatant obtained after centrifugation at 40,000 × g was dialyzed against distilled water for 18 h.
Preparation of genomic DNA and DNA-cloning
Genomic DNA was isolated with the QIAGEN Genomic-tip 100 kit according to the manufacturer's instruction. Plasmid DNA was prepared with the "Wizard® SV Gel and PCR Claen-up system kit" (Promega).
Cloning in pET17b
DNA encoding mSbsC(31–1099) was PCR-amplified with primers #1 and #2, using total DNA of G. stearothermophilus ATCC 12980 as a template. An ATG initiation codon was introduced 5'to the open reading frame by primer #1. The PCR-product was cut with NdeI and XhoI and ligated with plasmid pET17b (Novagen) to yield plasmid pET17b-mSbsC_oT.
The XhoI and Bpu1102I sites of pET17b-mSbsC_oT were used to insert the EGFP reading frame, which was PCR-amplified from vector EGFP-N1 (Clontech) with primers #3 and #4. Primer #3 introduced a 5'-flanking XhoI-site in the PCR product, while primer #4 generated a 3'-flanking Bpu1102I-site and introduced a TAA termination codon. The PCR-product was cut with XhoI and Bpu1102I and ligated with pET17b-mSbsC_oT to yield plasmid pET17b-mSbsC_oT-EGFP.
Cloning in p426-GPD
A DNA fragment encoding mSbsC was PCR-amplified as described above with primers #5 and #2, cut with BamHI and XhoI, and ligated with the S. cerevisiae expression vector p426-GPD, which bears the strong GPD promoter [18]. The resulting plasmid p426-GPD-mSbsC_oT was used for in frame fusion the EGFP orf, which was amplified with primers #3 and #6. The respective PCR product harbours a 5' XhoI site and at the 3'-end three consecutive termination codons followed by a KpnI restriction site. Thereby plasmid p426-GPD-mSbsC_oT-EGFP was created.
Plasmid p426-GPD-EGFP for expression of EGFP in yeast was created by ligation of the PCR-fragment encoding the EGFP orf into p426-GPD via the flanking XhoI and KpnI sites. Plasmids p426-GPD-mSbsC_oT-EGFP and p426-GPD-EGFP were propagated in E. coli DH5α and transformed into the S. cerevisiae wild type strain BY4741.
Cloning in pEGFP-N1
The reading frame coding for mSbsC was amplified with primers #7 and #8 and subsequently integrated within the XhoI and SacII restriction sites in the vector pEGFP-N1 (Clontech) as an in frame fusion with the EGFP-ORF.
The resulting plasmid pmSbsC-EGFP-N1 was propagated in E. coli DH5α and transfected into HeLa cells.
Transfection of HeLa cells
For transient transfection with plasmid pmSbsC-EGFP-N1, 2,5 × 105 cells were plated on chamber slides (Nunc) and incubated after one day with TFX™-20 liposome reagent (Promega) and 3 μg DNA for 1 h. Subsequently DMEM with high glucose and 10 % (v/v) fetal calf serum (PAA laboratories) was supplemented. After over night cultivation transfected cells were analysed by Western blot and fluorescence microscopy.
Isolation and purification of recombinant S-layer proteins from E. coli
E. coli BL21(DE3) was used as a prokaryotic host for the expression of mSbsC-EGFP. E. coli transformants bearing pET17b-mSbsC-EGFP were cultivated in LB medium containing 100 mg/l ampicillin at 30°C to an OD600 of 0,4. S-layer synthesis was initiated by adding IPTG to a final concentration of 0,5 mM.
8–16 h after induction cells obtained from a 500 ml culture were harvested by centrifugation at 10,000 × g for 10 min; 4°C. The cell pellet was washed 3 times with ice-cold distilled water and resuspended in 10 ml Tris-buffer (10 mM Tris, pH 7,5; 1 mM AEBSF hydrochloride (AppliChem)) containing 1% (v/v) Triton X-100. After disruption of cells by French press (25,000 PSI pressure, at 4°C), soluble and insoluble fractions were separated by centrifugation at 20,000 × g for 30 min at 4°C. Nucleic acids were removed from the pellet fraction by treatment with DNase I (1 mg/ml) and RNase A (1 mg/ml) in Tris-buffer. After three washing steps with Tris-buffer the insoluble material was denatured in 5 ml 5 M Gua-HCl (in 10 mM Tris pH 7,5). The suspension was stirred at room temperature (RT) for 30 min and centrifuged for 30 min at 40,000 × g (10°C). Renatured S-layer proteins were obtained by dialyzing the supernatant twice against 5 l of distilled water or Tris-buffer over night at 4–8°C using dialysis tubes with an exclusion size of 14 kDa.
For further purification the suspension containing renatured S-layer proteins was centrifuged at 12,000 × g for 30 min, the pellet washed 3 times with distilled water or Tris-buffer and dissolved in 3 M Gua-HCl in Tris-buffer. After a final centrifugation step (40,000 × g; 30 min) the clear supernatant was subjected to FPLC using a Sephacryl™ S-300 column (Pharmacia Biotech). Size exclusion chromatography was performed under denaturing conditions (3 M Gua-HCl in Tris-buffer) at RT with a flow rate of 0,5 ml/min. Fractions containing mSbsC-EGFP were dialyzed against dH2O over night and used for further analysis.
Isolation and purification of recombinant S-layer proteins from S. cerevisiae
Constitutive expression of S-layer protein coding genes was obtained by growing transformants of S. cerevisiae strain BY4741 bearing plasmid p426-GPD-mSbsC-EGFP in YNB medium (Invitrogen) supplemented with Leu, His, Met and glucose as carbon source. 500 ml of an exponential culture (OD600 = 1) were centrifuged at 5,000 × g for 3 min and 20°C. Cells were washed 3 times with Tris-buffer and subsequently lysed by osmotic swelling of protoplasts obtained by treatment with zymolyase-20T (ICN) as described [19]. S-layer self assembly products were pelleted from the suspension by centrifugation at 20,000 × g.
Isolation and purification of recombinant S-layer proteins from HeLa cells
4 × 106 HeLa cells were transiently transfected with pmSbsC-EGFP-N1. After cultivation for 16 h cells were harvested by trypsin treatment followed by three washing steps with 1 × PBS w/o Ca2+, Mg2+ (PAA-Laboratories). Cell lysis was performed by 30 strokes in a dounce homogenizer in the presence of 0,1 % (v/v) Triton X-100, 1 mM AEBSF and Pi-cocktail (protease inhibitor mix; Roche). The cell lysate was centrifuged at 20,000 × g, and the resulting supernatant and pellet fractions were analysed in a Western blot.
Protein analysis
Sample preparation and SDS-polyacrylamide gelelectrophoresis were carried out as described by Laemmli [20]. Unless otherwise indicated 10 μg of protein were separated on 7,5% (w/v) acrylamide gels. For Westernblot analysis proteins were transferred onto a PVDF membrane (Millipore) and probed with monoclonal mouse antibody directed against the GFP epitope (Boehringer Mannheim). Detection of bound antibodies was performed with horseradish peroxidase (HRP)-conjugated secondary antibodies and the ECL-Plus Kit (Amersham Pharmacia Biotech). As positive control for immunreactivity of anti-GFP antibodies whole cell extract of EGFP expressing S. cerevisiae transformants was used (Fig. 4B).
Electron microscopy
Whole yeast cells and spheroplasts were fixed and prepared for electron microscopy as described by Waterham et al. [21]. Immunolabeling was performed on ultrathin sections of Unicryl-embedded cells using polyclonal rabbit anti-GFP antibodies and 15 nm colloidal gold-conjugated goat anti-rabbit secondary antibodies (Amersham). Negative staining of osmotically swollen protoplasts and FPLC-purified recrystallized mSbsC-EGFP was done with 3% (w/v) ammonium-molybdate, pH 7,2 and examined in a Philips CM10 Transmission Electron Microscope.
Miscellaneous procedures
Standard DNA techniques were as described [22]. Yeast cells were transformed by the lithium acetate method [23]. GENOMED™ columns were used for isolation of DNA fragments from agarose gels. The correct sequence of all constructs was confirmed by DNA sequencing with the dideoxy-chain termination method [24] using 5'-IRD800 labelled primers (MWG-BIOTECH). A Thermo Sequenase fluorescent labelled primer cycle sequencing kit with 7-deaza-dGTP (Amersham) was employed for sequencing with the LI-COR DNA sequencer 4000 (MWG-BIOTECH). Protein concentrations were determined by the Lowry method (BioRad).
Authors' contributions
AB carried out the molecular genetic studies, performed protein purification and drafted the manuscript.
KZ performed part of the DNA cloning and sequencing and carried out the transfection studies of HeLa cells.
KAS performed the immunogold labeling of yeast protoplasts and the TEM analysis. He provided the TEM images for publication.
MV contributed to the design of the study and conducted the electron microscopy analysis.
GR conceived of the study, participated in its design and coordination, and participated in drafting of the manuscript.
All authors read and approved the final manuscript.
Acknowledgements
This work was supported by a grant of the Saxonian Ministry of Science and the Fine Arts. We thank S. Selenska-Pobell, FZ Rossendorf for providing strain ATCC 12980 and S. Tokalov, Institut für Zoologie, Technische Universität Dresden for laser scan microscopy.
==== Refs
Sleytr UB Beveridge TJ Bacterial S-layers Trends Microbiol 1999 7 253 260 10366863 10.1016/S0966-842X(99)01513-9
Sleytr UB Basic and applied S-layer research: an overview FEMS Microbiol Rev 1997 20 5 12 10.1016/S0168-6445(97)00039-9
Thomas SR Trust TJ Tyrosine phosphorylation of the tetragonal paracrystalline array of Aeromonas hydrophila: molecular cloning and high-level expression of the S-layer protein gene J Mol Biol 1995 245 568 581 7844827 10.1006/jmbi.1994.0047
Messner P Prokaryotic glycoproteins: unexplored but important J Bacteriol 2004 186 2517 2519 15090489 10.1128/JB.186.9.2517-2519.2004
Schäffer C Messner P Glycobiology of surface layer proteins Biochimie 2001 83 591 599 11522387 10.1016/S0300-9084(01)01299-8
Sara M Sleytr UB S-Layer proteins J Bacteriol 2000 182 859 868 10648507 10.1128/JB.182.4.859-868.2000
Jarosch M Egelseer EM Huber C Moll D Mattanovich D Sleytr UB Sára M Analysis of the structure-function relationship of the S-layer protein SbsC of Bacillus stearothermophilus ATCC 12980 by producing truncated forms Microbiology 2001 147 1353 1363 11320138
Sleytr UB Bayley H Sara M Breitwieser A Kupcu S Mader C Weigert S Unger FM Messner P Jahn-Schmid B Schuster B Pum D Douglas K Clark NA Moore JT Winningham TA Levy S Frithsen I Pankovc J Beale P Gillis HP Choutov DA Martin KP Applications of S-layers FEMS Microbiol Rev 1997 20 151 175 9276930 10.1016/S0168-6445(97)00044-2
Bahl H Scholz H Bayan N Chami M Leblon G Gulik-Krzywicki T Shechter E Fouet A Mesnage S Tosi-Couture E Gounon P Mock M Conway de Macario E Macario AJ Fernandez-Herrero LA Olabarria G Berenguer J Blaser MJ Kuen B Lubitz W Sára M Pouwels PH Kolen CP Boot HJ Resch S Molecular biology of S-layers FEMS Microbiol Rev 1997 20 47 98 9276928 10.1016/S0168-6445(97)00050-8
Egelseer EM Schocher I Sleytr UB Sára M Evidence that an N-terminal S-layer protein fragment triggers the release of a cell-associated high-molecular-weight amylase in Bacillus stearothermophilus ATCC 12980 J Bacteriol 1996 178 5602 5609 8824603
Novotny R Scheberl A Giry-Laterriere M Messner P Schäffer C Gene cloning, functional expression and secretion of the S-layer protein SgsE from Geobacillus stearothermophilus NRS 2004/3a in Lactococcus lactis FEMS Microbiol Lett 2005 242 27 35 15675069 10.1016/j.femsle.2004.10.036
Jarosch M Egelseer EM Mattanovich D Sleytr UB Sára M S-layer gene sbsC of Bacillus stearothermophilus ATCC 12980: molecular characterization and heterologous expression in Escherichia coli Microbiology 2000 146 ( Pt 2) 273 281 10708365
Breitwieser A Egelseer EM Moll D Ilk N Hotzy C Bohle B Ebner C Sleytr UB Sára M A recombinant bacterial cell surface (S-layer)-major birch pollen allergen-fusion protein (rSbsC/Bet v1) maintains the ability to self-assemble into regularly structured monomolecular lattices and the functionality of the allergen Protein Eng 2002 15 243 249 11932495 10.1093/protein/15.3.243
Ilk N Küpcü S Moncayo G Klimt S Ecker RC Hofer-Warbinek R Egelseer EM Sleytr UB Sára M A functional chimaeric S-layer-enhanced green fluorescent protein to follow the uptake of S-layer-coated liposomes into eukaryotic cells Biochem J 2004 379 441 448 14725506 10.1042/BJ20031900
Archibald JM Logsdon JMJ Doolittle WF Origin and evolution of eukaryotic chaperonins: phylogenetic evidence for ancient duplications in CCT genes Mol Biol Evol 2000 17 1456 1466 11018153
Pum D Neubauer A Györvary E Sára M Sleytr UB S-layer proteins as basic building blocks in a biomolecular construction kit Nanotechnology 2000 11 100 107 10.1088/0957-4484/11/2/310
Schäffer C Kahlig H Christian R Schulz G Zayni S Messner P The diacetamidodideoxyuronic-acid-containing glycan chain of Bacillus stearothermophilus NRS 2004/3a represents the secondary cell-wall polymer of wild-type B. stearothermophilus strains Microbiology 1999 145 ( Pt 7) 1575 1583 10439396
Mumberg D Müller R Funk M Yeast vectors for the controlled expression of heterologous proteins in different genetic backgrounds Gene 1995 156 119 122 7737504 10.1016/0378-1119(95)00037-7
Daum G Gasser SM Schatz G Import of proteins into mitochondria. Energy-dependent, two-step processing of the intermembrane space enzyme cytochrome b2 by isolated yeast mitochondria J Biol Chem 1982 257 13075 13080 6752147
Laemmli UK Cleavage of structural proteins during the assembly of the head of bacteriophage T4 Nature 1970 227 680 685 5432063 10.1038/227680a0
Waterham HR Titorenko VI Haima P Cregg JM Harder W Veenhuis M The Hansenula polymorpha PER1 gene is essential for peroxisome biogenesis and encodes a peroxisomal matrix protein with both carboxy- and amino-terminal targeting signals J Cell Biol 1994 127 737 749 7962056 10.1083/jcb.127.3.737
Sambrook J Fritsch EF T M Molecular Cloning. A Laboratory Manual. 1989 Cold Spring Harbor Laboratory Press: New York.,
Schiestl RH Gietz RD High efficiency transformation of intact yeast cells using single stranded nucleic acids as a carrier Curr Genet 1989 16 339 346 2692852 10.1007/BF00340712
Sanger F Nicklen S Coulson AR DNA sequencing with chain-terminating inhibitors Proc Natl Acad Sci U S A 1977 74 5463 5467 271968
|
16202167
|
PMC1262761
|
CC BY
|
2021-01-04 16:24:35
|
no
|
Microb Cell Fact. 2005 Oct 4; 4:28
|
utf-8
|
Microb Cell Fact
| 2,005 |
10.1186/1475-2859-4-28
|
oa_comm
|
==== Front
Microb Cell FactMicrobial Cell Factories1475-2859BioMed Central London 1475-2859-4-291621266510.1186/1475-2859-4-29ReviewDirected evolution strategies for improved enzymatic performance Hibbert Edward G [email protected] Paul A [email protected] The Advanced Centre for Biochemical Engineering, Department of Biochemical Engineering, University College London, Torrington Place, London WC1E 7JE, UK2005 7 10 2005 4 29 29 26 7 2005 7 10 2005 Copyright © 2005 Hibbert and Dalby; licensee BioMed Central Ltd.2005Hibbert and Dalby; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The engineering of enzymes with altered activity, specificity and stability, using directed evolution techniques that mimic evolution on a laboratory timescale, is now well established. However, the general acceptance of these methods as a route to new biocatalysts for organic synthesis requires further improvement of the methods for both ease-of-use and also for obtaining more significant changes in enzyme properties than is currently possible. Recent advances in library design, and methods of random mutagenesis, combined with new screening and selection tools, continue to push forward the potential of directed evolution. For example, protein engineers are now beginning to apply the vast body of knowledge and understanding of protein structure and function, to the design of focussed directed evolution libraries, with striking results compared to the previously favoured random mutagenesis and recombination of entire genes. Significant progress in computational design techniques which mimic the experimental process of library screening is also now enabling searches of much greater regions of sequence-space for those catalytic reactions that are broadly understood and, therefore, possible to model.
Biocatalysis for organic synthesis frequently makes use of whole-cells, in addition to isolated enzymes, either for a single reaction or for transformations via entire metabolic pathways. As many new whole-cell biocatalysts are being developed by metabolic engineering, the potential of directed evolution to improve these initial designs is also beginning to be realised.
==== Body
Introduction
Natural enzymes can catalyse reactions with up to 1017 fold rate accelerations [1], and with exquisite control of regio- and stereo-chemistry. This along with their compatibility with mild aqueous conditions has led to their increasing use as biocatalysts in synthetic chemistry, especially in cases where chemical routes are difficult to implement [2,3]. Enzymes used in such biotransformations are frequently prepared as isolated enzymes in solution or immobilised onto resins, and used in the presence of organic solvents, harsh chemical compounds, or under conditions of temperature or pH that are suboptimal for enzyme activity. Such non-natural conditions also often result in poor enzyme activity, or complete deactivation due to denaturation or chemical modifications. Developments in protein engineering over the past ten years have enabled enzymes to be evolved in vitro for properties that favour the required process conditions, and also to obtain enzyme variants with altered substrate specificity or enantioselectivity [4,5]. Despite the significant advances to date on many industrially relevant enzymes, there still remains a need to improve directed evolution strategies and develop generic screening or selection tools which make the process of identifying novel enzyme activities more efficient, and also to access much greater changes to enzyme function.
Just as Nature evolves enzymes without any 'knowledge' of enzyme structure and function, the techniques of in vitro directed evolution mimic natural processes such as random mutagenesis [6] and sexual recombination [7-10] to improve enzymes without understanding them in great detail.
This review discusses recent advances in the techniques and strategies used for the directed evolution of biocatalytic enzymes, including the development of new genetic and computational strategies aimed at improving the quality and potential of enzyme libraries, as well as new screening tools that broaden the range of targets for directed evolution. Also discussed is the use of directed evolution for enhancing metabolically engineered pathways, and some future applications arising from novel pathway engineering in E. coli. Examples of the application of established methods to new enzymes are not discussed.
New mutagenic strategies for directed enzyme evolution
Despite the rapid growth of published examples of directed evolution, there is still a clear need for alternative and improved methods for the directed evolution of enzymes. Current constraints include the difficulty in optimising ligation steps when large libraries (>106 variants) are sought for selection-based methods, practical limitations to library sizes that can be screened, and barriers to technology licensing [5]. Some improvements have been made in circumventing the need for ligations by adopting PCR-based approaches [11], and more recently by directing in vivo hypermutation with B cells to target genes delivered by retroviral infection [12]. However, ligation free approaches for DNA shuffling have yet to be demonstrated. Practical limitations to library sizes have been partially addressed recently by improving the proportion of non-redundant or degenerate variants in libraries. For example, the analysis of the frequency and distribution of beneficial single mutants obtained from initial libraries, can be used to define the ratio of templates to be used in recombination by overlap extension PCR [13]. The effect is to improve the diversity of the multiple mutation variants obtained in the shuffled library. Another recent study has examined why libraries constructed using high-mutation frequencies, tend to yield a higher than expected number of functional variants [14]. Increased mutation rates permit the synergistic effects of multiple mutations to be identified more frequently, though they also lead to an increased likelihood of negating positive mutations, or non-functional variants. It was demonstrated that increased mutation rates lead to more unique variants in each library, whereas single mutant libraries can contain many copies of the same variant. This work leads to the possibility of finding an optimal mutagenic load for a given mutagenic method or application.
Another strategy, previously suggested as having the potential to obtain more useful variants from restricted library sizes, is to focus mutations in regions of the enzyme more likely to result in beneficial mutations [4]. In the same review it was also noted that recent examples of the directed evolution of properties traditionally associated with the active site, were producing the majority of mutations in regions that contribute to substrate binding, catalysis or the conformation and dynamics of the active site environment. A more recent and extensive study of previously published rational design and directed evolution experiments demonstrates that indeed by far the majority of mutations that improve the enantioselectivity of enzymes, occur within 10 Å of the enzyme active site [15]. The authors also compared random mutagenesis of five residues in the P. fluorescens esterase active site, to single random mutations across the entire enzyme, demonstrating a five-fold improvement of enantioselectivity enhancement for the focussed approach compared to the more random method. In a similar study, four residues which in tetrameric form comprise the sixteen-residue active-site of dihydrofolate reductase (DHFR), were mutated by cassette mutagenesis [16]. The resulting library yielded three mutants with entirely altered active-sites and showing increased activity. More recently, a technique dubbed CASTing (combinatorial active-site saturation), in which pair-wise saturation mutagenesis of residues adjacent in sequence, was focussed into the active site of a lipase from Pseudomonas aeruginosa, yielded a number of mutants active on substrates not previously accepted by the wild type [17]. In another striking example, Parikh et al. compared the site-saturation mutagenesis of three carefully chosen active-site residues in E. coli β-galactosidase, to a previous DNA shuffling experiment for the same enzyme [18]. The previous DNA shuffling experiment enhanced the kcat/Km for β-fucosidase by 10-fold after seven rounds, whereas the saturation mutagenesis technique resulted in a 180-fold improvement in a single round. Not all enzyme properties can be expected to improve through active-site mutations alone, however. Indeed, thermostability was shown to be improved equally by mutations close to and distant from the active site [19].
One further promising approach for obtaining more efficient searches of sequence-space is the use of consensus-sequence data for constructing libraries [20]. By aligning the target gene of β-lactamase from Enterobacter cloacae with the consensus sequence from 38 homologues, 29 residues were identified as differing from the consensus. Each of these sites was simultaneously mutated back to the consensus sequence, using the QuikChange multi site-directed mutagenesis kit (QCMS) (Stratagene), to produce a combinatorial library. Screening of just 90 variants yielded 15 variants with improved thermostability and subsequent recombination led to further improvements. This demonstrated the potential power of refined library design, though it is yet to be seen whether this type of approach can be applied to properties other than thermostability.
New screening and selection strategies for directed enzyme evolution
The available methods for screening of, and selection from enzyme libraries have recently been reviewed [21]. New screens are required to enable the identification of improved enzymes from larger libraries, and also to obtain the desired properties with generic methods that measure it directly. The latter issue addresses the often quoted first law of directed evolution, ie. 'you get what you screen for' [22]. A frequent target for the directed evolution of enzymes is the improvement of thermostability which leads to more robust biocatalysts [4], and increased stability in organic solvents, as shown in a recent study on fructose bisphosphate aldolase [23]. Most screens for thermostability have made use of indirect measures, such as resistance to thermoinactivation at high temperatures [23]. Such a screen, though effective in many cases, is not a direct measure of protein stability and is unsuitable for proteins that are reversibly unfolded, or those that are likely to become reversibly unfolded upon mutation [24], thus leading to false positives. To enable a more direct screen for protein stability, the measurement of protein denaturation curves using tryptophan fluorescence in microplates has been explored [25]. The results have shown that using autotitration of denaturant directly into the microwells can yield transition midpoints (C1/2) with an accuracy of ± 0.15 M and a throughput of up to 1000 samples per day. Linkage with automated protein purification has the potential to enable application of the screen to directed evolution libraries.
Screening for improved enzyme activity often leads to loss of substrate selectivity (or vice versa). The ability to screen directly for both activity and selectivity would enable more useful enzyme variants to be found. Recently, a cell-surface display approach that uses multiparameter flow cytometry (FACS), demonstrates the benefits of simultaneously screening for activity and selectivity, using the E. coli endopeptidase Omp T as an example [26]. FACS-based methods permit the screening of up to 107 cells per hour, enabling large areas of sequence-space to be searched. Combined with the ability to cell sort based upon two different fluorescent reporters, the authors have shown that large numbers of enzyme variants can be screened under simultaneous positive and negative selection pressures to obtain protease mutants with both improved activity to the new Ala-Arg substrate and reduced activity to the Arg-Arg substrate preferred by wild type. Although the particular FACS technique used is limited to cell-surface displayed proteases, the concept of dual screening in this manner could potentially be applied to other enzymes screened by more traditional methods.
The use of selection-based methods has the potential to identify novel enzyme variants from much larger libraries (106-1013 variants), than for screening methods (up to 105-107) and have been reviewed in detail [21,27,28]. Consequently, deeper searches of sequence-space can be performed to access better enzyme variants. Selection of enzymes has been achieved using complementation of the deleted activities in auxotrophic strains [29], enrichment of active beta-lactamase from a background of an inactive point mutant by ribosome display and selection for binding of a mechanism-based inhibitor [30], enrichment of phage display libraries by affinity capture of the phage-bound turnover product [31], and enrichment of phage display libraries by selection for transition-state analogue or suicide inhibitor binding [32-34]. While genetic selection with auxotrophs is limited to the range of activities found to be essential to survival of the host cells, display methods have to potential to explore more novel activities perhaps not previously found in Nature as they rely mostly on the design of good transition-state analogues (TSA) or suicide inhibitors for the desired activity. However, mechanism-based inhibitors do not necessarily represent a full catalytic turnover, and TSA structures may not accurately reflect the true transition state structure in the desired reaction mechanism. Recently, selection for a complete turnover has been achieved using phage display [35]. The product formed after phosphatase turnover is spontaneously converted into an electrophilic reagent that can capture the nearby phage particle. While this approach has enabled significant enhancement of catalytic activity compared to the TSA binding methods it is still to be seen how generally applicable such methods can become and will presumably require a good deal of inventive chemistry to identify suitable capture reagents that spontaneously form after the turnover of other desired reactions.
Computational approaches for enzyme evolution and design
The last two years have seen a revolution in the use of computational approaches to search sequence space in a combinatorial manner that is analogous to experimental screening of directed evolution libraries. Previously, computational design was used to eliminate the vast proportion of sequences that were incompatible with the protein fold, before experimentally screening the remaining variants for improved activity [36]. The introduction of new activity into protein scaffolds was also achieved using the careful placement of potentially catalytic residues into models and the computational search of variants with improved binding affinity to high-energy reaction intermediates [37,38]. Since these groundbreaking efforts, the use of computational design has expanded to include the thermostabilization of enzymes [39] and the redesign of an enzyme active-site for improved catalytic activity [40]. As computational processing power continues to increase, and protein modelling algorithms become further refined, computational design should soon be capable of tackling more complex enzyme mechanisms and also of dramatically refining experimental library approaches.
Directed evolution of metabolic pathways
The use of whole cell biocatalysts potentially enables novel molecular synthesis via entire metabolic pathways, involving multiple enzymes. Consequently, directed evolution could also be applied to natural or engineered metabolic pathways, multiple enzyme systems (mini-pathways), and even whole organisms [41]. The directed evolution of the three-enzyme arsenate resistance pathway in E. coli, for increased resistance to arsenate, was the first example of using DNA shuffling within a metabolic pathway [42]. Since then, the approach has been combined with that of metabolic engineering to obtain entirely new pathways and products. For example, carotenoids can be synthesised in E. coli by expression of genes from the isoprenoid pathways of Archaeoglobus fulgidus and Agrobacterium aurantiacum. Directed evolution was successfully applied to optimize the expression level of the geranylgeranyl diphosphate (GGPP) synthase gene, increasing the production of astaxanthin [43]. In parallel work, neurosporene was produced by co-exression of isoprenoid pathway genes from R.sphaeroides, in E.coli. Directed evolution of the R.sphaeroides phytoene desaturase yielded variants capable of producing lycopene [44]. A similar independent study yielded a metabolically engineered E. coli that could produce lycopene. DNA shuffling of the phytoene desaturase genes from E. uredovora and E. herbicola resulted in variants of E. coli capable of tetradehydrolycopene production. Further extension of the pathway with shuffled lycopene cyclase (crtY) genes from E. uredovora and E. herbicola yielded the production of torulene [45]. Previously unknown C-45 and C-50 carotenoids have also been synthesised by directed evolution of the C-30 carotenoid synthase (crtM) gene [46]. More recently, a carotenoid desaturase (CrtOx) homolog from S. aureus has been coexpressed in the previously evolved tetradehydrolycopene producing strain, to yield the C-40 carotenoid tetradehydrolycopendial [47].
The directed evolution of enzymes within metabolically engineered pathways can benefit product yields via a number of mechanisms, including the reduction of product inhibition at a key enzyme step. Metabolic engineering has been used to enhance glucosamine production in E. coli by 15-fold to 60 mg L-1 [48]. The directed evolution by error-prone PCR of the overexpressed glucosamine synthase gene (GlmS), and screening for reduced product inhibition by glucosamine-6-P, yielded an E. coli strain capable of producing glucosamine to 17 g L-1.
Conclusion
Directed evolution using genetic techniques has led the way for engineering altered proteins during the last decade, yet there is still considerable scope for developing experimentally simpler and also more efficient techniques and strategies. Recent advances have addressed the issue of library redundancy in terms of the numbers of unique sequences as a function of mutation frequency, and also in terms of focussing random mutagenesis to regions of enzymes more likely to elicit the desired effect. An alternative to improving such library quality is to screen larger libraries by more efficient means. The power of cell surface-display techniques for the selection of novel enzymes is greatly improving the library size and hence the sequence space that can be searched. The ability of this method to perform multiple simultaneous measurements will also potentially improve the quality and usefulness of the enzyme variants isolated from libraries. The use of phage- and ribosome-display methods also has the potential to search much larger variant libraries. While, mechanisms for the affinity-based capture of active enzyme variants are continually being developed and improved, there is still some way to go in widening the general applicability of these methods to more useful enzyme activities.
Computational approaches are also improving rapidly and will become very useful in either creating novel enzymes as starting points for directed evolution, or for defining smarter libraries that contain fewer redundant enzyme variants. While for simple reactions the capability of computational methods is approaching that of genetic techniques, there is still considerable effort required to extend its use towards obtaining novel enzymes with more complex catalytic mechanisms.
Finally, advances in the metabolic engineering of whole cell biocatalysts is stimulating the use of directed evolution to improve metabolic pathways. Considerable progress has been made, especially in the synthesis of novel carotenoids. This area also opens up new targets for directed evolution, such as the reduction of product inhibition on a single enzyme which in turn improves the yield of product from the whole pathway, as demonstrated for the production of glucosamine.
List of abbreviations
CAST Combinatorial active-site saturation
DHFR Dihydrofolate reductase
FACS Fluorescence-activated cell sorting
GGPP Geranylgeranyl diphosphate
PCR Polymerase chain reaction
QCMS QuikChange multi site
TSA Transition state analogue
Authors' contributions
EH drafted the manuscript
PAD helped draft and critically revised the manuscript
Both authors have read and approved the final manuscript
Acknowledgements
The authors would like to thank the UK Engineering and Physical Sciences Research Council (EPSRC) for support of the multidisciplinary Biocatalysis Integrated with Chemistry and Engineering (BiCE) programme (GR/S62505/01). The authors would also like to thank the UK Joint Infrastructure Fund (JIF), the Science Research Investment Fund (SRIF) and the Gatsby Charitable Foundation for funds to establish the UCL Centre for Micro Biochemical Engineering.
==== Refs
Radzicka A Wolfenden R A proficient enzyme Science 1995 267 90 93 7809611
Schoemaker HE Mink D Wubbolts MG Dispelling the myths--biocatalysis in industrial synthesis Science 2003 299 1694 1697 12637735 10.1126/science.1079237
Straathof AJ Panke S Schmid A The production of fine chemicals by biotransformations Curr Opin Biotechnol 2002 13 548 556 12482513 10.1016/S0958-1669(02)00360-9
Dalby PA Optimising enzyme function by directed evolution Curr Opin Struct Biol 2003 13 500 505 12948780 10.1016/S0959-440X(03)00101-5
Hibbert EG Baganz F Hailes HC Ward JM Lye GJ Woodley JM Dalby PA Directed evolution of biocatalytic processes Biomolecular Engineering 2005 22 11 19 15857779 10.1016/j.bioeng.2004.09.003
Chen K Arnold FH Tuning the activity of an enzyme for unusual environments: sequential random mutagenesis of subtilisin E for catalysis in dimethylformamide Proc Natl Acad Sci U S A 1993 90 5618 5622 8516309
Stemmer WPC Rapid evolution of a protein in vitro by DNA shuffling Nature 1994 370 389 391 8047147 10.1038/370389a0
Zhao H Giver L Shao Z Affholter JA Arnold FH Molecular evolution by staggered extension process (StEP) in vitro recombination [see comments] Nat Biotechnol 1998 16 258 261 9528005 10.1038/nbt0398-258
Coco WM Levinson WE Crist MJ Hektor HJ Darzins A Pienkos PT Squires CH Monticello DJ DNA shuffling method for generating highly recombined genes and evolved enzymes Nature Biotech 2001 19 354 359 10.1038/86744
Ostermeier M Shim JH Benkovic SJ A combinatorial approach to hybrid enzymes independent of DNA homology Nat Biotechnol 1999 17 1205 1209 10585719 10.1038/10850
Miyazaki K Takenouchi M Creating random mutagenesis libraries using megaprimer PCR of whole plasmid Biotechniques 2002 33 1033 1038 12449380
Wang CL Yang DC Wabl M Directed molecular evolution by somatic hypermutation Protein Engineering Design & Selection 2004 17 659 664 10.1093/protein/gzh080
Hamamatsu N Aita T Nomiya Y Uchiyama H Nakajima M Husimi Y Shibanaka Y Biased mutation-assembling: an efficient method for rapid directed evolution through simultaneous mutation accumulation Protein Engineering Design & Selection 2005 18 265 271 10.1093/protein/gzi028
Drummond DA Iverson BL Georgiou G Arnold FH Why high-error-rate random mutagenesis libraries are enriched in functional and improved proteins J Mol Biol 2005 350 806 816 15939434 10.1016/j.jmb.2005.05.023
Park S Morley KL Horsman GP Holmquist M Hult K Kazlauskas RJ Focusing mutations into the P. fluorescens esterase binding site increases enantioselectivity more effectively than distant mutations Chemistry & Biology 2005 12 45 54 15664514 10.1016/j.chembiol.2004.10.012
Schmitzer AR Lepine F Pelletier JN Combinatorial exploration of the catalytic site of a drug- resistant dihydrofolate reductase: creating alternative functional configurations Protein Engineering Design & Selection 2004 17 809 819 10.1093/protein/gzh090
Reetz MT Bocola M Carballeira JD Zha DX Vogel A Expanding the range of substrate acceptance of enzymes: Combinatorial active-site saturation test Angewandte Chemie-International Edition 2005 44 4192 4196 10.1002/anie.200500767
Parikh MR Matsumura I Site-saturation mutagenesis is more efficient than DNA shuffling for the directed evolution of b-fucosidase from b-galactosidase J Mol Biol 2005 352 621 628 16095606 10.1016/j.jmb.2005.07.020
Morley KL Kazlauskas RJ Improving enzyme properties: when are closer mutations better? Trends in Biotechnology 2005 23 231 237 15866000 10.1016/j.tibtech.2005.03.005
Amin N Liu AD Ramer S Aehle W Meijer D Metin M Wong S Gualfetti P Schellenberger V Construction of stabilized proteins by combinatorial consensus mutagenesis Protein Engineering Design & Selection 2004 17 787 793 10.1093/protein/gzh091
Aharoni A Griffiths AD Tawfik DS High-throughput screens and selections of enzyme-encoding genes Current Opinion in Chemical Biology 2005 9 210 216 15811807 10.1016/j.cbpa.2005.02.002
Schmidt-Dannert C Arnold FH Directed evolution of industrial enzymes Trends Biotechnol 1999 17 135 136 10203769 10.1016/S0167-7799(98)01283-9
Hao JJ Berry A A thermostable variant of fructose bisphosphate aldolase constructed by directed evolution also shows increased stability in organic solvents Protein Engineering Design & Selection 2004 17 689 697 10.1093/protein/gzh081
Gray KA Richardson TH Kretz K Short JM Bartnek F Knowles R Kan L Swanson PE Robertson DE Rapid evolution of reversible denaturation and elevated melting temperature in a microbial haloalkane dehalogenase Advanced Synthesis & Catalysis 2001 343 607 617 10.1002/1615-4169(200108)343:6/7<607::AID-ADSC607>3.3.CO;2-D
Aucamp JP Cosme AM Lye GJ Dalby PA High-throughput measurement of protein stability in microtiter plates Biotechnology and Bioengineering 2005 89 599 607 15672379 10.1002/bit.20397
Varadarajan N Gam J Olsen MJ Georgiou G Iverson BL Engineering of protease variants exhibiting high catalytic activity and exquisite substrate selectivity Proc Natl Acad Sci U S A 2005 102 6855 6860 15867160 10.1073/pnas.0500063102
Forrer P Jung S Pluckthun A Beyond binding: using phage display to select for structure, folding and enzymatic activity in proteins Curr Opin Struct Biol 1999 9 514 520 10449374 10.1016/S0959-440X(99)80073-6
Fernandez-Gacio A Uguen M Fastrez J Phage display as a tool for the directed evolution of enzymes Trends in Biotechnology 2003 21 408 414 12948674 10.1016/S0167-7799(03)00194-X
Evnin LB Vasquez JR Craik CS Substrate specificity of trypsin investigated by using a genetic selection Proc Natl Acad Sci U S A 1990 87 6659 6663 2204062
Amstutz P Pelletier JN Guggisberg A Jermutus L Cesaro-Tadic S Zahnd C Pluckthun A In vitro selection for catalytic activity with ribosome display Journal of the American Chemical Society 2002 124 9396 9403 12167034 10.1021/ja025870q
Demartis S Huber A Viti F Lozzi L Giovannoni L Neri P Winter G Neri D A strategy for the isolation of catalytic activities from repertoires of enzymes displayed on phage J Mol Biol 1999 286 617 633 9973575 10.1006/jmbi.1998.2476
Arkin MR Wells JA Probing the importance of second sphere residues in an esterolytic antibody by phage display J Mol Biol 1998 284 1083 1094 9837728 10.1006/jmbi.1998.2234
Goud GN Artsaenko O Bols M Sierks M Specific glycosidase activity isolated from a random phage display antibody library Biotechnol Prog 2001 17 197 202 11170499 10.1021/bp0001258
Widersten M Mannervik B Glutathione Transferases with Novel Active-Sites Isolated by Phage Display from A Library of Random Mutants Journal of Molecular Biology 1995 250 115 122 7608963 10.1006/jmbi.1995.0362
Cesaro-Tadic S Lagos D Honegger A Rickard JH Partridge LJ Blackburn GM Pluckthun A Turnover-based in vitro selection and evolution of biocatalysts from a fully synthetic antibody library Nature Biotechnology 2003 21 679 685 12754520 10.1038/nbt828
Hayes RJ Bentzien J Ary ML Hwang MY Jacinto JM Vielmetter J Kundu A Dahiyat BI Combining computational and experimental screening for rapid optimization of protein properties Proc Natl Acad Sci U S A 2002 99 15926 15931 12446841 10.1073/pnas.212627499
Bolon DN Mayo SL Enzyme-like proteins by computational design Proc Natl Acad Sci U S A 2001 98 14274 14279 11724958 10.1073/pnas.251555398
Dwyer MA Looger LL Hellinga HW Computational design of a biologically active enzyme Science 2004 304 1967 1971 15218149 10.1126/science.1098432
Korkegian A Black ME Baker D Stoddard BL Computational thermostabilization of an enzyme Science 2005 308 857 860 15879217 10.1126/science.1107387
Lassila JK Keeffe JR Oelschlaeger P Mayo SL Computationally designed variants of Escherichia coli chorismate mutase show altered catalytic activity Protein Engineering Design & Selection 2005 18 161 163 10.1093/protein/gzi015
Schmidt-Dannert C Directed evolution of single proteins, metabolic pathways, and viruses Biochemistry 2001 40 13125 13136 11683620 10.1021/bi011310c
Crameri A Dawes G Rodriguez EJ Silver S Stemmer WP Molecular evolution of an arsenate detoxification pathway by DNA shuffling [see comments] Nat Biotechnol 1997 15 436 438 9131621 10.1038/nbt0597-436
Wang CW Oh MK Liao JC Directed evolution of metabolically engineered Escherichia coli for carotenoid production Biotechnology Progress 2000 16 922 926 11101317 10.1021/bp000124f
Wang CW Liao JC Alteration of product specificity of Rhodobacter sphaeroides phytoene desaturase by directed evolution Journal Of Biological Chemistry 2001 276 41161 41164 11526111 10.1074/jbc.M105786200
Schmidt-Dannert C Umeno D Arnold FH Molecular breeding of carotenoid biosynthetic pathways Nat Biotechnol 2000 18 750 753 10888843 10.1038/77319
Umeno D Arnold FH Evolution of a pathway to novel long-chain carotenoids Journal of Bacteriology 2004 186 1531 1536 14973014 10.1128/JB.186.5.1531-1536.2004
Mijts BN Lee PC Schmidt-Dannert C Identification of a carotenoid oxygenase synthesizing acyclic xanthophylls: Combinatorial biosynthesis and directed evolution Chemistry & Biology 2005 12 453 460 15850982 10.1016/j.chembiol.2005.02.010
Deng MD Severson DK Grund AD Wassink SL Burlingame RP Berry A Running JA Kunesh CA Song L Jerrell TA Rosson RA Metabolic engineering of Escherichia coli for industrial production of glucosamine and N-acetylglucosamine Metabolic Engineering 2005 7 201 214 15885618 10.1016/j.ymben.2005.02.001
|
16212665
|
PMC1262762
|
CC BY
|
2021-01-04 16:24:35
|
no
|
Microb Cell Fact. 2005 Oct 7; 4:29
|
utf-8
|
Microb Cell Fact
| 2,005 |
10.1186/1475-2859-4-29
|
oa_comm
|
==== Front
Nucl ReceptNuclear Receptor1478-1336BioMed Central London 1478-1336-3-21619754710.1186/1478-1336-3-2ResearchEvolutionary selection across the nuclear hormone receptor superfamily with a focus on the NR1I subfamily (vitamin D, pregnane X, and constitutive androstane receptors) Krasowski Matthew D [email protected] Kazuto [email protected] Lee R [email protected] Erin G [email protected] Department of Pathology, Children's Hospital of Pittsburgh, 5834 Main Tower, 200 Lothrop Street, University of Pittsburgh, Pittsburgh, PA, 15213 USA2 Department of Pharmaceutical Sciences, St. Jude Children's Research Hospital, Memphis, TN, 38105 USA3 Department of Medicine, University of California, San Diego, CA, 92093, USA2005 30 9 2005 3 2 2 8 7 2005 30 9 2005 Copyright © 2005 Krasowski et al; licensee BioMed Central Ltd.2005Krasowski et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The nuclear hormone receptor (NR) superfamily complement in humans is composed of 48 genes with diverse roles in metabolic homeostasis, development, and detoxification. In general, NRs are strongly conserved between vertebrate species, and few examples of molecular adaptation (positive selection) within this superfamily have been demonstrated. Previous studies utilizing two-species comparisons reveal strong purifying (negative) selection of most NR genes, with two possible exceptions being the ligand-binding domains (LBDs) of the pregnane X receptor (PXR, NR1I2) and the constitutive androstane receptor (CAR, NR1I3), two proteins involved in the regulation of toxic compound metabolism and elimination. The aim of this study was to apply detailed phylogenetic analysis using maximum likelihood methods to the entire complement of genes in the vertebrate NR superfamily. Analyses were carried out both across all vertebrates and limited to mammals and also separately for the two major domains of NRs, the DNA-binding domain (DBD) and LBD, in addition to the full-length sequences. Additional functional data is also reported for activation of PXR and the vitamin D receptor (VDR; NR1I1) to gain further insight into the evolution of the NR1I subfamily.
Results
The NR genes appear to be subject to strong purifying selection, particularly in the DBDs. Estimates of the ratio of the non-synonymous to synonymous nucleotide substitution rates (the ω ratio) revealed that only the PXR LBD had a sub-population of codons with an estimated ω ratio greater than 1. CAR was also unusual in showing high relative ω ratios in both the DBD and LBD, a finding that may relate to the recent appearance of the CAR gene (presumably by duplication of a pre-mammalian PXR gene) just prior to the evolution of mammals. Functional analyses of the NR1I subfamily show that human and zebrafish PXRs show similar activation by steroid hormones and early bile salts, properties not shared by sea lamprey, mouse, or human VDRs, or by Xenopus laevis PXRs.
Conclusion
NR genes generally show strong sequence conservation and little evidence for positive selection. The main exceptions are PXR and CAR, genes that may have adapted to cross-species differences in toxic compound exposure.
==== Body
Background
Nuclear hormone receptors (NRs) are ligand-activated transcription factors that work in concert with co-activators and co-repressors to regulate gene expression [1-3]. NRs share a modular domain structure, which includes, from N-terminus to C-terminus, a modulatory A/B domain, the DNA-binding domain (DBD; C domain), the hinge D domain, the ligand-binding domain (LBD; E domain) and a variable C-terminal F domain that is absent in some NRs [3]. Examples of ligands for NRs include a range of endogenous compounds such as steroid hormones, thyroid hormone, and retinoids [3,4]. A few NRs, such as the 'xenobiotic sensors' pregnane X receptor (PXR, NR1I2) and constitutive androstane receptor (CAR or NR1I3), are activated by structurally diverse exogenous ligands [5-7].
The NR superfamily in mammals is composed of approximately 50 functional genes, with 48 genes in humans, 47 in rats, and 49 in mice [8]. Bony fish have a somewhat larger complement of NR genes due to gene duplication, exemplified by the 68 NR genes found in the genome of the pufferfish Fugu rubripes [9]. The current official nomenclature for NRs divides the superfamily into 7 families (NR0-6) [10,11]. The NR0 family, represented in humans by DAX-1 (dosage-sensitive sex and AHC critical region on the X chromosome; NR0B1) and SHP (small heterodimer partner; NR0B2) are unusual in essentially being 'domain singletons' that lack a DBD [12,13]. NRs have been the focus of a number of evolutionary studies including detailed investigations into the origins of the superfamily [11,14-16] and the development of ligand selectivity by the sex and adrenocortical steroid hormone receptors [17-20].
A major focus of molecular phylogenetics has been a search for evidence of positive selection (molecular adaptation) [21]. A variety of computational techniques have been developed over the last several decades to detect nucleotide variation between different genes suggestive of positive selection [21,22]. For comparisons within coding regions, the most common approach is to compare nucleotide variation that is non-synonymous (i.e., changes amino acid sequence encoded for by codons) or synonymous (does not changes amino acid sequence). Synonymous variation is considered to be neutral, an assumption which is generally true although there are exceptions [23]. The ratio of the rate of non-synonymous versus the rate of synonymous nucleotide variation (i.e., how many non-synonymous or synonymous changes have occurred in comparison to the total number of non-synonymous or synonymous changes possible; dN/dS or ω) provides some indication into selective forces acting on a given gene. For most gene comparisons, ω is less than one, often less than 0.1, reflective of negative or purifying selection to maintain a conserved amino acid sequence. ω = 1 reflects neutral selection (a ratio that would be expected for a non-functional pseudogene) while ω > 1 suggests positive selection. A large-scale comparison of 3,595 groups of homologous genes revealed that less than 0.5% had ω ratios greater than 1, with many of these genes being found in microorganisms [24]. Given that comparisons between full-length gene sequences rarely result in ω ratios greater than 1, techniques have been developed to detect sub-populations of codons that have elevated ω ratios. Different mathematical approaches have been applied to achieve this goal, including maximum likelihood [25,26] and Bayesian [27] methods. In this study, we employed the PAML (Phylogenetic Analysis by Maximum Likelihood) software, developed by Yang and colleagues [28], as this methodology is robust and has an extensive published literature associated with its application in biomedical research [21].
Most of the NR genes are strongly conserved between vertebrate species. Not surprisingly, previous studies utilizing two-species comparisons between human, mouse, and rat or humans, chimpanzee, and mouse genomes revealed that the NR genes are in general subject to negative selection [8,29], with only a few possible exceptions such as the LBDs of PXR and CAR [8,30]. A more detailed phylogenetic analysis of PXR, CAR, and the other member of the NR1I subfamily, the 1,25-(OH)2-vitamin D3 receptor (VDR; NR1I1), within mammals and across vertebrates, showed ω ratios for the CAR and PXR LBDs markedly higher than that for the VDR LBD [30]. For the PXR LBD analyses within mammals, the ω ratio exceeded one for a sub-population of codons comprising approximately 5% of the total codons in the LBD [30].
Given that a major function of PXR and CAR is to detect toxic endogenous and xenobiotic compounds that likely differ between species [5,6], these two genes may represent unusual examples of NR genes that have undergone positive selection in their LBDs for functional advantage. The aim of this study was to apply detailed phylogenetic analysis to the entire superfamily of NR genes in vertebrates to detect possible signatures of positive selection. This phylogenetic analysis, combined with functional analyses of PXRs and VDRs, provides a detailed context into how unusual the nucleotide variation of PXR and CAR is to the rest of the superfamily.
Results
Sequences available for phylogenetic analysis
Data from genome sequencing projects (e.g., human, chimpanzee, mouse, rat, dog, chicken, Xenopus tropicalis, Fugu rubripes, and Tetraodon nigroviridis) has greatly increased the number of NR coding sequences publicly available for phylogenetic analysis across vertebrates. The complete set of species and accession numbers for the NR genes analyzed in this study is provided in Additional file 1: Genes used for phylogenetic analysis. Complete nucleotide sequences in PAML format are provided in Additional file 2: Sequences used for phylogenetic analysis by PAML. To improve the power of accurately detecting positive selection and to minimize the risk of false positives, PAML analyses were only performed if sequence data from at least six species from at least six separate genera were available [22,31-33]. Given that some of the sequence data was partial and only contained complete data for the DBD or LBD (and not full-length sequence in those instances), the number of species available for the various analyses for each gene (i.e., full-length, DBD only, and LBD only) differs in some cases. The number of sequences varies widely across the NR superfamily, mainly because some receptors have been more intensively studied than others. For example, 33 full-length sequences are available for analysis of the estrogen receptor-α (ERα, NR3A1), while only 6 are available for estrogen-related receptor-β (ERRβ, NR3B2).
PAML analyses of the NR superfamily
The PAML analyses used correspond to models M0, M3 (with ncatG = 2, 3, 4, where ncatG is the number of populations of codons with distinct ω ratios), M7, and M8 within the PAML software (see Materials and Methods) [25,26]. Models M0 and M3 are 'discrete' in that they assign codons to population(s) of distinct ω ratios. For instance, an analysis of a particular gene may assign 95% of codons to an ω ratio of 0.05 (consistent with purifying selection) and the remaining 5% to an ω ratio exceeding 1 (suggestive of positive selection). For each M0/M3 analysis, the 'best minimum model' was determined. An M3 analysis was chosen only if it was statistically superior to the closest simpler model (e.g., M3 with 2 ω ratios versus M0). The ω ratios in the M0 or M3 analyses may be any value 0 or greater. M7 assigns codons within a gene to ω ratios along a β distribution function between 0 and 1 with parameters α and β. Depending on the parameters α and β, the β distribution may have most ω ratios clustered near 0 or be distributed more evenly between 0 and 1. M8 is a model where some codons have ω ratios that fall along a β distribution but the remaining codons form a separate population with a discrete ω ratio that may be any value 0 or greater (including greater than 1). M8 can detect the presence of positive selection (i.e., the extra ω class can be greater than 1) whereas M7 cannot.
M0 and M3 analyses
The complete results for all PAML analyses are provided in Additional file 3: Results of PAML analysis and treefiles. The distribution of frequency and ω ratios for all NR genes is plotted in Figure 1 (parts A, B, and C correspond to full-length sequence, DBD only, and LBD only analyses, respectively; the open circles are for analyses of all vertebrates while the closed circles are for analyses confined to mammals only). Each point on the plots in Figure 1 corresponds to the frequency and ω ratios for the best minimum model that provides a statistically superior fit to the data over the closest simpler model (see legend to Figure 1 for more details).
Figure 1 Summary of PAML discrete ω ratio variation models. Each point on the plots in (A), (B), and (C) corresponds to the frequency and ω ratios for the best minimum model (e.g., M0, M3/ncatG = 2, M3/ncatG = 3, etc.) that provides a statistically superior fit to the data (i.e., a more complex model with additional codon ω ratio classes that does not provide a statistically better fit to the next simplest model is rejected). For example, the analysis of the full-length sequence of NR1A1 (TRα) for all available vertebrate species shows that M3/ncatG = 3 is superior to M3/ncatG = 2 but statistically equivalent to M3/ncatG = 4. Consequently, plotted on Figure 1 are three points for the NR1A1 M3/ncatG = 3 analysis corresponding to frequency and ω ratios for three classes of codons – 80.6% (frequency = 0.806) of codons have an estimated ω ratio of 0.004, 14.9% have an ω ratio of 0.094, and 4.5% have an ω ratio of 0.259. An analysis that shows M0 is the best minimum model will have 100% of codons (frequency = 1.0) with a particular ω ratio. (A), (B), and (C) apply to analyses of full-length sequences, DBD only, and LBD only, respectively. The open circles are for analyses of all available vertebrate sequences while the closed circles are for analyses of mammals only. For part (C), the red open and closed triangles represent data for the LBD of the AHR gene (a non-NR gene that encodes a protein with similar function to PXR and CAR).
Overall, the M0/M3 analyses confirm original observations from two-species comparisons that the NR genes are in general subject to strong negative selection, particularly in the DBD [8,29]. The DBDs show lower ω ratios than the LBD or full-length sequences (Figure 1B). Only NR1C1 (peroxisome proliferator-activated receptor-α; PPARα) has a sub-population of codons within the DBD with an estimated ω ratio of greater than 0.5, and this sub-population corresponds to only 1 codon out of 66 analyzed. A number of genes, including NR1B1 (retinoic acid receptor-α; RARα), NR2B1 (retinoid X receptor-α; RXRα), NR3B1 (ERRα), and NR3B2 (ERRβ), show virtually no non-synonymous nucleotide differences between different species in the DBD and have ω ratios close to 0 (< 0.01). NR1A1 (TRα) and NR1I3 (CAR) are somewhat unusual relative to the other NR genes in having at least 15% of codons with an estimated ω ratio of greater than 0.1.
The LBDs of the NR genes clearly show higher ω ratios than the DBDs (Figure 1C). Yet, despite this, the majority of receptors (41 of 48) have ω ratios for all sub-populations of codons less than 0.5. Only PXR genes, analyzed for mammals only, have a sub-populations of codons with an ω ratio greater than 1; this ω class corresponds to 5% of all analyzed codons in the PXR LBD. The analyses for the full-length receptor sequences generally follow the trends seen for the LBD only analysis with minor differences (compare Figure 1A and 1C). To provide another comparison to PXR and CAR, PAML analysis was applied to the LBD of the aryl hydrocarbon receptor-1 (AHR), a non-NR that has a similar function to CAR and PXR, namely to respond to ligands (including xenobiotics) and regulate expression of genes involved in metabolism and elimination of potentially toxic compounds [34]. In contrast to PXR and CAR, the ω ratios associated with the AHR LBD were more similar to the majority of NR genes than to PXR or CAR (see red open and closed triangles in Figure 1C).
M7 and M8 analyses
For most analyses, M8 was not statistically superior to the neutral model M7. Only 10 of 132 analyses of all vertebrate species and 3 of 65 analyses for mammals-only revealed M8 results statistically superior to M7. None of the M8 analyses identified a sub-population of codons with an ω ratio exceeding 1. Once again, however, analysis did identify a sub-population of codons within the PXR LBD with a high ω ratio relative to other NRs (e.g., for analysis of the full-length PXR receptors for all vertebrates, the M8 analysis found 3.4% of the PXR codons, all within the LBD, with an ω ratio of 0.97).
Figure 2 shows plots derived from the estimated β-distribution parameters for the M7 (or M8, if statistically superior to M7) analyses for the NR genes. The β-distribution is a continuous function that for the PAML analysis is restricted to values between 0 and 1. Depending on the parameters α and β for the β-distributions, the values may be clustered more towards 0 or be more evenly distributed between 0 and 1. Figure 2 plots for a particular gene how many codons have estimated ω ratios equal to or less than a particular ω ratio on the abscissa. For example, for analysis of the LBD of NR3B1 (ERRα), the M7 analysis produces a β-distribution where all variation in ω ratios across codons is accounted for by ω ratios < 0.01 (see Figure 2C). In contrast, PAML analyses for the DBD and LBD of CAR show β-distributions that span a range of ω ratios greater than all other NR genes; the same analyses for PXR are second only to CAR in the range of ω ratios spanned (Figure 2B,C). The two LBD domain 'singletons', DAX-1 and SHP also show a wider span of ω ratios than most other NR genes (Figure 2A,C; DBD analysis for DAX-1 and SHP is not possible as these are domain singletons). Similar to the M0/M3 analyses described above, PAML M7/M8 analysis was also applied to the LBD of the AHR LBD. In contrast to PXR and CAR, the β-distribution of ω ratios across codons in the AHR LBD was more similar to the majority of NR genes than to PXR or CAR (see red line in Figure 2C).
Figure 2 Summary of PAML β-distribution ω ratio variation models. The plots are derived from the estimated β-distribution parameters for the M7 (or M8, if statistically superior to M7) models for the NR genes and show for a particular gene how many codons have estimated ω ratios equal to or less than a particular ω ratio on the abscissa. In contrast to Figure 1, only data derived from analyses of all available species are included in Figure 2 (i.e., mammals-only comparisons are not included). (A), (B), and (C) apply to analyses of full-length sequences, DBD only, and LBD only, respectively. For part (C), the red curve represents data for the LBD of the AHR gene. Analysis in part (A) is for NR1A1, 1B1, 1C1, 1F2, 1H3, 1I1, 1I2, 1I3, 2A1, 2B1, 3A1, 3A2, 3B1, 3C1, 3C3, 4A1, 5A1, 6A1, 0B1, and OB2; for part (B), analysis is for NR1A1, 1B1, 1C1, 1H3, 1I1, 1I2, 1I3, 2A1, 2B1, 3A1, 3A2, 3B1, 3C1, 3C3, 4A1, 5A1, and 6A1; and for part (C), analysis is for NR1A1, 1B1, 1F2, 1H3, 1I1, 1I2, 1I3, 2A1, 2B1, 3A1, 3A2, 3B1, 3C1, 3C3, 4A1, 5A1, 6A1, 0B1, OB2, and AHR.
Individual variation among codons in selected NR genes
The M3 PAML analyses also provide estimates of the mean ω ratio for each individual codon in a gene or gene domain. This provides some indication of which specific codons are potentially subject to positive selection. Figure 3 shows a plot of calculated mean ω ratios for the LBDs of 7 NR genes and the AHR gene. Comparison across the LBDs of NR1I subfamily genes shows that PXR has very diverse variation of ω ratios across codons with 5.1% of codons in the mammals only analysis having estimated mean ω ratios exceeding 1 (Figure 3C). In contrast, VDR, whose major function is to respond to 1,25-(OH)2-vitamin D3 (calcitriol), a ligand conserved across all vertebrate species, shows much lower inter-codon variation of ω ratios than CAR or PXR (compare Figures 3A,C,E). The pattern of ω variation across codons for CAR and the AHR gene are descriptively similar (Figure 3E,H). For the mammals-only analysis, higher mean ω ratios are clearly associated with the Helix1-3 (H1-3) 'insert' region of PXR. The H1-3 insert is very divergent across PXR genes and, similar to the CAR genes, the two Xenopus laevis BXRs lack this sequence entirely. This stretch of sequence was excluded from the analyses of VDR and PXR for all vertebrate species due to extreme sequence divergence and difficulties in alignment.
Figure 3 Estimates of ω ratios for individual codons in the LBDs of 7 NR genes and the AHR gene. The graphs in (A) through (H) plot the estimated ω ratios for individual codons of the LBDs of 7 NR genes and the AHR gene derived from the 'best minimum' PAML discrete model. The location of the α-helices in the LBDs of the NR genes that correspond to codons are indicated in the abscissas (e.g., 'H1' denotes α-helix-1; 'H1-H3 insert' denotes the insertion region in the NR1I subfamily proteins between helix-1 and helix-3); the location of the PAS-B domain is also shown for the AHR gene. CAR lacks the H1-H3 insert but this region is plotted in (E) to keep the alignment consistent between (A) VDR, (C) PXR, and (E) CAR. Due to difficulties in alignment and extreme sequence divergence for VDR and PXR in the H1-H3 insertion region, PAML analysis for this region could be performed for mammals only for the PXR genes. For NR1I1, NR2B2, and NR3C4, analysis restricted to mammals resulted in a best minimum PAML discrete model of only one ω ratio population (i.e., the M0 model); therefore, only data for all vertebrate species is plotted for those three genes (note also that the CAR gene is only found in mammals). The plots in (A), (C), and (E) show data for all three NR1I subfamily members and reveal that PXR has the widest variation of ω ratios across codons both within this subfamily (with CAR intermediate between PXR and VDR) and compared to the other NR genes.
The analyses of other NR genes also show some variation of nucleotide diversity across different receptors. RXRβ (NR2B2) is illustrative of a group of NRs whose LBDs show very low ω ratios across all codons (Figure 3B). There are also differences between the LBDs of the sex and adrenocortical steroid receptors with the androgen receptor (AR, NR3C4) showing low ω ratios, the estrogen receptor-β (ERβ, NR3A2) somewhat higher, and the glucocorticoid receptor (GR, NR3C1) the highest ω ratios of the three (Figure 3D,F,G). The somewhat higher ω ratios for select codons in the GR may relate to this receptor likely being one of the evolutionarily 'newer' classical steroid receptors, the estrogen and progesterone receptors being the most ancestral [17-20].
The receptors described above are all ligand-dependent. A contrasting group of receptors are the 'ligand-independent' NRs, of which the estrogen-related receptors (ERRα, NR3B1; ERRβ, NR3B2; ERRγ, NR3B3) [35], steroidogenic factor 1 (SF-1; NR5A1) [36], and liver receptor homolog 1 (LRH-1, NR5A2) [36] are examples. These NRs are activated in the absence of ligand, although recent work has shown that phosphatidyl inositols are likely endogenous ligands for SF-1 and LRH-1 [36]. The evolutionary history of ligand-independent NRs is incompletely understood [11,20], with a proposal that ligand-binding is actually the ancestral state for SF-1 and LRH-1 [36]. The ω ratios for ERRα, ERRβ, ERRγ, SF-1, and LRH-1 were all on the lower end for the NR superfamily, both across all vertebrates and within mammals (Additional file 3: Results of PAML analysis and treefiles), with patterns of ω ratio variation across codons very similar to that seen with the androgen receptor (Figure 3G). While it is possible limited positive selection has occurred at a small number of codons in the five ligand-independent receptors discussed above, this was not detected by the PAML analysis.
Figure 4 shows ω ratio variation across the codons of the DBDs of six NR genes. In general, the ω ratios are generally much lower than 1, consistent with strong purifying of the DBD. Overall, CAR and PXR show higher ω ratios across codons than the other four NR genes, but the calculated ω ratios are still less than 0.2 for all codons (Figure 4C,E).
Figure 4 Estimates of ω ratios for individual codons in the DBDs of 6 NR genes. The graphs in (A) through (F) plot the estimated ω ratios for individual codons of the DBDs of 6 NR genes derived from the 'best minimum' PAML discrete model, utilizing sequence data from all available vertebrate species (note that the CAR gene is found only in mammals). In contrast to the analyses of the LBDs in Figure 3, the ω ratio variation in the DBDs for the six NR genes shown in (A) through (F) is limited and restricted to low ω ratios.
Comparisons of the NR1I subfamily genes
The NR1I subfamily includes VDR, PXR, and CAR. Sequence alignments of the LBDs of selected genes from this subfamily are presented in Figure 5. Amino acid residues identified as directly interacting with ligands in high-resolution crystal structures are indicated in bold type. As can be seen from inspection of Figure 5 and highlighting the shared history of PXRs, CARs, and VDRs, a number of orthologous amino acid residues are involved in ligand binding at more than one receptor. For example, leucine-240 of the human PXR directly interacts with the antibiotic rifampicin [37] and hyperforin (active component of the herbal anti-depressant St. John's wort) [38]; the orthologous residue in human VDR (L227) directly interacts with calcitriol analogs [39] while the same position in mouse CAR (L168) binds the pesticide contaminant 1,4-bis [2-(3,5-dichloropyridoxyl)]benzene (TCPOBOP) [40].
Figure 5 Sequence alignment of the LBD of PXR, VDR, and CAR genes. The locations of the α-helices above the amino acid sequences are based on the structures determined from x-ray crystallography of human PXR and human VDR [73, 88]. Amino acid resides highlighted in bold type are residues in human PXR, human VDR, mouse CAR, and human CAR shown to directly interact with structurally diverse ligands. These residues have been determined by x-ray crystallography and, in some cases, by additional molecular modelling for human VDR [39, 88-90], rat VDR [91], human PXR [37, 38, 73], mouse CAR [40, 92], and human CAR [81]. The ligands for the various receptors are: human VDR – calcitriol [39, 88, 89], 20-epi calcitriol analogs [89], calcipotriol, seocalcitol [39], 1α,25-lumisterol [90]; rat VDR – 2-carbon substituted vitamin D3 analogs [91]; human PXR – SR12813 [73], hyperforin [38], rifampicin [37]; mouse CAR – 5α-androst-16-en-3α-ol (androstenol) [92], TCPOBOP [40]; and human CAR – 5β-pregnan-3,20-dione and 6-(4-chlorophenyl)imidazo[2,1-b][1,3]thiazole-5-carbaldehyde O-(3,4-dichlorobenzyl)oxime (CITCO) [81]. The amino acid residues highlighted in red underlined boldin Xenopus laevis BXRα and BXRβ correspond to codons that show evidence of positive selection in a previously published phylogenetic analysis of nucleotide variation in the BXRα and/or BXRβ lineages [30]. Note that of the 23 amino acid residue positions identified as having high probability of having experienced positive selection in the BXRα and/or BXRβ lineages, 9 are orthologous to or adjacent to residues that are orthologous to human PXR residues shown to directly interact with the ligands SR12813, hyperforin, and/or rifampicin in x-ray crystallographic structures of the human PXR [37, 38, 73]; an additional two residues are orthologous to ligand-binding residues in human VDR [39, 88-90] and, also in one case, human and mouse CAR as well [40, 81, 92].
In a previous publication, the authors have determined codon positions in the Xenopus laevis BXRs that show evidence for positive selection in the evolution of this unusual lineage of PXRs; 23 such codons were identified, and all were located in the LBD of the BXRs [30]. The BXRs are notable in the PXRs for having lost the ability to be activated (and presumably to bind) structurally diverse ligands. In addition, these receptors have a tissue expression pattern markedly different than other PXRs, being found at high levels in gonadal tissue but not xenobiotic-metabolizing organs such as liver or the intestines [41-43], and are activated efficaciously only by benzoate compounds that have a role in frog development [41,42,44]. Interestingly, 9 of the 23 amino acid residue positions in the BXR lineages that have evidence of positive selection are orthologous to or directly adjacent to residues that are orthologous to human PXR residues shown to directly interact with the ligands SR12813, hyperforin, and/or rifampicin in x-ray crystallographic structures of the human PXR (Figure 5); an additional two residues are orthologous to ligand-binding residues in human VDR and, also in one case, to ligand-binding residues in human and mouse CAR as well. The evidence is consistent with positive selection in the evolution of the BXRs being directed at the LBD to alter ligand specificity, in large part by targeting amino acid residues capable of directly interacting with ligands. This has presumably been a significant factor underlying the much narrower ligand selectivity observed in the modern BXRs.
The NR1I subfamily members also differ markedly in conservation of ligand-binding residues across vertebrate species. Figure 6 shows for VDRs, PXRs, and CARs how many species differ from the human receptor at the amino acid residue positions known to interact directly with ligands. VDRs show tight conservation of ligand-binding residues. Only 4 of 22 amino acid residues show any variation at all across species, ranging from jawless fish (sea lamprey), teleost fish, reptiles, frog, birds, and mammals (Figure 6A). PXRs, on the other hand, show extensive amino acid sequence divergence at ligand-binding residues with only 3 of 23 positions conserved across all 13 vertebrate PXRs currently sequenced; for 9 of 23 positions, over half of the non-human PXRs have an amino acid residue different from the human sequence (Figure 6B). CARs also show more divergence at ligand-binding positions than human VDR but not as great as the PXRs, although the analysis is limited due to the presence of CAR genes only in mammals (Figure 6C).
Figure 6 Conservation of ligand-binding residues in VDR, PXR, and CAR. From high-resolution, x-ray crystallographic structures of human VDR, rat VDR, human PXR, mouse CAR, and human CAR bound to various ligands, the amino acid residues that directly interact with ligands are known (see Figure 5; also see Additional file 1: Genes used for phylogenetic analysis for complete list of species available and their accession numbers). (A) Of the 22 amino acid residues shown to interact with ligands at human and/or rat VDRs, only 4 residues show any sequence variation across vertebrate species. The remaining 18 of 22 residues show complete conservation across all vertebrate VDRs (from sea lamprey to human VDRs). Eighteen VDRs were used for the analysis. Due to partial sequence, data for the chimpanzee VDR was only available for the first two ligand-binding residues (corresponding to human VDR Y143 and F150); in addition, data was missing for the four most C-terminal ligand-binding residues (corresponding to human VDR H397, L414, V418, and F422) for crocodile, snake, turtle, lizard, frog, and fugu-β VDRs. (B) In contrast to the VDRs, the PXRs show extensive amino acid sequence divergence at the residues shown to interact with ligands at the human PXR. Only 3 of 23 positions are conserved throughout the 13 vertebrate PXRs while for 9 of 23 residues, over half of the PXRs have an amino acid residue that differs from that at the human PXR. Also indicated are the 9 amino acid residues in the BXRα and/or BXRβ lineages that show evidence for positive selection (see Figure 5 legend for more details; * indicates BXRα and/or BXRβ residue directly orthologous to human PXR ligand-binding residue; ** indicates residue adjacent to such a ligand-binding residue). (C) CARs also show much more divergence at ligand-binding positions than human VDR but not as great as the PXRs. The data is based on eight complete mammalian CAR sequences.
Functional studies of VDRs and PXRs and a proposed phylogeny of the NR1I subfamily
Previous studies have revealed the broad ligand specificity of mammalian and chicken PXRs; these receptors are activated by a structurally diverse array of xenobiotics and endogenous compounds including bile salts, steroid hormones, and prescription drugs [30,44-50]. A recent study has revealed that while zebrafish PXR does not respond to mammalian 24-carbon (C24) bile acids such as cholic acid, chenodeoxycholic acid, or lithocholic acid, this PXR is activated well by the zebrafish 27-carbon (C27) bile alcohol sulfate (cyprinol sulfate) [30], a biliary bile salt typical of the earliest biliary detergents to evolve in vertebrates [51]. VDRs, on the other hand, have a much more restricted ligand specificity, having adapted to bind calcitriol at nanomolar or subnanomolar affinity [52]. Mammalian VDRs do, however, have the ability to be activated by the toxic secondary bile salt lithocholic acid and its metabolites, a function that confers a protective role in the intestine against this toxic secondary bile acid [53].
Figure 7 examines the response of four PXRs and two VDRs to pregnenolone (a pregnane steroid), scymnol sulfate (a C27 bile alcohol sulfate from cartilaginous fish), petromyzonol sulfate (an unusual C24 bile alcohol sulfate from the sea lamprey), 3-ketolithocholic acid (a metabolite of lithocholic acid), calcitriol, and a benzoate analog (n-propyl-p-hydroxybenzoate). Human PXR is activated by micromolar concentration of all 6 compounds except calcitriol (Figure 7A). As previously proposed [30], activation of human PXR by 'early' bile salts such as petromyzonol sulfate and scymnol sulfate likely represents an 'ancestral' property retained in mammalian PXRs (in contrast, unconjugated scymnol, the precursor to the excreted scymnol sulfate, was inactive at all receptors tested). Similar to human PXR, zebrafish PXR is activated by pregnenolone, scymnol sulfate, and the benzoate analog, but not by the other compounds (Figure 7B). The lack of activation of zebrafish PXR by petromyzonol sulfate may be a result of this compound being an unusual C24 bile alcohol sulfate. The sea lamprey has apparently independently evolved the ability to cleave the cholesterol side-chain because C24 bile alcohol sulfates are not found in other fish, and the peroxisomal mechanisms used to cleave the cholesterol side-chain of C24 bile acids evolved more recently than our last common ancestor with lampreys [51,54]. The Xenopus laevis BXRs do not share steroid or bile salt ligands with human or zebrafish PXR, but are activated well by the benzoate analog (Figure 7C, D).
Figure 7 Concentration-response curves of activation of PXRs and VDRs by endogenous ligands or their analogs. The ordinate represents activation of the PXR or VDR, relative to vehicle control, and normalized to the maximal activator (rifampicin for human PXR, 5α-androst-3α-ol for zebrafish PXR, n-butyl-p-aminobenzoate for Xenopus laevis BXRα, n-propyl-p-hydroxybenzoate for Xenopus laevis BXRβ, and calcitriol for human and sea lamprey VDRs; see Materials and Methods for more details). The drugs tested were pregnenolone (●), petromyzonol sulfate (sea lamprey bile salt; ○), scymnol sulfate (cartilaginous fish bile salt; □), 3-ketolithocholic acid (mammalian bile acid metabolite; △), n-propyl-p-hydroxybenzoate (▲), and 1,25-(OH)2-vitamin D3 (calcitriol; ■). (A) Human PXR is activated by micromolar concentrations of the steroid pregenonolone, the early bile salts petromyzonol sulfate and scymnol sulfate, 3-ketolithocholic acid, and the benzoate analog. Calcitriol does not activate human PXR. (B) Similar to human PXR, zebrafish PXR is activated by the steroid pregnenolone, the cartilaginous fish bile salt scymnol sulfate, and the benzoate analog, but not by the other compounds. (C, D) The Xenopus laevis BXRs do not share any ligands with human and zebrafish PXRS other than the benzoate analog n-propyl-p-hydroxybenzoate, which activates BXRα and BXRβ robustly. (E, F) Human and sea lamprey VDRs are both activated robustly by nanomolar concentrations of calcitriol. Similar to human PXR, 3-ketolithocholic acid activates human VDR at micromolar concentrations, with an efficacy of only 15% relative to calcitriol. Sea lamprey was not activated at all by 3-ketolithocholic acid. Weak concentration-dependent activation of sea lamprey VDR by petromyzonol sulfate (○) was observed; however, the efficacy of this bile salt was only ~5–6% relative to that of calcitriol. In panels (A, E, F), full-length receptors for human PXR, human VDR, and sea lamprey VDR were used, with the reporter plasmid being CYP3A4-PXRE-Luc. In panels (B, C, D), GAL4-LBD fusion constructs were used for zebrafish PXR and the Xenopus laevis BXRs, with the reporter plasmid being tk-UAS-Luc.
Despite over 500 million years since the last common ancestor of humans and jawless fish, the human and sea lamprey VDRs are activated very similarly by calcitriol (Figure 7E, F). The major difference in ligands between these two receptors is the activation of human VDR by lithocholic acid and its metabolites (Figure 7E)[30,53]. Of the six compounds tested, only calcitriol activated the sea lamprey VDR efficaciously. The sea lamprey biliary surfactant petromyzonol sulfate produced a weak but concentration-dependent activation that had a maximal effect only 5–6% that of calcitriol (Figure 7F). The possible physiologic significance of this weak in vitro effect needs to be correlated with in vivo experiments in sea lampreys.
Overall, VDRs are activated by nanomolar concentrations of calcitriol, a ligand that does not activate PXRs. In contrast, zebrafish PXR and human PXR share similar activation by pregnane and androstane steroids along with early bile salts, as indicated by other studies [30,44]. Recent work has also revealed that activation of the zebrafish PXR in vivo upregulates the expression of cytochrome P450 (CYP) 3A and multi drug resistance 1 (MDR1) genes [55], properties shared by other PXRs (except in the frog) [46,49,56,57].
Discussion
The search for genes that show evidence for positive Darwinian selection has been an important focus of molecular phylogenetics [21,22]. The genes in the NR superfamily generally show nucleotide variation across species consistent with strong purifying selection, particularly in the DBDs. This study applied phylogenetic analysis by a maximum likelihood method to the entire NR superfamily in vertebrates to analyze patterns of nucleotide variation that may be suggestive of positive selection. The results extend previous more limited phylogenetic analysis of PXR and CAR genes [8,29,30,58] and clearly show that the LBDs of the PXR and CAR genes have ω ratios at the extreme high end for the NR superfamily. Two other genes that have similar variation in ω ratios are DAX-1 and SHP, two NRs that lack a defined DBD and are classified as LBD singletons [12,13]. The elevated ω ratios in DAX-1 and SHP, relative to other NR genes, may be related to the unique evolutionary pressures these two genes face as domain singletons [8,11].
This possible signature of positive selection in the LBDs of the PXR and CAR genes is consistent with the role of PXR and CAR as sensors of toxic endogenous and exogenous compounds that may vary across species [5,6]. For PXR and CAR, evolutionary selection would be directed at fine-tuning ligand specificity towards the most important toxic compounds (or class of compounds) for a given species. PXR and CAR clearly differ markedly in nucleotide variation from VDR, the other member of the NR1I subfamily; in terms of nucleotide variation, VDR behaves like other NRs and not PXR or CAR. The elevated ω ratios of PXR and CAR are not simply a result of synonymous and non-synonymous substitutions being increased in tandem. As with two gene comparisons within mammals [8], synonymous substitutions rates for PXR and CAR genes across vertebrates are average when compared to other NR genes [30,58]. The estimated ω ratios for the codons of the LBDs of CAR and PXR genes are also higher than those for the AHR gene, a non-NR that qualitatively has very similar functions to CAR and PXR, including the ability to upregulate the expression of CYP genes involved in detoxification of xenobiotics [34].
An additional finding with CAR was that the ω ratios for the CAR DBD are on the higher end for the NR superfamily. This may be a reflection of the recent divergence of the CAR gene following duplication of a pre-mammalian PXR gene [59]. Functional diversification with higher rates of non-synonymous to synonymous substitutions is common in a gene that has recently duplicated [60,61]. Once two copies of a gene exist, one of the two genes is free to diversify function to evolutionary advantage although the two genes still may share overlapping functions. The CAR and PXR DBDs show considerable cross-reactivity with one another in terms of interactions with target gene response elements [62-67], extending even to the chicken PXR [68]. This cross-reactivity highlights the close evolutionary history of PXR and CAR and may indicate that CAR provides physiologically important redundancy to important functions of PXR in mammals. (Note that even PXR and VDR, which are more distantly related than CAR and PXR, show considerable overlap in target gene regulation. VDR is able to upregulate expression of detoxifying proteins such as CYP3A4 [53,69,70] while PXR has recently been shown to regulate expression of proteins involved in the metabolism of calcitriol and related molecules [71,72]). The relatively high ratio of non-synonymous to synonymous substitutions of the CAR DBD (at least as compared to other NRs) suggests, however, that the CAR DBD is not evolutionarily 'static' and is diversifying to recognize other binding elements or to interact differently than PXR with shared binding elements.
The phylogenetic analyses presented here do have some limitations. The number of sequences available for analysis varies markedly across the NR superfamily. For those genes that have few sequences available, the phylogenetic analysis will have more limited power to detect small sub-populations of codons with elevated ω ratios [22,32,33]. On the other hand, analysis of too few genes runs the risk of increasing false positives [22,31-33]; this was the reason a minimum of number of six genes from six different genera was applied to the NR superfamily dataset. In addition, difficulties in sequence alignment are a potential problem for some genes and necessitated the removal of some codons from phylogenetic analysis. For the NR1I subfamily, this was particularly an issue with the H1-H3 insert. This sequence was not analyzed for the PXR and VDR genes when applied to all vertebrates. Interestingly, the H1-H3 insert region of PXR genes within mammals was identified as a region where a number of the codons have estimated ω ratios greater than one. Given the role of this region in expanding the ligand-binding pocket of human PXR relative to other NRs [37,38,73], variation in the H1-H3 insert region within mammalian PXRs may have allowed for evolutionarily advantageous changes in ligand specificity across species.
Figure 8 shows a proposed phylogeny of the NR1I subfamily, taking into account data from this study and others. The evolutionary history of PXR and CAR has not been fully resolved. Experimental analysis and ongoing genome sequencing projects have not revealed a CAR gene in teleost fish, with zebrafish, Fugu rubripes, and Tetraodon nigroviridis each possessing a single gene classified as a PXR due to closer sequence and functional similarity to mammalian PXR genes than CAR genes [9,44]. Similarly, the chicken possesses only a single PXR/CAR-like gene (the 'chicken X receptor' or CXR) with sequencing of the chicken genome, extensive attempts to clone an additional NR1I receptor gene, and RNA interference of the CXR gene coupled with functional assays failing to show direct or indirect evidence of an additional NR1I subfamily gene member [59]. Unlike telost fish PXRs, the CXR has equal sequence similarity to mammalian PXRs and CARs and shares a number of properties with mammalian CARs, including high constitutive activity in in vitro assays and lack of sequence in the H1-H3 insert region, leading to the possibility that the CXR gene is actually a CAR gene and not a PXR gene as currently classified [44,48,59]. The phylogeny in Figure 8 classifies the CXR as a PXR and presents the parsimonious explanation that fish, birds, and their ancestors possessed a single PXR gene that duplicated in an ancestral mammal or pre-mammal. One of the two resulting genes then diverged to the CAR gene [59]. This hypothetical phylogeny will be strengthened or modified by analysis of PXR/CAR genes in reptiles and additional mammals such as marsupials or monotremes.
Figure 8 Proposed phylogeny of the NR1I subfamily. The phylogenetic tree is based on known phylogenetic relationships between the species combined with functional data from this study and others. Features included are activation by pregnane and androstane steroids, C27 bile alcohol sulfates (like cyprinol and scymnol sulfate), C24 bile acids (such as cholic and lithocholic acid), benzoates, and calcitriol; ability to increase (upregulate) the expression of the CYP3A enzymes; and high constitutive (baseline) activity. It is possible that activation of PXRs by benzoates is an ancestral property as all PXRs can be activated by at least some benzoate compounds [44], although functional roles of these compounds have so far only been demonstrated in frogs [41, 42]. The study of ligand effects on CARs is complicated by the high constitutive activity of these receptors; many ligands act as inverse agonists of CARs. The possible developmental role of zebrafish PXR is highlighted by its strong expression in early life stages of zebrafish [74].
As described above, the Xenopus laevis BXRs are an especially intriguing example of extensive divergence of a gene from other vertebrate orthologs. The BXRs have dramatically altered ligand specificity and tissue expression relative to other PXRs. The BXR lineage shows strong evidence of positive selection [30], directed particularly at likely ligand-binding residues, based on comparison to known ligand-binding residues of human PXR, human VDR, mouse CAR, and human CAR.
An unanswered question in the NR1I subfamily is the functional properties of the closest common ancestor to PXR and VDR. Clues to this may be revealed if PXR can be cloned and characterized from a jawless fish such as lampreys or hagfish or with exploration of invertebrate ortholog(s) to these receptors. Sequencing of the genome of the chordate Amphioxus, one of the closest invertebrate relatives to vertebrates, and sea lamprey may be insightful in this regard. There are suggestions that zebrafish PXR may have important roles in early zebrafish development [74], which, if true, would provide a functional link between the zebrafish PXR and the frog BXRs [41,43].
The presence of a high-affinity VDR in the sea lamprey, a jawless fish with a non-calcified cartilaginous skeleton, suggests that regulation of calcium and phosphate levels was not a function (or a critical function) of the 'ancestral' VDR [75]. The high concentrations of calcium in sea water mean that maintaining adequate calcium levels in tissues and body fluids is mainly a problem for terrestrial or fresh water animals. In mammals, VDRs mediate a number of functions other than regulation of calcium and phosphate levels, with critical roles in the immune system and skin development [76,77]. Ciona intestinalis (sea squirt), a urochordate that is the closest invertebrate relative to vertebrates for which relatively complete genome information is available, has a single NR equally related to VDR/PXR/CAR and a Drosophila melanogaster NR [78]. The properties of this invertebrate gene remain uncharacterized. These findings highlight how little is known of the physiology of early fish and chordate invertebrates and of the role of NRs in these animals.
The marked diversity of the PXR and CAR LBDs across vertebrates contrasts markedly with detailed resequencing studies of the human PXR and CAR genes that reveal that non-synonymous mutations in the LBDs of these genes are rare. Resequencing of 70 individuals from three different ethnic groups for the CAR gene [58] and approximately 100 individuals from several different ethnic groups for the PXR gene [79] showed very low nucleotide diversity and no nonsynonymous substitutions in the LBD of either gene. Sequencing of 253 Japanese subjects revealed only a single non-synonymous substitution in the CAR LBD [80]; the residue identified (valine-133) is adjacent to a ligand-binding residue of human CAR [81] and mouse CAR [40]. Sequencing of 205 Japanese subjects found two non-synonymous PXR substitutions (R381W and I403V) as single alleles in separate individuals [82]; these mutations caused modest reductions in transactivation of a CYP3A4-based reporter [83]. Another PXR re-sequencing study found a single D163G substitution in 1/74 Africans and 0/418 Caucasians and a single A370T substitution in 1/64 Africans and 0/312 Caucasians [84]. In addition, the nucleotide divergence of the human and chimpanzee CAR and PXR genes are lower than the average for other genes in the human genome [29,58,85]. This suggests that important functions of the LBDs of PXR and CAR, including ligand specificity, do not vary significantly across human populations, and perhaps not between chimpanzees and humans as well, but do vary between humans and other mammals. Future studies should identify the ligands that have shaped the variation of PXR and CAR across species.
Conclusion
NR genes generally show strong sequence conservation and little evidence for positive selection. The main exceptions are PXR and CAR, genes that may have adapted to cross-species differences in toxic compound exposure. Future studies will be directed at precisely defining the cross-species structural variation in CARs and PXRs and relating this to evolutionarily relevant differences in toxic compound exposure.
Methods
Phylogenetic analysis
Sequences for NR genes were downloaded from public databases National Center for Biotechnology Information (NCBI; ) and Ensembl . Complete listing of all genes, species, and accession numbers are provided in Additional file 1: Genes used for phylogenetic analysis. Fragmentary sequences missing 10 or more codons were excluded from the analysis. In situations where supporting functional data was not available, orthology was confirmed by reciprocal BLAST searches. Sequences were aligned with Clustal X. Regions of sequences that could not be aligned between species were excluded from analyses. This was primarily an issue when attempting to align certain non-mammalian NRs with mammalian NRs, a difficult problem especially for the PXRs [30,44]. Estimation of dN/dS (ω) ratios was carried out by maximum likelihood using a codon-based substitution model in PAML (Phylogenetic Analysis by Maximum Likelihood) version 3.13 [25,26,28]. The input to PAML is a treefile of the phylogeny of the sequences to be studied and a file with aligned sequences (see Additional file 2: Sequences used for phylogenetic analysis by PAML and Additional file 3: Results of PAML analysis and treefiles). The phylogeny is based on known phylogenetic relationships between the species to be studied, determined by a consensus of morphological and molecular data [86]. The treefiles for all analyses are in Additional file 3: Results of PAML analysis and treefiles.
PAML determines estimates of ω ratios for models of varying complexity. The most commonly applied models are as follows (the PAML model numbers are shown in parestheses; the 'sites' refers to codons) [25,26]: model M0 (null model with a single ω ratio among all sites), M3 ("discrete" model, with 2 or more categories of sites with the ω ratio free to vary for each site at any value from 0 to greater than 1). M7 ("β model", ten categories of sites, with ten ω ratios in the range 0–1 taken from a discrete approximate of the β distribution), and M8 ("β plus ω " model, ten categories of sites from a β distribution as in M7 plus an additional category of sites with an ω ratio that is free to vary from 0 to greater than 1). PAML estimates the ω ratios that are allowed to vary under these models, as well as the proportion of sites (codons) with each ratio.
Of the PAML models listed above, M0, M3 and M8 can detect positive selection (i.e., ω > 1), although it would be unlikely that M0 would show ω > 1 for any NR gene given the rarity of such high ω ratios across all codons of an entire vertebrate gene or gene domain [24]. Each PAML model generates a log-likelihood, indicating how well the model fits the input data. Some PAML models are "nested" within each other (e.g., M0 within M3, M7 within M8). In those cases, twice the log-likelihood difference between the two models is compared with a X2 distribution with degrees of freedom equal to the difference in degrees of freedom between the two models [M0 has 1 degrees of freedom; M3 with two categories of ω sites (defined in the PAML software as ncatG = 2) has 3 degrees of freedom; M3, ncatG = 3, has 5 degrees of freedom; M3, ncatG = 4, has 7 degrees of freedom; M7 has 2 degrees of freedom; M8 has 4 degrees of freedom] [25,26]. P values for sites potentially under positive selection are obtained using a Bayesian approach in PAML [87]. The accuracy and power of PAML models increases with more sequences and longer length sequences [32,33]. Analyses were only performed if at least six species from six separate genera were available. Below six taxa, the power of PAML to detect positive selection is limited, and the risk of false positives increases [22,31-33]. Simpler PAML models are preferred unless a more complex model fits the data significantly better. Data from a more complex PAML model is presented only if twice the log-likelihood difference between that model and the closest simpler model (e.g., M0 compared with M3, ncatG = 2; M7 with M8) differs significantly with a P < 0.05 according to a X2 distribution.
Functional assays of PXR and VDR
Ligand activation of PXRs and VDRs was determined by a luciferase-based functional assay using methods previously described [30]. Briefly, HepG2 (human liver) cells stably expressing human Na+-taurocholate cotransporter (NTCP; SLC10A1) were used [30]. For experiments involving sulfated bile salts, human OATP (OATP; SLC21) was co-transfected at 10 ng/well to facilitate bile salt uptake. Mouse VDR (IMAGE clone 3710866) and pCMV-sport6 vectors were obtained from Invitrogen (Carlsbad, CA, USA). The zebrafish RXRβ cDNA clone (IMAGE clone 5410111) was obtained from ATCC (Manassus, VA, USA). The expression vectors were either full-length receptors (i.e., containing both a DBD and LBD; human PXR, human VDR, mouse VDR, sea lamprey VDR) or GAL4/PXR chimeras that contain only the LBD of the PXR receptor (BXRα, BXRβ, and zebrafish PXR). For the full-length expression vectors, the reporter plasmid was CYP3A4-PXRE-Luc, a construct that contains a promoter element from CYP3A4 (recognized by PXR and VDR DBDs) driving luciferase expression. For the GAL4/LBD expression constructs, the reporter plasmid was tk-UAS-Luc, which contains GAL4 DNA binding elements driving luciferase expression. The sea lamprey VDR cDNA was co-transfected with zebrafish RXRβ (15 ng/well) for more robust expression [30,75].
It should be pointed out that cross-species differences in the DBDs of various PXRs could impact the ability of a particular PXR to activate the human CYP3A4-based promoter driving luciferase expression. However, this is unlikely to affect the pharmacology of the various ligands studied in this report, particularly as ligand activation of a particular receptor was normalized to a specific maximal activator. The most distantly related PXRs to the human PXR, zebrafish PXR and frog BXRα and BXRβ, were studied using GAL4-LBD fusion constructs, so issues of cross-species differences in the DBD do not affect those receptors in this study. Although the sea lamprey is evolutionarily distant from mammals, the full-length sea lamprey VDR robustly activates the CYP3A4-PXRE-Luc reporter in response to calcitriol.
Activation of receptor by ligand was compared to receptor exposed to identical conditions without ligand ('vehicle control'). In general, dimethyl sulfoxide (Sigma) was used as vehicle and was adjusted to be 1% (v/v) in all wells. A control was also run with transfection of 'empty' vector (i.e., lacking the receptor cDNA) and reporter vector to control for activation of reporter vector by endogenous receptor(s). In experiments with a variety of activators, activation by endogenous receptors was not seen. Experiments were performed in quadruplicate and repeated for a total of at least three times. Concentration-response curves were fitted using Kaleidagraph software (Synergy Software, Reading, PA, USA). Data are presented throughout as mean ± S.E.M. In combining data from multiple experiments, the pooled variance was calculated by the formula spooled = {[(n1-1)s12 + (n2-1)s22 + ... + (nk-1)sk2]/[N-k]}}-1/2, where there are N total data points among k groups, with n replicates in the ith group.
Each PXR or VDR construct was tested with compounds previously shown to be robust activators of the respective receptors [30]. To facilitate more reliable cross-species comparisons, complete concentration-response curves for ligands were determined in the same microplate as determination of response to a maximal activator. This allows for determination of relative efficacy, ε, defined as the maximal response to test ligand divided by maximal response to a reference maximal activator (note that ε can exceed 1). Compounds with ε < 1 were considered 'inactive.' The maximal activators and their concentrations for the PXRs and VDRs studied are as follows: human PXR – 10 μM rifampicin; zebrafish PXR – 20 μM 5α-androstan-3α-ol; Xenopus laevis BXRα – 30 μM n-butyl-p-aminobenzoate; Xenopus laevis BXRβ – 50 μM n-propyl-p-hydroxybenzoate; human VDR – 1 μM calcitriol; and sea lamprey VDR – 0.3 μM calcitriol.
Scymnol sulfate was isolated from bile of the Spotted eagle ray (Aetobatus narinari) by extraction and Flash column chromatography. Scymnol sulfate was deconjugated using a solution of 2,2-dimethoxypropane:1.0 N HCl, 7:1 v/v, and incubating 2 hours at 37°C, followed by the addition of water and extraction into ether. Completeness of deconjugation and assessment of purity was performed by thin-layer chromatography using known standards. Other bile salts and steroids were from Steraloids (Newport, RI, USA). All other chemicals were from Sigma (St. Louis, MO, USA).
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
M.D.K. conceived of the study, carried out the majority of experiments, and drafted the manuscript. K.Y. contributed to the experiment, performed some of the cloning and molecular biology, and generated the stable HepG2 cell line used for functional assays. E.G.S. participated in the design of the study and helped edit the manuscript. L.R.H. isolated and purified the cartilaginous fish bile salt and helped edit the manuscript. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Genes used for phylogenetic analysis. Excel file containing accession numbers for all files used in the phylogenetic analysis.
Click here for file
Additional File 2
Sequences used for phylogenetic analysis by PAML. Text file with complete, full-length nucleotide sequences for all genes used in the PAML analysis. The nucleotide numbers corresponding to the DBD and LBD are also indicated.
Click here for file
Additional File 3
Results of PAML analysis and treefiles. Excel file containing summary data for all PAML analyses. The log-likelihood for each analysis is given along with additional relevant summary parameters. The methods for statistically comparing nested models are described in Materials and Methods. In addition, the phylogenetic treefile for each analysis is also included.
Click here for file
Acknowledgements
We thank Dr. Anna Di Rienzo (University of Chicago, Department of Human Genetics) for hosting MDK in her laboratory and for helpful discussions. MDK also gratefully acknowledges supply of plasmids by SA Kliewer, JT Moore, and LB Moore (GlaxoSmithKline, Research Triangle Park, NC) and supply of the sea lamprey VDR clone by GK Whitfield (University of Arizona College of Medicine, Tucson, AZ).
This work was supported by a Robert Priest Fellowship award to M.D.K. EGS was supported by National Institutes of Health Grants GM60346 and P40 CA21765 Cancer Center Support grant, by the National Institutes of Health/National Institute of General Medical Sciences Pharmacogenetics Research Network and Database (U01 GM61374, ) under grant U01 GM61393, and American Lebanese Syrian Associated Charities (ALSAC). LRH was supported by funding from the Zoological Society of San Diego, CA, USA.
==== Refs
Gronemeyer H Laudet V Transcription factors 3: nuclear receptors Protein Profile 1995 2 1173 1308 8681033
Mangelsdorf DJ Thummel C Beato M Herrlich P Schutz G Umesano K Blumberg B Kastner P Mark M Chambon P Evans R The nuclear receptor superfamily: the second decade Cell 1995 83 835 839 8521507 10.1016/0092-8674(95)90199-X
Steinmetz ACU Renaud JP Moras D Binding of ligands and activation of transcription by nuclear receptors Annual Review of Biophysics & Biomolecular Structure 2001 30 329 359 10.1146/annurev.biophys.30.1.329
Moras D Gronemeyer H The nuclear receptor ligand-binding domain: structure and function Current Opinion in Cell Biology 1998 10 384 391 9640540 10.1016/S0955-0674(98)80015-X
Xie W Evans RM Orphan nuclear receptors: the exotics of xenobiotics Journal of Biological Chemistry 2001 276 37739 37742 11459851
Kliewer SA Goodwin B Willson TM The nuclear pregnane X receptor: a key regulator of xenobiotic metabolism Endocrine Reviews 2002 23 687 702 12372848 10.1210/er.2001-0038
Kliewer SA Lehmann JM Willson TM Orphan nuclear receptors: shifting endocrinology into reverse Science 1999 284 757 760 10221899 10.1126/science.284.5415.757
Zhang Z Burch PE Cooney AJ Lanz RB Pereira FA Wu J Gibbs RA Weinstock G Wheeler DA Genomic analysis of the nuclear receptor family: new insights into structure, regulation, and evolution from the rat genome Genome Research 2004 14 580 590 15059999 10.1101/gr.2160004
Maglich JM Caravella JA Lambert MH Willson TM Moore JT Ramamurthy L The first completed genome sequence from a telost fish (Fugu rubripes) adds significant diversity to the nuclear receptor superfamily Nucleic Acids Research 2003 31 4051 4058 12853622 10.1093/nar/gkg444
Nuclear Receptors Nomenclature Comittee A unified nomenclature system for the nuclear receptor superfamily Cell 1999 97 161 163 10219237 10.1016/S0092-8674(00)80726-6
Bertrand S Brunet F Escriva H Parmentier G Laudet V Robinson-Rechavi M Evolutionary genomics of nuclear receptors: from twenty-five ancestral genes to derived endocrine systems Molecular Biology and Evolution 2004 21 1923 1937 15229292 10.1093/molbev/msh200
Seol W Choi HS Moore DD An orphan nuclear hormone receptor that lacks a DNA binding domain and heterodimerizes with other receptors Science 1996 272 1336 1339 8650544
Zanaria E Muscatelli F Bardoni B Strom TM Guioli S Guo W Lalli E Moser C Walker AP McCabe ERB Meitinger T Monaco AP Sassone-Corsi P Camerino G An unusual member of the nuclear hormone receptor superfamily responsible for X-linked adrenal hypoplasia congenita Nature 1994 372 635 641 7990953 10.1038/372635a0
Escriva H Langlois MC Mendonca RL Pierce R Laudet V Evolution and diversification of the nuclear receptor superfamily Annals of the New York Academy of Sciences 1998 839 143 146 9629140
Escriva H Delauny F Laudet V Ligand binding and nuclear receptor evolution BioEssays 2000 22 717 727 10918302 10.1002/1521-1878(200008)22:8<717::AID-BIES5>3.0.CO;2-I
Laudet V Evolution of the nuclear receptor superfamily: early diversification from an ancestral orphan receptor Journal of Molecular Endocrinology 1997 19 207 226 9460643 10.1677/jme.0.0190207
Baker ME Recent insights into the origins of adrenal and sex steroid receptors Journal of Molecular Endocrinology 2002 28 149 152 12063181 10.1677/jme.0.0280149
Baker ME Evolution of adrenal and sex steroid action in vertebrates: a ligand-based mechanism for complexity BioEssays 2003 25 396 400 12655646 10.1002/bies.10252
Thornton JW Evolution of vertebrate steroid receptors from an ancestral estrogen receptor by ligand exploitation and serial genome expansions Proceedings of the National Academy of Sciences of the United States of America 2001 98 5671 5676 11331759 10.1073/pnas.091553298
Thornton JW Need E Crews D Resurrecting the ancestral steroid receptor: ancient origin of estrogen signaling Science 2003 301 1714 1717 14500980 10.1126/science.1086185
Yang Z Bielawski JP Statistical methods for detecting molecular adaptation Trends in Ecology & Evolution 2000 15 496 503 11114436 10.1016/S0169-5347(00)01994-7
Wong WSW Yang Z Goldman N Nielsen R Accuracy and power of statistical methods for detecting adaptive evolution in protein coding sequences and for identifying positively selected sites Genetics 2004 168 1041 1051 15514074 10.1534/genetics.104.031153
Pagani F Raponi M Baralle FE Synonymous mutations in CFTR exon 12 affect splicing and are not neutral in evolution Proceedings of the National Academy of Sciences of the United States of America 2005 102 6368 6372 15840711 10.1073/pnas.0502288102
Endo T Ikeo K Gojobori T Large-scale search for genes on which positive selection may operate Molecular Biology and Evolution 1996 13 685 690 8676743
Yang Z Nielsen R Goldman N Pedersen AK Codon-substitution models for heterogeneous selection pressure at amino acid sites Genetics 2000 155 431 449 10790415
Yang Z Nielsen R Codon-substitution models for detecting molecular adaptation at individual sites along specific lineages Molecular Biology and Evolution 2002 19 908 917 12032247
Huelsenbeck JP Ronquist F Nielsen R Bollback JP Bayesian inference on phylogeny and its impact on evolutionary biology Science 2001 294 2310 2314 11743192 10.1126/science.1065889
Yang Z PAML: a program package for phylogenetic analysis by maximum likelihood Computer Applications in the Biosciences 1997 13 555 556 9367129
Clark AG Glanowski S Nielsen R Thomas PD Kejariwal A Todd MA Tanenbaum DM Civello D Lu F Murphy B Ferriera S Wang G Zheng X White TJ Sninsky JJ Adams MD Cargill M Inferring nonneutral evolution from human-chimp-mouse orthologous gene trios Science 2003 302 1960 1963 14671302 10.1126/science.1088821
Krasowski MD Yasuda K Hagey LR Schuetz EG Evolution of the pregnane X receptor: adaptation to cross-species differences in biliary bile salts Molecular Endocrinology 2005 19 1720 1739 15718292 10.1210/me.2004-0427
Zhang J Frequent false detection of positive selection by the likelihood method with branch-site models Molecular Biology and Evolution 2004 21 1332 1339 15014150 10.1093/molbev/msh117
Anisimova M Bielawski JP Yang Z Accuracy and power of the likelihood ratio test in detecting adaptive molecular evolution Molecular Biology and Evolution 2001 18 1585 1592 11470850
Anisimova M Bielawski JP Yang Z Accuracy and power of Bayes prediction of amino acid sites under positive selection Molecular Biology and Evolution 2002 19 950 958 12032251
Hahn ME Aryl hydrocarbon receptors: diversity and evolution Chemico-Biological Interactions 2002 141 131 160 12213389 10.1016/S0009-2797(02)00070-4
Bardet PL Obrecht-Pflumio S Thisse C Laudet V Thisse B Vanacker JM Cloning and developmental expression of five estrogen-receptor related genes in the zebrafish Development Genes and Evolution 2004 214 240 249 15088146 10.1007/s00427-004-0404-6
Krylova IN Sablin EP Moore J Xu RX Waitt GM MacKay JA Juzumiene D Bynum JM Madauss K Montana V Lebedeva L Suzawa M Williams JD Williams SP Guy RK Thornton JW Fletterick RJ Willson TM Ingraham HA Structural analyses reveal phosphatidyl inositols as ligand for the NR5 orphan receptors SF-1 and LRH-1 Cell 2005 120 343 355 15707893 10.1016/j.cell.2005.01.024
Chrencik JE Orans J Moore LB Xue Y Peng L Collins JL Wisely GB Lambert MH Kliewer SA Redinbo MR Structural disorder in the complex of human PXR and the macrolide antibiotic rifampicin Molecular Endocrinology 2005 19 1125 1134 15705662 10.1210/me.2004-0346
Watkins RE Maglich JM Moore LB Wisely GB Noble SM Davis-Searles PR Lambert MH Kliewer SA Redinbo MR 2.1 Å crystal structure of human PXR in complex with the St. John's wort compound hyperperforin Biochemistry 2003 42 1430 1438 12578355 10.1021/bi0268753
Tocchini-Valentini G Rochel N Wurtz JM Moras D Crystal structure of the vitamin D receptor liganded with the vitamin D side chain analogues calicotriol and seocalcitol, receptor agonists of clinical importance. Insights into a structural basis for the switching of calcipotriol to a receptor antagonist by a further side chain modification Journal of Medicinal Chemistry 2004 47 1956 1961 15055995 10.1021/jm0310582
Suino K Peng L Reynolds R Li Y Cha JY Repa JJ Kliewer SA Xu HE The nuclear xenobiotic receptor CAR: structural determinants of constitutive activation and heterodimerization Molecular Cell 2004 16 893 905 15610733 10.1016/j.molcel.2004.11.036
Blumberg B Kang H Bolado J Chen H Craig AG Moreno TA Umesano K Perlmann T De Robertis EM Evans RM BXR, an embryonic orphan nuclear receptor activated by a novel class of endogenous benzoate metabolites Genes & Development 1998 12 1269 1277 9573044
Grün F Venkatesan RN Tabb MM Zhou C Cao J Hemmati D Blumberg B Benzoate X receptors a and b are pharmacologically distinct and do not function as xenobiotic receptors Journal of Biological Chemistry 2002 277 43691 43697 12198127 10.1074/jbc.M206553200
Heath LA Jones EA Old RW Expression pattern of BXR suggests a role for benzoate ligand-mediated signalling in hatching gland function International Journal of Developmental Biology 2000 44 141 144 10761858
Moore LB Maglich JM McKee DD Wisely B Willson TM Kliewer SA Lambert MH Moore JT Pregnane X receptor (PXR), constitutive androstane receptor (CAR), and benzoate X receptor (BXR) define three pharmacologically distinct classes of nuclear receptors Molecular Endocrinology 2002 16 977 986 11981033 10.1210/me.16.5.977
Kliewer SA Moore JT Wade L Staudinger JL Watson MA Jones SA McKee DD Oliver BB Willson TM Zetterstrom RH Perlmann T Lehmann JM An orphan nuclear receptor activated by pregnanes defines a novel steroid signaling pathway Cell 1998 92 73 82 9489701 10.1016/S0092-8674(00)80900-9
Bertilsson G Heidrich J Svensson K Asman M Jendeberg L Sydow-Backman M Ohlsson R Postlind H Blomquist P Berkenstram A Identification of a human nuclear receptor defines a new signaling pathway for CYP3A induction Proceedings of the National Academy of Sciences of the United States of America 1998 95 12208 12213 9770465 10.1073/pnas.95.21.12208
Blumberg B Sabbagh W Juguilon H Bolado J van Meter CM Ong ES Evans RM SXR, a novel steroid and xenobiotic-sensing nuclear receptor Genes & Development 1998 12 3195 3205 9784494
Handschin C Podvinec M Meyer UA CXR, a chicken xenobiotic-sensing orphan nuclear receptor, is related to both mammalian pregnane X receptor (PXR) and constitutive androstane receptor (CAR) Proceedings of the National Academy of Sciences of the United States of America 2000 97 10769 10774 11005856 10.1073/pnas.97.20.10769
Lehmann JM McKee DD Watson MA Willson TM Moore JT Kliewer SA The human orphan nuclear receptor PXR is activated by compounds that regulate CYP3A4 gene expression and cause drug interactions Journal of Clinical Investigation 1998 102 1016 1023 9727070
Moore LB Parks DJ Jones SA Bledsoe RK Consler TG Stimmel JB Goodwin B Liddle C Blanchard SG Willson TM Collins JL Kliewer SA Orphan nuclear receptors constitutive androstane receptor and pregnane X receptor share xenobiotic and steroid ligands Journal of Biological Chemistry 2000 275 15122 15127 10748001 10.1074/jbc.M001215200
Goto T Holzinger F Hagey LR Cerrè C Ton-Nu HT Schteingart CD Steinbach JH Shneider BL Hofmann AF Physicochemical and physiological properties of 5a-cyprinol sulfate, the toxic bile salt of cyprinid fish Journal of Lipid Research 2003 44 1643 1651 12810826 10.1194/jlr.M300155-JLR200
Baker AR McDonnell DP Hughes M Crisp TM Mangelsdorf DJ Haussler MR Pike JW Shine J O'Malley BW Cloning and expression of full-length cDNA encoding human vitamin D receptor Proceedings of the National Academy of Sciences of the United States of America 1988 85 3294 3298 2835767
Makishima M Lu TT Xie W Whitfield GK Domoto H Evans RM Haussler MR Mangelsdorf DJ Vitamin D receptor as an intestinal bile acid sensor Science 2002 296 1313 1316 12016314 10.1126/science.1070477
Haslewood GAD Tökés L Comparative studies of bile salts: bile salts of the lamprey Petromyzon marinus L. Biochemical Journal 1969 114 179 184 5810077
Bresolin T de Freitas Rebelo M Bainy ACD Expression of PXR, CYP3A, and MDR1 genes in liver of zebrafish Comparative Biochemistry and Physiology Part C 2005 140 403 407 15914091 10.1016/j.cca.2005.04.003
Geick A Eichelbaum M Burk O Nuclear receptor response elements mediate induction of intestinal MDR1 by rifampin Journal of Biological Chemistry 2001 276 14581 14587 11297522 10.1074/jbc.M010173200
Schuetz EG Brimer C Schuetz JD Environmental xenobiotics and the antihormones cyproterone acetate and spironolactone use the nuclear hormone pregnenolone X receptor to activate the CYP3A23 hormone response element Molecular Pharmacology 1998 54 1113 1117 9855641
Thompson EE Kuttab-Boulos H Krasowski MD Di Rienzo A Functional constraints in the constitutive androstane receptor inferred from human sequence variation and cross-species comparisons Human Genomics 2005 2 168 178 16197734
Handschin C Blaettler S Roth A Looser R Oscarson M Kaufmann MR Podvinec M Gnerre C Meyer UA The evolution of drug-activated nuclear receptors: one ancestral gene diverged into two xenosensor genes in mammals Nuclear Receptor 2004 2 7 15479477 10.1186/1478-1336-2-7
Wagner A Birth and death of duplicated genes in completely sequenced eukaryotes Trends in Genetics 2001 17 237 239 11335019 10.1016/S0168-9525(01)02243-0
Lynch M Conery JS The evolutionary fate and consequences of duplicated genes Science 2000 290 1151 1155 11073452 10.1126/science.290.5494.1151
Xie W Barwick JL Simon CM Pierce AM Safe S Blumberg B Guzelian PS Evans RM Reciprocal activation of xenobiotic response genes by nuclear receptors SXR/PXR and CAR Genes & Development 2000 14 3014 3023 11114890 10.1101/gad.846800
Jones SA Moore LB Shenk JL Wisely GB Hamilton GA McKee DD Tomkinson NCO LeCluyse EL Lambert MH Willson TM Kliewer SA Moore JT The pregnane X receptor: a promiscuous xenobiotic receptor that has diverged during evolution Molecular Endocrinology 2000 14 27 39 10628745 10.1210/me.14.1.27
Goodwin B Moore LB Stoltz CM McKee DD Kliewer SA Regulation of the human CYP2B6 gene by the nuclear pregnane X receptor Molecular Pharmacology 2001 60 427 431 11502872
Goodwin B Hodgson E D'Costa DJ Robertson GR Liddle C Transcriptional regulation of the human CYP3A4 gene by the constitutive androstane receptor Molecular Pharmacology 2002 62 359 365 12130689 10.1124/mol.62.2.359
Smirlis D Muangmoonchai R Edwards M Phillips IR Shephard EA Orphan receptor promiscuity in the induction of cytochromes P450 by xenobiotics Journal of Biological Chemistry 2001 276 12822 12826 11278292 10.1074/jbc.M005930200
Mäkinen J Frank C Jyrkkärinne J Gynther J Carlberg C Honkakoski P Modulation of mouse and human phenobarbital-responsive enhancer module by nuclear receptors Molecular Pharmacology 2002 62 366 378 12130690 10.1124/mol.62.2.366
Handschin C Podvinec M Stockli J Hoffmann K Meyer UA Conservation of signaling pathways of xenobiotic-sensing orphan nuclear receptors, chicken xenobiotic receptor, constitutive adrostane receptor, and pregnane X receptor, from birds to humans Molecular Endocrinology 2001 15 1571 1585 11518807 10.1210/me.15.9.1571
Thummel KE Brimer C Yasuda K Thottassery J Senn T Lin Y Ishizuka H Kharasch E Schuetz J Schuetz E Transcriptional control of intestinal cytochrome P-4503A by 1a,25-dihydroxy vitamin D3 Molecular Pharmacology 2001 60 1399 1406 11723248
Drocourt L Ourlin JC Pascussi JM Maurel P Vilarem MJ Expression of CYP3A4, CYP2B6, and CYP2C9 is regulated by the vitamin D receptor pathway in primary human hepatocytes Journal of Biological Chemistry 2002 277 25125 25132 11991950 10.1074/jbc.M201323200
Holick MF Stay tuned to PXR: an orphan actor that may not be D-structive only to bone Journal of Clinical Investigation 2005 115 32 34 15630441 10.1172/JCI200523995
Pascussi JM Robert A Nguyen M Walrant-Debray O Garabedian M Martin P Pineau T Saric J Navarro F Maurel P Vilarem MJ Possible involvement of pregnane X receptor-enhanced CYP24 expression in drug-induced osteomalacia Journal of Clinical Investigation 2005 115 177 186 15630458 10.1172/JCI200521867
Watkins RE Wisely GB Moore LB Collins JL Lambert MH Williams SP Willson TM Kliewer SA Redinbo MR The human nuclear xenobiotic receptor PXR: structural determinants of directed promiscuity Science 2001 292 2329 2333 11408620 10.1126/science.1060762
Bainy ACD Stegeman JJ Cloning and identification of a full length pregnane X receptor and expression in vivo in zebrafish (Danio rerio) Marine Environmental Research 2004 58 133 134
Whitfield GK Dang HTL Schluter SF Bernstein RM Bunag T Manzon LA Hsieh G Dominguez CE Youson JH Haussler MR Marchalonis JJ Cloning of a functional vitamin D receptor from the lamprey (Petromyzon marinus), an ancient vertebrate lacking a calcified skeleton and teeth Endocrinology 2003 144 2704 2716 12746335 10.1210/en.2002-221101
Holick MF Sunlight and vitamin D for bone health and prevention of autoimmune diseases, cancers, and cardiovascular disease American Journal of Nutrition 2004 80 1678S 1688S
Nagpal S Na S Rathnachalam R Noncalcemic actions of vitamin D receptor ligands Endocrine Reviews 2005 26 662 687 15798098 10.1210/er.2004-0002
Dehal P Satou Y Campbell RK Chapman J Degnan B De Tomaso A Davidson B Di Gregorio A Gelpke M Goodstein DM Harafuji N Hastings KEM Ho I Hotta K Huang W Kawashima T Lemaire P Martinez D Meinertzhagen IA Necula S Nonaka M Putnam N Rash S Saiga H Satake M Terry A Yamada L Wang HG Awazu S Azumi K Boore J Branno M Chin-bow S DeSantis R Doyle S Francina P Keys DN Haga S Hayashi H Hino K Imai KS Kano S Kobayashi K Kobayashi M Lee BI Makabe KW Manohar C Matassi G Medina M Mochizuki Y Mount S Morishita T Miura S Nakayama A Nishizaka S Nomoto H Ohta F Oishi K Rigoutsos I Sano M Sasaki A Sasakura Y Shoguchi E Shin-i T Spagnuolo A Stainier D Suzuki MM Tassy O Takatori N Tokuoka M Yagi K Yoshizaki F Wada S Zhang C Hyatt PD Larimer F Detter C Doggett N Galvina T Hawkins T Richardson P Lucas S Kohara Y Levine M Satoh N Rokhsar DS The draft genome of Ciona intestinalis: insights into vertebrate origins Science 2002 298 2157 2167 12481130 10.1126/science.1080049
Zhang J Kuehl P Green ED Touchman JW Watkins PB Daly A Hall SD Maurel P Relling M Brimer C Yasuda K Wrighton SA Hancock M Kim RB Strom S Thummel K Russell CG Hudson JR Scheutz EG Boguski MS The human pregnane X receptor: genomic structure and identification and functional characterization of natural allelic variants Pharmacogenetics 2001 11 555 572 11668216 10.1097/00008571-200110000-00003
Ikeda S Kurose K Ozawa S Sai K Hasegawa R Komamura K Ueno K Kamakura S Kitakaze M Tomoike H Nakajima T Matsumoto K Saito H Goto Y Kimura H Katoh M Sugai K Minami N Shirao K Tamura T Yamamoto N Minami H Ohtsu A Yoshida T Saijo N Saito Y Sawada J Twenty-six novel single nucleotide polymorphisms and their frequencies of the NR1I3 (CAR) gene in a Japanese population Drug Metabolism and Pharmacokinetics 2003 18 413 418 15618763 10.2133/dmpk.18.413
Xu RX Lambert MH Wisely BB Warren EN Weinert EE Waitt GM Williams JD Collins JL Moore LB Willson TM Moore JT A structural basis for constitutive activity in the human CAR/RXRa heterodimer Molecular Cell 2004 16 919 928 15610735 10.1016/j.molcel.2004.11.042
Koyano S Kurose K Ozawa S Saeki M Nakajima Y Hasegawa R Komamura K Ueno K Kamakura S Nakajima T Saito H Kimura H Goto Y Saitoh O Katoh M Ohnuma T Kawai M Sugai K Ohtsuki T Suzuki C Minami N Saito Y Sawada J Eleven novel single nucleotide polymorphisms in the NR1I2 (PXR) gene, four of which induce non-synonymous amino acid alterations Drug Metabolism and Pharmacokinetics 2002 17 561 565 15618712 10.2133/dmpk.17.561
Koyano S Kurose K Saito Y Ozawa S Hasegawa R Komamura K Ueno K Kamakura S Kitakaze M Nakajima T Matsumoto K Akasawa A Saito H Sawada JI Functional characterization of four naturally occurring variants of human pregnane X receptor (PXR): one variant causes dramatic loss of both DNA binding activity and the transactivation of the CYP3A4 promoter/enhancer region Drug Metabolism and Disposition 2004 32 149 154 14709632 10.1124/dmd.32.1.149
Hustert E Zibat A Presecan-Siedel E Eiselt R Mueller R Fub C Brehm I Brinkmann U Eichelbaum M Wojnowski L Burk O Natural protein variants of pregnane X receptor with altered transactivation activity toward CYP3A4 Drug Metabolism and Disposition 2001 29 1454 1459 11602521
Ebersberger I Metzler D Schwarz C Pääbo S Genomewide comparison of DNA sequences between humans and chimpanzees American Journal of Human Genetics 2002 70 1490 1497 11992255 10.1086/340787
Carroll RL Vertebrate paleontology and evolution 1988 New York, Freeman
Yang Z Likelihood ratio tests for detecting positive selection and application to primate lysozyme evolution Molecular Biology and Evolution 1998 15 568 573 9580986
Rochel N Wurtz JM Mitschler A Klaholz B Moras D The crystal structure of the nuclear receptor for vitamin D bound to its natural ligand Molecular Cell 2000 5 173 179 10678179 10.1016/S1097-2765(00)80413-X
Tocchini-Valentini G Rochel N Wurtz JM Mitschler A Moras D Crystal structures of the vitamin D receptor complexed to superagonist 20-epi ligands Proceedings of the National Academy of Sciences of the United States of America 2001 98 5491 5496 11344298 10.1073/pnas.091018698
Mizwicki MT Keidel D Bula CM Bishop JE Zanello LP Wurtz JM Moras D Norman AW Identification of an alternative ligand-binding pocket in the nuclear vitamin D receptor and its functional importance in 1a,25(OH)2-vitamin D3 signaling Proceedings of the National Academy of Sciences of the United States of America 2004 101 12876 12881 15326291 10.1073/pnas.0403606101
Vanhooke JL Benning MM Bauer CB Pike JW DeLuca HF Molecular structure of the rat vitamin D receptor ligand binding domain complexed with 2-carbon-substituted vitamin D3 hormone analogues and a LXXLL-containing coactivator peptide Biochemistry 2004 43 4101 4110 15065852 10.1021/bi036056y
Shan L Vincent J Brunzelle JS Dussault I Lin M Ianculescu I Sherman MA Forman BM Fernandez EJ Structure of the murine constitutive androstane receptor complexed to androstenol: a molecular basis for inverse agonism Molecular Cell 2004 16 907 917 15610734 10.1016/j.molcel.2004.11.037
|
16197547
|
PMC1262763
|
CC BY
|
2021-01-04 16:37:44
|
no
|
Nucl Recept. 2005 Sep 30; 3:2
|
utf-8
|
Nucl Recept
| 2,005 |
10.1186/1478-1336-3-2
|
oa_comm
|
==== Front
Nucl ReceptNuclear Receptor1478-1336BioMed Central London 1478-1336-3-31619755810.1186/1478-1336-3-3ResearchGene expression profiling of potential peroxisome proliferator-activated receptor (PPAR) target genes in human hepatoblastoma cell lines inducibly expressing different PPAR isoforms Tachibana Keisuke [email protected] Yumi [email protected] Toshiya [email protected] Masayuki [email protected] Akira [email protected] Tatsuya [email protected] Chihiro [email protected] Daisuke [email protected] Kenji [email protected] Mikako [email protected] Yasutoshi [email protected] Takahide [email protected] Juro [email protected] Takao [email protected] Tatsuhiko [email protected] Takefumi [email protected] Graduate School of Pharmaceutical Sciences, Osaka University, Osaka, Japan2 Laboratory for System Biology and Medicine, The Research Center for Advanced Science and Technology, the University of Tokyo, Tokyo, Japan3 Perseus Proteomics Inc, Tokyo, Japan4 Graduate School of Medicine, Osaka University, Osaka, Japan2005 3 10 2005 3 3 3 24 5 2005 3 10 2005 Copyright © 2005 Tachibana et al; licensee BioMed Central Ltd.2005Tachibana et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Peroxisome proliferator-activated receptors (PPARs) are ligand-activated transcription factors and commonly play an important role in the regulation of lipid homeostasis. To identify human PPARs-responsive genes, we established tetracycline-regulated human hepatoblastoma cell lines that can be induced to express each human PPAR and investigated the gene expression profiles of these cells.
Results
The expression of each introduced PPAR gene was investigated using the various concentrations of doxycycline in the culture media. We found that the expression of each PPAR subtype was tightly controlled by the concentration of doxycycline in these established cell lines. DNA microarray analyses using these cell lines were performed with or without adding each subtype ligand and provided much important information on the PPAR target genes involved in lipid metabolism, transport, storage and other activities. Interestingly, it was noted that while ligand-activated PPARδ induced target gene expression, unliganded PPARδ repressed these genes. The real-time RT-PCR was used to verify the altered expression of selected genes by PPARs and we found that these genes were induced to express in the same pattern as detected in the microarray analyses. Furthermore, we analysed the 5'-flanking region of the human adipose differentiation-related protein (adrp) gene that responded to all subtypes of PPARs. From the detailed analyses by reporter assays, the EMSAs, and ChIP assays, we determined the functional PPRE of the human adrp gene.
Conclusion
The results suggest that these cell lines are important tools used to identify the human PPARs-responsive genes.
==== Body
Background
The peroxisome proliferator-activated receptors (PPARs) are ligand-activated transcription factors that belong to the nuclear hormone receptor superfamily [1]. Three subtypes, PPARα, PPARβ/δ and PPARγ, have been identified and these subtypes with a high degree of sequence conservation of each subtype across various species, have been characterized. The DNA binding domains of the three subtypes are 80% identical, while their ligand-binding domains exhibit a lower degree (approx. 65%) of identity. Consistent with this relatively high divergence among the subtype-specific ligand binding domains, differential activation of PPARs by endogenous and exogenous compounds may account for the specific biological activity of the three PPAR subtypes [2,3].
PPARα is expressed in the liver, kidney, heart and muscle where it regulates energy homeostasis. PPARα activates fatty acid catabolism, stimulates gluconeogenesis and ketone body synthesis and is involved in the control of lipoprotein assembly [4]. Although PPARα is well characterized, the functional differences of PPARα derived from species are not clear. For example, sustained PPARα activation has carcinogenic consequences in the liver of rodents, but long-term usage of PPARα activators in epidemiological data, has proven that similar effects are unlikely to occur in humans [5,6]. PPARδ is expressed ubiquitously, and is implicated in fatty acid oxidation, in keratinocyte differentiation and wound healing, and in mediating very low density lipoprotein signalling of the macrophage [7-11]. However, the function of PPARδ is less understood than PPARα and γ. There are two PPARγ isoforms, PPARγ1 and γ2. PPARγ2, which is generated by alternative splicing, contains an additional 28 amino acids at the N-terminal end relative to PPARγ1. PPARγ3 is a splicing variant of PPARγ1 and gives rise to the same protein. PPARγ2 is expressed exclusively in adipose tissue and has a pivotal role in adipocyte differentiation, lipid storage in the white adipose tissue and energy dissipation in the brown adipose tissue [12,13]. On the other hand, PPARγ1 is expressed in the liver and other tissues, and the expression of hepatic PPARγ is increased in some obese and diabetic model mice [14-17]. PPARγ is involved in glucose metabolism through the improvement of insulin sensitivity; however, its function is not well defined.
All PPARs bind to a direct repeat of two hexanucleotides, spaced by one or two nucleotides (the DR1 or DR2 motif) as heterodimers with the retinoid X receptor (RXR), and activate several target genes [18-20]. These peroxisome proliferator responsive elements (PPREs) are found in various genes that are involved in lipid metabolism and energy homeostasis, including lipid storage or catabolism, and fatty acid transport, uptake and intracellular binding [21].
One of the approaches for investigating target genes of PPARs is to construct stable cell lines that can be induced to express PPARs. The tet-off system is a well-established system for inducible gene expression. In this system, transcription is turned on or off in response to doxycycline (Dox; a tetracycline derivative) in a strictly dose-dependent manner. Therefore, background or leaky expression in the absence of induction is extremely low. The tet-off system enables us to compare the same cell line before and after induction of the gene of interest. Previously, we established the human hepatoblastoma cell lines (HepG2 cells) which were strictly induced to express the hepatitis C non-structural proteins by removing Dox from the media, and we characterized the changes in mRNA expression profile using DNA microarray analyses [22].
In the present study, to identify human PPARs-responsive genes in the liver cell line, we established tightly tet-regulatable HepG2 cells which can be induced to express each human PPAR. We demonstrated that human PPARs are important regulators of lipid homeostasis in these cell lines using DNA microarray and real-time RT-PCR technologies. Subsequent analyses revealed that all PPARs induced human adipose differentiation-related protein (adrp) gene expression through the same PPRE of the ADRP promoter, while unliganded PPARδ repressed this gene. Our results imply that these cell lines are important tools, which can be used to identify the human PPARs-responsive genes.
Results
Establishment of HepG2 cells that can be induced to express PPARs
To identify human PPARs-responsive genes, we established each cell line that expresses any one of the PPAR subtypes (α, β/δ, γ1 or γ2) using the tet-regulated system. Previously, we established the tightly tet-regulatable HepG2 cell clone (HY-Toff) which was transfected with the pTet-off vector, and that this clone had a large induction/repression rate [22]. The tet-off regulatable HY-Toff cells were transfected with the pBabepuro (for puromycin resistance) and pBI-EGFP vector harbouring the cDNA for human PPARα, PPARδ, PPARγ1 or PPARγ2. We picked out puromycin-resistant clones. Among these clones, we selected the strictly responsive cell lines (HepG2-tet-off-hPPAR cells) to the concentration of Dox. These cell lines were cultured in the presence of Dox at 0, 0.01, 0.1 or 1000 ng/ml for 5 days, and we found that PPARs expression in these cell lines was induced in a dose-dependent manner (Figure 1A–D). We then examined the time course experiment of each PPAR expression after the removal of Dox. The results showed that both the mRNA (Figure 2A–D) and protein (Figure 2E–H) levels of each PPAR were increased after the removal of Dox from the culture medium. Under these conditions, cell proliferation in these cell lines was not affected until 7 days after the removal of Dox. Based on these results, we decided that the expression of each PPAR on the 5th day was suitable for further analyses, such as the determination of the target genes of each PPAR subtype.
Figure 1 Induction of the expression of PPAR by doxycycline in established cell lines. A, B, C and D, Nuclear extracts (50 μg protein/lane) from each cell line cultured in the presence of the indicated amounts of Dox for 5 days were subjected to SDS-PAGE and immunoblots were performed with anti-PPARα (A), anti-PPARδ (B) or anti-PPARγ (C and D).
Figure 2 Time course of PPAR expression in established cell lines. A, B, C and D, The cell lines for the expression of PPARα (A), PPARδ (B), PPARγ1 (C) or PPARγ2 (D) were cultured in the presence of Dox. At time point 0, Dox was removed from the medium. The total RNAs were extracted from the cells cultured for the indicated days after removal of Dox and the amount of mRNAs of PPARs were measured by real time RT-PCR. Values are expressed as the mean ± S.E.M. (n = 3), relative to the control (0 day) set as 1. E, F, G and H, Nuclear extracts (50 μg protein/lane) from the cell lines expressing PPARα (E), PPARδ (F), PPARγ1 (G) or PPARγ2 (H) cultured for the indicated days after removal of Dox were subjected to SDS-PAGE and immunoblots were performed with anti-PPARα (E), anti-PPARδ (F) or anti-PPARγ (G and H).
Changes in mRNA expression profiles by the induction of each PPAR subtype with the PPAR ligands
To characterize the regulation of gene expression by each PPAR subtype, we performed oligonucleotide microarray analyses. HepG2-tet-off-hPPAR cells were cultured in the presence or absence of Dox for 5 days, then treated with the PPAR ligands (100 μM fenofibric acid for PPARα, 100 nM GW501516 for PPARδ or 10 μM ciglitizone for PPARγ) or vehicle for 24 h. Total RNA samples were prepared from these cells, and oligonucleotide microarray analyses were performed using the Affymetrix HG-U133A arrays that contain >22,000 probe sets.
To identify human PPARs-responsive genes, we analysed the genes induced by ligand-activated PPARs. The criteria for selecting genes are described in the Methods section. Based on these criteria, 29, 21, 60 and 107 genes were up-regulated by ligand-activated PPARα, PPARδ, PPARγ1 and PPARγ2, respectively. As shown in Table 1 and the additional file 1, each PPAR affects the expression levels of several genes involved in lipid metabolism, glucose homeostasis, etc. PPARα tends to induce the expression of a number of genes involved in the β-oxidation of fatty acids by mitochondria, peroxisomal fatty acid oxidation, antioxidant and ketogenesis. Most of the genes that were induced by PPARδ were similar to those exhibited by PPARα. We observed that induced expression of PPARδ increased the expression in the presence of the ligand, while unliganded PPARδ repressed the expression of its target genes instead. On the other hand, PPARγ tends to induce the expression of several genes involved in gluconeogenesis, lipid storage, transport and metabolism. Moreover, PPARγ is likely to induce several genes involved in angiogenesis, cytoskeleton organization, protein modification, regulation of transcription, signal transduction, etc.
Table 1 Changes in mRNA expression levels of metabolism-related genes in HepG2-tet-off-hPPAR cells by ligands. Microarray analyses were performed on HepG2-tet-off-hPPAR cells; the cells were cultured in the presence (Dox) or absence of Dox for 5 days. In the absence of Dox, the cells were treated with PPAR ligands (100 μM fenofibric acid for PPARα (Feno), 100 nM GW501516 for PPARδ (GW) or 10 μM ciglitizone for PPARγ (Cig)) or vehicle (DMSO) for 24 h. Gene expression profiles were compared between DMSO and Dox (DMSO versus Dox), ligands and Dox (Feno versus Dox, GW versus Dox, and Cig versus Dox were indicated in the case of PPARα, PPARδ and PPARγ, respectively) or ligands and DMSO (Feno versus DMSO, GW versus DMSO, and Cig versus DMSO were indicated in the case of PPARα, PPARδ and PPARγ, respectively). Samples were analysed using GeneChip® software Microarray Suite (MAS) Ver.5.0 (Affymetrix).
Gene Genbank accession No. PPARα PPARδ PPARγ1 PPARγ2
Fold change Fold change Fold change Fold change
DMSO vs. Dox Feno vs. Dox Feno vs. DMSO DMSO vs. Dox GW vs. Dox GW vs. DMSO DMSO vs. Dox Cig vs. Dox Cig vs. DMSO DMSO vs. Dox Cig vs. Dox Cig vs. DMSO
Fatty acid metabolism
acyl-CoA synthetase long-chain family member 1 NM_001995 1.84 3.49 1.90 0.82 1.72 2.09 0.88 2.30 2.60 0.82 1.00 1.23
carnitine palmitoyltransferase 1A (liver) NM_001876 1.48 2.86 1.93 0.37 2.05 5.60 0.74 1.35 1.83 1.12 1.24 1.10
solute carrier family 25 (carnitine/acylcarnitine translocase), member 20 NM_000387 1.20 2.23 1.85 0.91 1.36 1.50 0.83 0.93 1.12 1.04 1.27 1.22
acyl-Coenzyme A dehydrogenase, very long chain NM_000018 2.16 3.46 1.60 0.92 2.69 2.93 1.36 1.97 1.45 1.55 2.33 1.50
acyl-Coenzyme A dehydrogenase, C-4 to C-12 straight chain NM_000016 1.62 2.52 1.56 0.64 2.03 3.16 1.24 1.48 1.20 1.40 1.80 1.29
hydroxyacyl-Coenzyme A dehydrogenase/3-ketoacyl-Coenzyme A thiolase/enoyl-Coenzyme A hydratase (trifunctional protein), alpha subunit NM_000182 1.48 2.43 1.65 0.89 1.49 1.66 1.34 1.62 1.21 1.41 2.29 1.62
hydroxyacyl-Coenzyme A dehydrogenase/3-ketoacyl-Coenzyme A thiolase/enoyl-Coenzyme A hydratase (trifunctional protein), beta subunit NM_000183 1.74 2.66 1.53 0.95 1.30 1.36 1.17 1.49 1.27 1.14 1.62 1.42
acetyl-Coenzyme A acyltransferase 2 (mitochondrial 3-oxoacyl-Coenzyme A thiolase) NM_006111 1.58 1.92 1.21 1.06 1.63 1.53 0.89 1.31 1.47 1.18 2.44 2.06
acetyl-Coenzyme A acyltransferase 1 (peroxisomal 3-oxoacyl-Coenzyme A thiolase) NM_001607 1.42 1.62 1.14 0.68 1.47 2.15 1.22 2.07 1.70 1.42 2.01 1.41
enoyl Coenzyme A hydratase 1, peroxisomal NM_001398 2.28 2.61 1.15 0.99 2.05 2.07 1.16 1.46 1.26 1.44 1.98 1.38
Antioxidant
catalase NM_001752 1.51 2.36 1.56 0.68 1.45 2.13 1.06 1.54 1.45 1.69 2.59 1.53
vanin 1 NM_004666 5.07 12.92 2.55 0.68 1.91 2.80 2.42 6.02 2.49 1.77 5.70 3.23
Ketogenesis
3-hydroxy-3-methylglutaryl-Coenzyme A synthase 2 (mitochondrial) NM_005518 32.40 156.62 4.83 1.13 2.51 2.22 2.14 6.57 3.07 2.20 6.53 2.96
Transport/strage
fatty acid binding protein 1, liver NM_001443 1.39 2.08 1.50 0.18 1.33 7.57 2.73 6.52 2.39 3.52 10.44 2.96
adipose differentiation-related protein NM_001122 1.96 3.44 1.75 0.22 2.90 13.02 3.33 6.06 1.82 3.40 5.73 1.68
C-terminal linking and modulating protein/PDZ domain containing 1 NM_002614 2.16 3.02 1.40 0.52 2.85 5.45 1.03 2.42 2.34 1.98 3.21 1.62
lipase, hepatic NM_000236 8.34 10.39 1.25 0.72 2.89 4.00 1.08 2.47 2.28 1.64 3.99 2.44
Gluconeogenesis
aquaporin 3 NM_004925 3.00 5.86 1.95 0.26 1.91 7.38 1.58 3.41 2.16 1.38 3.09 2.24
glycerol kinase NM_000167 0.87 1.57 1.79 0.73 0.79 1.08 1.00 1.21 1.21 1.84 3.17 1.73
phosphoenolpyruvate carboxykinase 1 (soluble) NM_002591 6.39 17.90 2.80 0.83 4.64 5.56 6.94 29.19 4.20 34.46 110.78 3.21
Metabolism
angiopoietin-like protein 4 NM_016109 1.71 17.97 10.54 3.00 25.67 8.56 1.24 7.27 5.84 3.88 33.56 8.66
heme oxygenase (decycling) 1 NM_002133 1.25 2.21 1.77 0.66 1.25 1.90 1.31 2.20 1.68 2.18 4.20 1.93
biliverdin reductase B (flavin reductase (NADPH)) NM_000713 1.07 1.65 1.55 0.84 0.86 1.03 0.87 1.58 1.81 1.36 4.03 2.97
sulfotransferase family, cytosolic, 2A, dehydroepiandrosterone (DHEA) -preferring, member 1 NM_003167 2.61 2.86 1.10 0.70 2.08 2.98 0.95 1.76 1.85 1.71 1.68 0.98
abhydrolase domain containing 3 NM_138340 1.68 3.51 2.08 0.54 2.16 3.96 1.76 3.30 1.88 2.01 3.99 1.99
Verification of gene expression profiles with quantitative PCR analyses
The expression of a subset of genes in response to treatment with PPAR ligands was further characterized using real-time quantitative PCR analyses. Total RNA samples were prepared from HepG2-tet-off-hPPAR cells, which were cultured in the presence or absence of Dox for 5 days, and subsequently treated with the PPAR ligands (100 μM fenofibric acid, 100 nM GW501516 or 10 μM ciglitizone for PPARα, PPARδ and PPARγ, respectively) or vehicle for 0, 8, 24, 48 or 72 h. Several PPAR-responsive genes (ADRP, 3-hydroxy-3-methylglutaryl-Coenzyme A synthase 2 (mitochondrial) (HMGCS2), hydroxyacyl-Coenzyme A dehydrogenase/3-ketoacyl-Coenzyme A thiolase/enoyl-Coenzyme A hydratase (trifunctional protein), α subunit (HADHA), phosphoenolpyruvate carboxykinase 1 (PEPCK), acyl-Coenzyme A dehydrogenase, C-4 to C-12 straight chain (MCAD), angiopoietin-like protein 4 (ANGPTL4), fatty acid binding protein 1, liver (L-FABP)) were up-regulated by the expression of PPARα and PPARγ (Figure 3). On the other hand, only ligand-activated PPARδ induced the expression of these genes, but unliganded PPARδ repressed them, in agreement with microarray analyses (Figure 3B and Table 1).
Figure 3 Eight gene expressions were modulated by PPARs in tet-regulated HepG2 cells. A, B, C and D, HepG2-tet-off-hPPAR cells were treated with DMSO (vehicle) or PPAR ligands (100 μM fenofibric acid for PPARα (A), 100 nM GW501516 for PPARδ (B), 10 μM ciglitizone for PPARγ1 (C) or PPARγ2 (D)) for 0, 8, 24, 48 or 72 h in the absence of Dox. Messenger RNA levels of human ADRP, HMGCS2, HADHA, PEPCK, MCAD, ANGPTL4, L-FABP and CLAMP/PDZK1 were measured by real time RT-PCR. Values are expressed as mean ± S.E.M. (n = 3) target mRNA levels normalized to GAPDH mRNA.
Although until now C-terminal linking and modulating protein/PDZ domain containing 1 (CLAMP/PDZK1) mRNA has not been reported as a target for PPARs, we observed that PPARs induced CLAMP/PDZK1 mRNA by microarray analyses (Table 1). We, therefore, verified the expression of CLAMP/PDZK1 mRNA using real-time quantitative PCR analysis. A similar result was observed on the expression pattern of CLAMP/PDZK1 mRNA (Figure 3).
Although all subtypes of PPARs could induce the expression of these genes, PPARα was more effective than other subtypes in the real-time PCR analyses (Figure 3).
PPARs modulate human ADRP promoter activity via a PPRE
To determine whether one of these genes was regulated by PPARs directly, we performed transient transfection experiments with the human ADRP promoter. It has been reported that the mouse ADRP promoter contains a functional PPRE [11]. Examination of the corresponding region of the human adrp gene indicated that the essential features of this element are likely to be conserved, and we found that the human ADRP promoter contained a potential PPRE at positions -2361 to -2345 (Figure 4A).
Figure 4 PPARs modulate human ADRP promoter activity via a PPRE located between -2361 and -2345 bp. A, A sequence corresponding to the -2366/-2339 region of the human adrp gene, was compared with a consensus PPRE and the analogous regions in the mouse adrp gene promoter (-2013/-1986). A human ADRPmut indicates a human ADRP promoter whose potential PPRE was mutated. Asterisks denote conserved bases and arrows represent response element half-sites. B, Schematic representation of the chimeric genes containing the human ADRP promoters; each wild type (hADRP-4K), point mutation (hADRP-mut), and deletion (hADRP-d1) of the ADRP promoter was cloned in front of the firefly luciferase reporter gene. Lowercase letters indicate mutations introduced in the human ADRP PPRE. C, D, E and F, HepG2 cells were co-transfected with a human ADRP reporter plasmid (50 ng), phRL-TK (50 ng) and either pcDNA3-hPPARα (5 ng) (C), pcDNA3-hPPARδ (5 ng) (D), pcDNA3-hPPARγ1 (5 ng) (E) or pcDNA3-hPPARγ2 (5 ng) (F). Transfected cells were treated with ligands (100 μM fenofibric acid (C), 100 nM GW501516 (D) or 10 μM ciglitizone (E and F)) for 24 h and the cells were used for reporter gene assays. Luciferase activities from reporter plasmids were normalized by internal Renilla luciferase activity. Values are expressed as fold induction of the control (the value when only reporter plasmid (ADRP-4K) was transfected) set at 1. Values represent the mean ± S.E.M. (n = 3).
To analyse whether this site can be regulated by PPARs, either the wild type 4-kb fragment of the human ADRP promoter (hADRP-4K), deletion promoter (hADRP-d1) whose potential PPRE is deleted or mutation promoter (hADRP-mut) whose potential PPRE is mutated, was cloned in front of the pGL3-luciferase reporter gene to construct reporter plasmids (Figure 4B). We co-transfected either the PPAR expression plasmid (pcDNA3-hPPARα, pcDNA3-hPPARδ, pcDNA3-hPPARγ1 or pcDNA3-hPPARγ2) or the vector plasmid (pcDNA3) as a control together with each reporter plasmid into HepG2 cells, and subsequently incubated these cells with or without ligands (100 μM fenofibric acid, 100 nM GW501516 or 10 μM ciglitizone for PPARα, PPARδ and PPARγ, respectively) for 24 h. Ligand-activated PPARs induced the promoter activity of the wild type 2 to 5-fold higher than the control (no PPAR expression without ligand) (Figure 4C–F). Interestingly, the expression of PPARδ without ligand diminished the wild type ADRP reporter gene expression to about 60% of the control in HepG2 cells (Figure 4D). We did not detect this observation when either deleted or mutated PPRE reporter plasmid was used. These data indicate that this PPRE in the ADRP promoter is a cis-acting element by which PPARs modulate human ADRP promoter activity.
To determine whether human PPARs/RXRα heterodimers bind this PPRE in the ADRP promoter, EMSAs were performed using this response element as a radiolabelled probe (Figure 5A–D). EMSAs revealed that all PPAR subtypes could bind the PPRE of the ADRP promoter in the presence of RXRα (closed arrowhead on lanes 2 to 4). Furthermore, these complexes were supershifted by anti-PPAR antibodies (open arrowhead on lane 5). This complex formation was competed by increasing amounts (10- and 100-fold excess) of unlabeled self-competitor (ADRP) and rat acyl CoA oxidase (ACO) PPRE fragments, but not by the mutated PPRE (ADRPmut) probes (lanes 6 to 11). Furthermore, no protein-DNA complex was observed when using the mutated PPRE (ADRPmut) probe (lane 12). Taken together, these data demonstrate that all PPARs bind to the same PPRE site at the position -2361 to -2345, and modulate human ADRP promoter activity.
Figure 5 PPARs bind to the PPRE in the -2366/-2339 region of the human adrp gene. A, B, C and D, EMSAs were performed with 32P-labelled either ADRP or mutated ADRP (ADRPmut) oligonucleotides in the presence of purified PPARα (A), purified PPARδ (B), in vitro transcribed/translated PPARγ1 (C), purified PPARγ2 (D) and/or purified RXRα proteins. Supershift experiments were carried out using anti-PPARα (A), anti-PPARδ (B) or anti-PPARγ (C and D) antibodies. Unlabelled oligonucleotides (ADRP, ADRPmut or ACO) were used at 10- or 100-fold molar excess to the labelled probe to perform competition assays. Closed and open arrowheads indicate the specific bands and the supershift bands, respectively.
To confirm that the PPARs/RXRα heterodimers bind to the PPRE in the ADRP promoter in vivo, we performed chromatin immunoprecipitation (ChIP) assays using an anti-PPARα, anti-PPARδ, anti-PPARγ, or anti-RXRα antibody. HepG2-tet-off-hPPAR cells were cultured in the absence of Dox for 5 days, and were subsequently treated with the PPAR ligands (100 μM fenofibric acid, 100 nM GW501516 or 10 μM troglitazone for PPARα, PPARδ and PPARγ, respectively) for 8 h and subsequently used for ChIP assays. PPARs/RXRα heterodimers bound to the PPRE in the ADRP promoter in vivo (Figure 6).
Figure 6 Chromatin immunoprecipitation assays of the ADRP promoter. HepG2-tet-off-hPPAR cells were treated with PPAR ligands (100 μM fenofibric acid for PPARα, 100 nM GW501516 for PPARδ, or 10 μM troglitazone for PPARγ) for 8 h in the absence of Dox and processed for the ChIP assays. Soluble chromatin was immunoprecipitated with pre-immune rabbit IgG (lanes 2 and 6), anti-PPAR antibodies (lanes 3 and 7), or anti-RXRα antibody (lanes 4 and 8). Immunoprecipitates were subjected to PCR with a primer-pair specific to the ADRP promoter. As a negative control, a second set of primers were used to amplify another genomic region that was not expected to interact with the PPARs. PCR was performed with total chromatin input (lanes 1 and 5).
Discussion
In the present study, in order to identify human PPARs-responsive genes in the liver, we have established tightly tet-regulatable human hepatoblastoma cell lines that can be induced to express each human PPAR (HepG2-tet-off-hPPAR cells). DNA microarray and real-time RT-PCR analyses using these stable cell lines indicate that PPARs activate gene expression involved in lipid metabolism, glucose homeostasis, and other activities (Table 1 and Figure 3). Cell proliferation of these cell lines was not affected until 7 days after removal of Dox. However, further investigation is required to define the influence of the expression of PPARs on the proliferation of these cell lines.
We identified a series of genes that are critical for many aspects of carbohydrate and lipid metabolism, including intracellular fatty acid transport, mitochondrial fatty acid β-oxidation, and ketogenesis by PPARα (Table 1). Indeed, PPARα regulates the expression of genes that encode for enzymes involved in peroxisomal proliferation and fatty acid oxidation in peroxisomes and mitochondria [2]. In addition, PPREs are present in the promoter region or the intronic sequence of these target genes [4]. However, our results show that PPARα and other PPAR subtypes induce expression of these genes. In fact, liver PPARγ regulated the expression levels of genes involved in lipogenesis, fatty acid transport, storage, and oxidation using both wild type mice and adipose-deficient mice [23]. Other reports also showed that PPARδ induced fatty acid oxidation by regulating genes involved in fatty acid transport, β-oxidation, and mitochondrial respiration in various tissues [7-9]. These reports support our results whereby each PPAR subtype induces fatty acid oxidation in the hepatoblastoma cell lines. In the real-time PCR analyses PPARα was more effective than other subtypes in the hepatoblastoma cell lines, although all subtypes could induce the expression of these genes (Figure 3). Thus, we assume that PPARα plays a major role in the liver, however, the mechanism of this effect remains uncertain.
Mammalian cells use glucose as a major energy source, and the origin of hepatic glucose production shifts from mainly glycogenolysis to gluconeogenesis as fasting prolongs to maintain blood glucose levels. The main precursors for hepatic gluconeogenesis are lactate, pyruvate, glycerol and alanine, which are converted into glucose via a series of reactions in the cytosol and mitochondria. PPARα plays a pivotal role in the management of energy stores during fasting [2]. Indeed, PPARα-null mice, fasted for 24 h, exhibit severe hypoglycaemia [24,25]. Recently, Patsouris et al. demonstrated that PPARα stimulates the expression of a set of genes involved in hepatic gluconeogenesis from glycerol [26]. In the present study, glycerol kinase and glycerol transporter aquaporin 3 were up-regulated by the liganded PPARs. Surprisingly, an unexpected finding was that the PEPCK was up-regulated by PPARs in these cell lines. PEPCK catalyses the rate-limiting step in liver gluconeogenesis. Way et al. showed that when Zucker diabetic fatty rats were treated with PPARγ agonists, PEPCK expression was decreased in the liver, and suggested that PPARγ agonists decrease gluconeogenesis [27]. Although we cannot fully explain this inconsistency, there are several reports, which support our results. An in vivo study showed that PEPCK mRNA levels are elevated in the livers of dexamethasone (DEX)-treated PPARα+/+ mice, and not DEX-treated PPARα-/- mice, and that PEPCK is increased when hepatic PPARα expression is reconstituted in DEX-treated PPARα-/-LDLR-/- mice. Furthermore, the combination of DEX and the potent PPARα ligand Wy14,643 increased the expression of PEPCK in human hepatocytes [28]. However, the reason for this differential induction remains unknown.
CLAMP, a four-PDZ-domain-containing protein also called PDZK1 [29], was identified as the scavenger receptor class B type I (SR-BI)-associated protein from rat liver membrane extracts [30]. CLAMP/PDZK1 regulates the stable SR-BI protein expression by a post-transcriptional mechanism [31]. The hepatic expression of SR-BI plays a critical role in lipoprotein metabolism, mainly due to its ability to mediate selective high density lipoproteins (HDL) cholesterol uptake [32,33]. HDL removes cholesterol from peripheral tissues and then transfers it to the liver. CLAMP/PDZK1 may modulate the intracellular transport and metabolism of cholesteryl esters taken up from HDL. Although until now, clamp/pdzk1 has not been reported as a direct PPAR target gene, we observed that CLAMP/PDZK1 mRNA was induced by PPARs in the established cell lines (Table 1 and Figure 3). These data suggest that CLAMP/PDZK1 is a target gene of PPARs. In the case of humans, PPRE exists in the CLAMP/PDZK1 promoter and PPARs can directly bind to this region (unpublished data). Therefore, PPAR activators may play a role in the regulation of HDL metabolism. However, further investigations are required to define the molecular mechanisms underlying the PPAR-mediated HDL metabolism.
Our microarray analyses indicated that there are many genes which are induced by all PPAR subtypes (Table 1, additional file 1). However, PPARα and PPARδ tend to induce the expression of a number of genes involved in the β-oxidation of fatty acids by mitochondria and peroxisomal fatty acid oxidation, and PPARγ tends to induce the expression of several genes involved in gluconeogenesis, lipid storage, transport and metabolism. Interestingly, we also observed that the genes involved in angiogenesis, cytoskeleton organization, signal transduction, protein modification and regulation of transcription were up-regulated in our HepG2-tet-off-hPPARγ cell lines. Indeed, there are several reports that PPARγ ligands modulate angiogenesis [34]. Further investigation is necessary to certify the subtype specific function of PPARs.
In the present study, we also observed that ligand-activated PPARδ induced target gene expression, however, unliganded PPARδ repressed these genes in the established cell lines (Table 1, Figures 3, 4). It has been reported that unliganded PPARδ binds to PPRE and recruits corepressors, such as a silencing mediator for retinoid and thyroid hormone receptor (SMRT), nuclear receptor corepressor (NCoR) etc. On the other hand, liganded PPARδ is thought to release the corepressor and form a complex with coactivators [35,36]. From this viewpoint, these established cell lines have a potential application to provide a high-throughput screening to detect the PPARδ ligands (see below).
To further investigate the regulation of target gene expression by PPARs, we analysed the human ADRP promoter in detail. ADRP, expressed ubiquitously, is a protein covering lipid droplets [37]. The mouse adrp gene exhibits a PPRE between -2004 and -1992 bp that binds PPARs/RXR heterodimers [11]. The sequence of the corresponding response element in the human gene between -2361 and -2345 bp is very similar to the mouse PPRE (Figure 4A). We performed EMSAs with this element and all PPAR subtypes bound to the same PPRE site of the ADRP promoter (Figure 5). Moreover, this human PPRE site confers transactivation by PPARs to a luciferase reporter gene, and mutations that disrupt the binding sequence of PPARs/RXR on this PPRE abolish transactivation (Figure 4). During the preparation of this manuscript, it was reported that the ADRP PPRE was the functional PPAR binding site and could be activated by human PPARα and δ [38]. In the present study, we showed the same result and that PPARγ1 and γ2 can also bind to PPRE and regulate the expression of the human adrp gene. Moreover, we were the first to confirm that PPARs/RXRα heterodimers bind to the PPRE in the ADRP promoter in vivo using ChIP assays (Figure 6).
PPARs are linked to metabolic disorders and, therefore, are interesting pharmaceutical targets. Among the synthetic ligands that activate these receptors, the fibrates are hypolipidemic compounds that activate PPARα. The thiazolidinediones, which selectively activate PPARγ, are hypoglycaemic molecules that are used to treat type II diabetes. Recently, we and other groups reported that PPARδ agonists might form effective drugs for obesity, diabetes, and cardiovascular disease [7-9,39,40]. From this viewpoint, these established cell lines have a potential application in various high-throughput assays. For instance, the reporter gene assay using the reporter plasmid containing the ADRP promoter would provide a useful detection system of PPARδ ligands as potential drug candidates in the HepG2-tet-off-hPPARδ cell line [41].
Conclusion
In conclusion, we established tightly tet-regulatable human hepatoblastoma cell lines that can be induced to express each human PPAR. These cell lines provided evidence that human PPARs are important regulators of lipid and glycerol homeostasis and new insights about the candidate target genes of PPARs. Therefore, these cell lines are powerful tools for analysing the function of human PPARs.
Methods
Materials
Fenofibric acid and GW501516 were synthesized as described previously [7]. Ciglitizone was purchased from Sigma. Troglitazone was purchased from Cayman Chemical.
Plasmid Constructs
Construction of pcDNA3-hPPARα, pcDNA3-hPPARδ, pcDNA3-hPPARγ1 and pcDNA3-hPPARγ2 expression plasmids were described previously [42]. Human PPARs fragments were excised from pcDNA3-hPPAR vectors. pBI-EGFP (Clontech) was digested with PvuII, and ligated with the PPAR fragment (termed pBI-EGFP-hPPAR).
To generate human ADRP promoter-reporter plasmids, human ADRP promoter containing -2981 to +1066 bp (hADRP-4K) and the deletion promoter containing -2345 to +1066 bp (hADRP-d1) were obtained by means of PCR with the bacterial artificial chromosome (BAC) clone plasmid (BACPAC resource center at Children's Hospital Oakland Research Institute) using a forward primer 5'-GGTACCTATCCCTGGTGCCAAAAAGGTTGGGGA-3' (including a KpnI site, underlined) for hADRP-4K or a forward primer 5'-GGTACCGAGAGTCTTCTGATGCAAAGTAAGAGG-3' (including a KpnI site, underlined) for hADRP-d1 and a reverse primer 5'-AGATCTTTTTCTTCCTGGAGAAAGAAATCTGCAGAAAAGAG-3' (including a BglII site, underlined). Each promoter was cloned into the KpnI-BglII sites of a pGL3-Basic vector (Promega). A point mutation was introduced into the ADRP promoter by PCR methodology. To generate a point-mutation construct (pGL3-hADRP-mut), a 101 bp mutant fragment was generated from forward primer (5'-GCAAAAAGAAGCTTGCTCAG-3') and reverse primer (5'-GACTCTCGCCCTTaagCtTgCggAATG-3') (Mutated bases are shown as lowercase letters). Then using this 101 bp mutant fragment for a forward primer, we amplified a 1273 bp mutant fragment using a reverse primer (5'-GTGCAGGGTTATGCATTGTT-3'). The 1273 bp mutant fragment was digested with Bpu1102I and EcoT22I, and then the fragment was inserted into the Bpu1102I/EcoT22I-digested pGL3-hADRP-4K vector.
Nucleotide sequences of these plasmids were confirmed by ABI PRISM® 310 Genetic Analyzer (Applied Biosystems).
Cell culture
HepG2 cells were cultured in DMEM (Nacalai tesque) containing 10% heat-inactivated charcoal/dextran treated foetal bovine serum (FBS) (HyClone), 100 IU/ml penicillin and 100 μg/ml streptomycin. The tightly tet-regulatable HepG2 cell clone (HY-Toff) which was transfected with the pTet-off vector, was isolated as previously described [22]. HY-Toff cells were co-transfected with pBI-EGFP-hPPAR and pBabepuro using TransIT®-LT1 Transfection Reagents (Mirus). The cells were cultured in a medium containing 600 μg/ml G418 (Nacalai tesque), 1 μg/ml puromycin (Sigma) and 1 μg/ml Dox (Clontech). G418- and puromycin-resistant clones were isolated. These clones were further screened for well-inducible clones by checking the expression of EGFP using fluorescence microscopy and the expression of PPAR proteins using immunoblot analysis in the absence of Dox.
Quantitative Real-Time PCR
Total RNA was isolated using an RNA preparation kit (Isogen; Nippon Gene Corp.). First strand cDNA was synthesized from 5 μg of total RNA of each cell sample using the SuperScript™ First-Strand Synthesis System for RT-PCR (Invitrogen) with oligo(dT)12–18 primer. The cDNAs were then used as templates for individual PCR reactions using specific primer sets (Table 2), which were designed by the Primer3 program written by the Whitehead Institute [43]. PCR reactions were carried out using QuantiTect™ SYBR® Green PCR Kit (Qiagen). The quantitative PCR analysis was performed using the DNA Engine Opticon™ System (Bio-Rad Laboratories). Amplification specificity was verified by visualizing PCR products on an ethidium bromide-stained 3% agarose gel. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used for normalizing each expression data.
Table 2 Primers used for Real-Time PCR Analysis. Sequences of forward (For.) and reverse (Rev.) primer for each target are shown. Sequences are 5' to 3'.
Gene Sequence Size (bp)
ADRP For. primer : TGAGATGGCAGAGAACGGTGTG 184
Rev. primer : GGCATTGGCAACAATCTGAGT
ANGPTL4 For. primer : GGGAGAGGCAGAGTGGACTATTT 96
Rev. primer : TTACTGTCCAGCCTCCATCTGA
CLAMP/PDZK1 For. primer : CTAAACTCTGCAGGCTGGCTAAA 101
Rev. primer : GCCCTTCTGTACCTCTTTGATGA
HADHA For. primer : CAAGGGCTTCCTAGGTCGTAAAT 155
Rev. primer : GGAACTGGATGTCTTCGTCTGAT
HMGCS2 For. primer : GGAACCCATATGGAGAATGTGT 168
Rev. primer : ATCGCTGCCAGCTTGCTT
GAPDH For. primer : TGGGTGTGAACCATGAGAAG 76
Rev. primer : GCTAAGCAGTTGGTGGTGC
L-FABP For. primer : TTGCCACCATGAGTTTCTCCG 82
Rev. primer : GGCAGACCGATTGCCTTCA
MCAD For. primer : TTCCAGAGAACTGTGGAGGTCTT 100
Rev. primer : TCAATAGCAGTCTGAACCCCTGT
PEPCK For. primer : TGCATGAAAGGTCGCACCA 192
Rev. primer : CACAGAATGGAGGCATTTGACA
PPARα For. primer : CTATCATTTGCTGTGGAGATCG 121
Rev. primer : AAGATATCGTCCGGGTGGTT
PPARδ For. primer : GTCACACAACGCTATCCGTTT 143
Rev. primer : AGGCATTGTAGATGTGCTTGG
PPARγ1 For. primer : CGTGGCCGCAGATTTGAA 166
Rev. primer : CTTCCATTACGGAGAGATCCAC
PPARγ2 For. primer : GGTGAAACTCTGGGAGATTCT 102
Rev. primer : CTCTGTGTCAACCATGGTCA
Immunoblot Analysis
Nuclear extracts were obtained as previously described [44]. Each nuclear extract (50 μg) was resolved by 10% SDS-PAGE, and electroblotted to nitrocellulose membranes. Western blot analyses were carried out using anti-human PPARα [H0723], PPARδ [K7701] (Perseus Proteomics Inc.) or PPARγ [E-8] antibodies (Santa Cruz). The signals were visualized with the ECL detection system (Amersham Biosciences).
Affymetrix Oligonucleotide Microarray Analysis
HepG2-tet-off-hPPAR cells were cultured in the presence (Dox) or absence of Dox for 5 days. In the case of the absence of Dox, the cells were treated with PPAR ligands (100 μM fenofibric acid for PPARα (Feno), 100 nM GW501516 for PPARδ (GW) or 10 μM ciglitizone for PPARγ (Cig)) or vehicle (DMSO) for 24 h. Total RNA samples were prepared from these cells using an RNA preparation kit (Isogen). Hybridization samples were prepared according to the Affymetrix protocol as previously described [7]. Briefly, 10 μg total RNA was used to generate first-strand cDNA. After second-strand cRNA synthesis, biotinylated and amplified RNAs were purified using RNeasy (Qiagen) and quantitated by a spectrophotometer. Biotinylated cRNA samples were then hybridized to Affymetrix Human Genome U133A arrays. These arrays contain probe sets for >22,000 transcripts and EST clones. After hybridization, microarrays were washed, scanned, and analysed with the GeneChip® software Microarray Suite (MAS) Ver.5.0 (Affymetrix). The criteria for selecting genes that were induced by ligand-activated PPARs were as follows: (1) in the presence of ligand, the average difference of gene was ≥ 100 and the gene represented as "presence"; (2) the ratio of the expression level in the presence of ligand to the expression level in the presence of Dox was greater than two; and (3) the ratio of the expression level in the presence of ligand to the expression level in the absence of ligand was greater than 1.5. Raw data are available at NCBI GEO, web page accession number GSE2699.
Luciferase Assay
HepG2 cells were transfected using Lipofectamine™ 2000 (Invitrogen) according to the manufacturer's instructions. HepG2 cells (3 × 104 cells/well) were seeded in 96-well plates 14–18 h before transfection. The cells were transfected with 50 ng of human ADRP reporter plasmid, 50 ng of phRL-TK (Promega) and either 5 ng of pcDNA3, pcDNA3-hPPARα, pcDNA3-hPPARδ, pcDNA3-hPPARγ1 or the pcDNA3-hPPARγ2 expression vector. Twenty-four hours following transfection, the cells were incubated with a medium containing dimethylsulfoxide (DMSO) (vehicle), 100 μM fenofibric acid, 100 nM GW501516 or 10 μM ciglitizone. Following a period of 24 h, both firefly and Renilla luciferase activities were quantified using a Dual-Luciferase® Reporter Assay System (Promega) according to manufacturer's instructions.
Electrophoretic Mobility Shift Assay (EMSA)
Human PPARα, human PPARδ, human PPARγ2 and human RXRα proteins were prepared using the IMPACT™-CN system (New England Biolabs). Human PPARγ1 protein was synthesized in vitro using the TNT® Quick Coupled Transcription/Translation Systems (Promega). Double-strand oligonucleotides were labelled with [α-32P]dCTP and a Klenow fragment and were used as probes. PPARα, PPARδ, PPARγ1, PPARγ2 and/or RXRα proteins were incubated with 32P-labelled probe in a total volume of 12.5 μl binding buffer (10 mM Tris-HCl (pH 7.5), 5% glycerol, 1 mM DTT, 1 mM EDTA, 1 μg poly (dI-dC), 30 μg BSA) at room temperature for 30 min, followed by an incubation at 4°C for 30 min. Supershift assays were performed by adding antibodies 1 h before incubation with an oligonucleotide probe at room temperature. All monoclonal antibodies (PPARα [H0723], PPARδ [K9418], PPARγ [A3408A]) were obtained from Perseus Proteomics. In competition studies, the proteins were pre-incubated with 10- or 100-fold molar excess of unlabelled wild-type or mutant oligonucleotides. Protein-DNA complexes were resolved on a 5% nondenaturing polyacrylamide gel in 1 × TAE buffer. The loaded gel was fixed with 10% methanol and 10% acetic acid, and then the gel was dried and autoradiographed. Double-stranded oligonucleotides composed of the following sequences were used for the binding and competition assays: human ADRP PPRE wild type, 5'-GCATTTTGTAGGTGAAAGGGCGAGAGTC-3'; ADRP PPRE mutant, 5'-GCATTccGcAaGcttAAGGGCGAGAGTC-3', and rat acyl-CoA oxidase (ACO) PPRE wild type, 5'-GCGGACCAGGACAAAGGTCACGTTC-3' (Mutated bases are shown as lowercase letters).
Chromatin Immunoprecipitation (ChIP) Assay
HepG2-tet-off-hPPAR cells were cultured in the absence of Dox for 5 days. The cells were treated with PPAR ligands (100 μM fenofibric acid for PPARα, 100 nM GW501516 for PPARδ, or 10 μM troglitazone for PPARγ) for 8 h. Cells were fixed in vivo at room temperature for 10 min by the addition of formaldehyde at a final concentration of 1% directly onto the cell culture media. Fixation was completed following the addition of glycine with a 0.125 M final concentration and the incubation was continued for a further 5 min. The cells were washed twice using ice-cold phosphate-buffered saline and collected. The cell pellets were washed with cell lysis buffer (10 mM Tris-HCl (pH 7.5), 10 mM NaCl, 3 mM MgCl2, 0.5% NP-40, and a protease inhibitor cocktail (Sigma)) three times, and dissolved in SDS lysis buffer (10 mM Tris-HCl (pH 7.5), 270 mM NaCl, 3 mM MgCl2, 1 mM CaCl2, 4% NP-40, 1.3% SDS, and a protease inhibitor cocktail) and remained on ice for 10 min. The cell lysates were sonicated to shear chromosomal DNA with an average length of 1000 bp. After centrifugation to remove insoluble materials, the chromatin solution was diluted 10-fold in an IP dilution buffer (20 mM Tris-HCl (pH 8.0), 150 mM NaCl, 2 mM EDTA, 1% Triton X-100, 0.01% SDS, and a protease inhibitor cocktail), and the diluted solution was pre-cleared with protein G Sepharose beads on a rotating wheel at 4°C for 1 h. Beads were removed by centrifugation and the supernatants were incubated with 2 μg of antibodies to PPARα (H-98, Santa Cruz), PPARδ (H-74, Santa Cruz), PPARγ (H-100, Santa Cruz), or RXRα (D-20, Santa Cruz) at 4°C overnight. For a negative control, pre-immune rabbit IgG (Santa Cruz) was incubated with the supernatant. The complexes were immunoprecipitated with protein G Sepharose beads. The beads were washed once with IP dilution buffer, twice with wash buffer 1 (20 mM Tris-HCl (pH 8.0), 150 mM NaCl, 2 mM EDTA, 1% Triton X-100, 0.1% SDS, and a protease inhibitor cocktail), once with wash buffer 2 (20 mM Tris-HCl (pH 8.0), 500 mM NaCl, 2 mM EDTA, 1% Triton X-100, 0.1% SDS, and a protease inhibitor cocktail), once with wash buffer 3 (10 mM Tris-HCl (pH 8.0), 1 mM EDTA, 0.25 M LiCl, 1% NP-40, 1% deoxycholate), and twice with TE buffer. Immune complexes were eluted from the beads in the elution buffer (25 mM Tris-HCl (pH 7.5), 5 mM EDTA, 0.5% SDS, 10 mM DTT) for 15 min. The proteins were removed from DNA by digesting with 1.5 mg/ml pronase at 42°C for 2 h. The crosslink was reversed by adding 5 M NaCl to a final concentration of 200 mM followed by incubation at 65°C for 6 h. The sample DNAs were then extracted with phenol-chloroform-isoamyl alcohol (25:24:1), precipitated with ethanol in the presence of glycogen, and resuspended in TE buffer. Similarly purified DNA fragments from the chromatin extracts (input) were used as a control for PCR reactions. Precipitated DNAs were analysed by PCR of 32 cycles using primers 5'-GCAAAAAGAAGCTTGCTCAG-3' and 5'-TGTTGCCATCTTCAGTGTTT-3' that flanked the PPRE of the ADRP promoter or primers 5'-ATGGTTGCCACTGGGGATCT-3' and 5'-TGCCAAAGCCTAGGGGAAGA-3' that are located about 6-kb upstream of the GAPDH promoter (negative control). PCR products were separated on a 2% agarose gel and stained with ethidium bromide.
List of Abbreviations
ACO, acyl-CoA oxidase; ADRP, adipose differentiation-related protein; ANGPTL4, angiopoietin-like protein 4; CLAMP/PDZK1, C-terminal linking and modulating protein/PDZ domain containing 1; DMSO, dimethylsulfoxide; Dox, doxycycline; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HADHA, hydroxyacyl-Coenzyme A dehydrogenase/3-ketoacyl-Coenzyme A thiolase/enoyl-Coenzyme A hydratase (trifunctional protein), α subunit; HDL, high density lipoprotein; HMGCS2, 3-hydroxy-3-methylglutaryl-Coenzyme A synthase 2 (mitochondrial); L-FABP, fatty acid binding protein 1, liver; MCAD, acyl-Coenzyme A dehydrogenase, C-4 to C-12 straight chain; PEPCK, phosphoenolpyruvate carboxykinase 1; PPAR, peroxisome proliferator-activated receptor; PPRE, peroxisome proliferator responsive element; RXR, retinoid X receptor; Tet, tetracycline.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
K. Tachibana carried out all the experiments and prepared the manuscript. Y. Kobayashi performed the reporter gene assays. T. Tanaka participated in designing the experiments. Y. Kobayashi and A. Sugiyama generated the human ADRP promoter-reporter plasmids. Y. Kobayashi, M. Tagami, T. Katayama, C. Ueda, D. Yamasaki, K. Ishimoto, and M. Sumitomo assisted with cell culture and the establishment of the stable cell lines. Y. Kobayashi, M. Tagami, T. Katayama, C. Ueda, and Y. Uchiyama performed the immunoblot analyses. T. Kohro assisted with the analysis of the microarray data. J. Sakai, T. Hamakubo, T. Kodama, and T. Doi developed the idea for the study, and participated in its design and coordination. All authors read and approved the final manuscript.
Supplementary Material
Additional File 1
Changes in mRNA expression levels in HepG2-tet-off-hPPAR cell lines by ligands. Microarray analyses were performed on HepG2-tet-off-hPPAR cells; the cells were cultured in the presence (Dox) or absence of Dox for 5 days. In the absence of Dox, the cells were treated with PPAR ligands (100 μM fenofibric acid for PPARα (Feno), 100 nM GW501516 for PPARδ (GW) or 10 μM ciglitizone for PPARγ (Cig)) or vehicle (DMSO) for 24 h. Gene expression profiles were compared between DMSO and Dox (DMSO versus Dox), ligands and Dox (Feno versus Dox, GW versus Dox, and Cig versus Dox were indicated in the case of PPARα, PPARδ and PPARγ, respectively) or ligands and DMSO (Feno versus DMSO, GW versus DMSO, and Cig versus DMSO were indicated in the case of PPARα, PPARδ and PPARγ, respectively). Average differences expressed the intensities of the mRNA levels in HepG2-tet-off-hPPARs cell lines. Samples were analysed using GeneChip® software Microarray Suite (MAS) Ver.5.0 (Affymetrix).
Click here for file
Acknowledgements
We thank Dr. Ying Huang and Dr. Yutaka Midorikawa for their help in establishing the stable cell lines, and Ms Akashi Izumi for the excellent technique used in the DNA microarray analysis experiments.
==== Refs
Mangelsdorf DJ Thummel C Beato M Herrlich P Schutz G Umesono K Blumberg B Kastner P Mark M Chambon P Evans RM The nuclear receptor superfamily: the second decade Cell 1995 83 835 839 8521507 10.1016/0092-8674(95)90199-X
Desvergne B Wahli W Peroxisome proliferator-activated receptors: nuclear control of metabolism Endocr Rev 1999 20 649 688 10529898 10.1210/er.20.5.649
Willson TM Brown PJ Sternbach DD Henke BR The PPARs: from orphan receptors to drug discovery J Med Chem 2000 43 527 550 10691680 10.1021/jm990554g
Mandard S Muller M Kersten S Peroxisome proliferator-activated receptor alpha target genes Cell Mol Life Sci 2004 61 393 416 14999402 10.1007/s00018-003-3216-3
Corton JC Lapinskas PJ Gonzalez FJ Central role of PPARalpha in the mechanism of action of hepatocarcinogenic peroxisome proliferators Mutat Res 2000 448 139 151 10725468
Cheung C Akiyama TE Ward JM Nicol CJ Feigenbaum L Vinson C Gonzalez FJ Diminished hepatocellular proliferation in mice humanized for the nuclear receptor peroxisome proliferator-activated receptor alpha Cancer Res 2004 64 3849 3854 15172993 10.1158/0008-5472.CAN-04-0322
Tanaka T Yamamoto J Iwasaki S Asaba H Hamura H Ikeda Y Watanabe M Magoori K Ioka RX Tachibana K Watanabe Y Uchiyama Y Sumi K Iguchi H Ito S Doi T Hamakubo T Naito M Auwerx J Yanagisawa M Kodama T Sakai J Activation of peroxisome proliferator-activated receptor delta induces fatty acid beta-oxidation in skeletal muscle and attenuates metabolic syndrome Proc Natl Acad Sci U S A 2003 100 15924 15929 14676330 10.1073/pnas.0306981100
Dressel U Allen TL Pippal JB Rohde PR Lau P Muscat GE The peroxisome proliferator-activated receptor beta/delta agonist, GW501516, regulates the expression of genes involved in lipid catabolism and energy uncoupling in skeletal muscle cells Mol Endocrinol 2003 17 2477 2493 14525954 10.1210/me.2003-0151
Wang YX Lee CH Tiep S Yu RT Ham J Kang H Evans RM Peroxisome-proliferator-activated receptor delta activates fat metabolism to prevent obesity Cell 2003 113 159 170 12705865 10.1016/S0092-8674(03)00269-1
Tan NS Michalik L Noy N Yasmin R Pacot C Heim M Fluhmann B Desvergne B Wahli W Critical roles of PPAR beta/delta in keratinocyte response to inflammation Genes Dev 2001 15 3263 3277 11751632 10.1101/gad.207501
Chawla A Lee CH Barak Y He W Rosenfeld J Liao D Han J Kang H Evans RM PPARdelta is a very low-density lipoprotein sensor in macrophages Proc Natl Acad Sci U S A 2003 100 1268 1273 12540828 10.1073/pnas.0337331100
Tontonoz P Hu E Graves RA Budavari AI Spiegelman BM mPPAR gamma 2: tissue-specific regulator of an adipocyte enhancer Genes Dev 1994 8 1224 1234 7926726
Tontonoz P Hu E Spiegelman BM Stimulation of adipogenesis in fibroblasts by PPAR gamma 2, a lipid-activated transcription factor Cell 1994 79 1147 1156 8001151 10.1016/0092-8674(94)90006-X
Vidal-Puig A Jimenez-Linan M Lowell BB Hamann A Hu E Spiegelman B Flier JS Moller DE Regulation of PPAR gamma gene expression by nutrition and obesity in rodents J Clin Invest 1996 97 2553 2561 8647948
Chao L Marcus-Samuels B Mason MM Moitra J Vinson C Arioglu E Gavrilova O Reitman ML Adipose tissue is required for the antidiabetic, but not for the hypolipidemic, effect of thiazolidinediones J Clin Invest 2000 106 1221 1228 11086023
Memon RA Tecott LH Nonogaki K Beigneux A Moser AH Grunfeld C Feingold KR Up-regulation of peroxisome proliferator-activated receptors (PPAR-alpha) and PPAR-gamma messenger ribonucleic acid expression in the liver in murine obesity: troglitazone induces expression of PPAR-gamma-responsive adipose tissue-specific genes in the liver of obese diabetic mice Endocrinology 2000 141 4021 4031 11089532 10.1210/en.141.11.4021
Bedoucha M Atzpodien E Boelsterli UA Diabetic KKAy mice exhibit increased hepatic PPARgamma1 gene expression and develop hepatic steatosis upon chronic treatment with antidiabetic thiazolidinediones J Hepatol 2001 35 17 23 11495036 10.1016/S0168-8278(01)00066-6
Kliewer SA Umesono K Noonan DJ Heyman RA Evans RM Convergence of 9-cis retinoic acid and peroxisome proliferator signalling pathways through heterodimer formation of their receptors Nature 1992 358 771 774 1324435 10.1038/358771a0
Hsu MH Palmer CN Song W Griffin KJ Johnson EF A carboxyl-terminal extension of the zinc finger domain contributes to the specificity and polarity of peroxisome proliferator-activated receptor DNA binding J Biol Chem 1998 273 27988 27997 9774413 10.1074/jbc.273.43.27988
Gervois P Chopin-Delannoy S Fadel A Dubois G Kosykh V Fruchart JC Najib J Laudet V Staels B Fibrates increase human REV-ERBalpha expression in liver via a novel peroxisome proliferator-activated receptor response element Mol Endocrinol 1999 13 400 409 10076997 10.1210/me.13.3.400
Schoonjans K Staels B Auwerx J Role of the peroxisome proliferator-activated receptor (PPAR) in mediating the effects of fibrates and fatty acids on gene expression J Lipid Res 1996 37 907 925 8725145
Huang Y Uchiyama Y Fujimura T Kanamori H Doi T Takamizawa A Hamakubo T Kodama T A human hepatoma cell line expressing hepatitis c virus nonstructural proteins tightly regulated by tetracycline Biochem Biophys Res Commun 2001 281 732 740 11237719 10.1006/bbrc.2001.4424
Gavrilova O Haluzik M Matsusue K Cutson JJ Johnson L Dietz KR Nicol CJ Vinson C Gonzalez FJ Reitman ML Liver peroxisome proliferator-activated receptor gamma contributes to hepatic steatosis, triglyceride clearance, and regulation of body fat mass J Biol Chem 2003 278 34268 34276 12805374 10.1074/jbc.M300043200
Kersten S Seydoux J Peters JM Gonzalez FJ Desvergne B Wahli W Peroxisome proliferator-activated receptor alpha mediates the adaptive response to fasting J Clin Invest 1999 103 1489 1498 10359558
Leone TC Weinheimer CJ Kelly DP A critical role for the peroxisome proliferator-activated receptor alpha (PPARalpha) in the cellular fasting response: the PPARalpha-null mouse as a model of fatty acid oxidation disorders Proc Natl Acad Sci U S A 1999 96 7473 7478 10377439 10.1073/pnas.96.13.7473
Patsouris D Mandard S Voshol PJ Escher P Tan NS Havekes LM Koenig W Marz W Tafuri S Wahli W Muller M Kersten S PPARalpha governs glycerol metabolism J Clin Invest 2004 114 94 103 15232616 10.1172/JCI200420468
Way JM Harrington WW Brown KK Gottschalk WK Sundseth SS Mansfield TA Ramachandran RK Willson TM Kliewer SA Comprehensive messenger ribonucleic acid profiling reveals that peroxisome proliferator-activated receptor gamma activation has coordinate effects on gene expression in multiple insulin-sensitive tissues Endocrinology 2001 142 1269 1277 11181544 10.1210/en.142.3.1269
Bernal-Mizrachi C Weng S Feng C Finck BN Knutsen RH Leone TC Coleman T Mecham RP Kelly DP Semenkovich CF Dexamethasone induction of hypertension and diabetes is PPAR-alpha dependent in LDL receptor-null mice Nat Med 2003 9 1069 1075 12847522 10.1038/nm898
Kocher O Comella N Tognazzi K Brown LF Identification and partial characterization of PDZK1: a novel protein containing PDZ interaction domains Lab Invest 1998 78 117 125 9461128
Ikemoto M Arai H Feng D Tanaka K Aoki J Dohmae N Takio K Adachi H Tsujimoto M Inoue K Identification of a PDZ-domain-containing protein that interacts with the scavenger receptor class B type I Proc Natl Acad Sci U S A 2000 97 6538 6543 10829064 10.1073/pnas.100114397
Silver DL A carboxyl-terminal PDZ-interacting domain of scavenger receptor B, type I is essential for cell surface expression in liver J Biol Chem 2002 277 34042 34047 12119305 10.1074/jbc.M206584200
Acton S Rigotti A Landschulz KT Xu S Hobbs HH Krieger M Identification of scavenger receptor SR-BI as a high density lipoprotein receptor Science 1996 271 518 520 8560269
Murao K Terpstra V Green SR Kondratenko N Steinberg D Quehenberger O Characterization of CLA-1, a human homologue of rodent scavenger receptor BI, as a receptor for high density lipoprotein and apoptotic thymocytes J Biol Chem 1997 272 17551 17557 9211901 10.1074/jbc.272.28.17551
Margeli A Kouraklis G Theocharis S Peroxisome proliferator activated receptor-gamma (PPAR-gamma) ligands and angiogenesis Angiogenesis 2003 6 165 169 15041792 10.1023/B:AGEN.0000021377.13669.c0
Shi Y Hon M Evans RM The peroxisome proliferator-activated receptor delta, an integrator of transcriptional repression and nuclear receptor signaling Proc Natl Acad Sci U S A 2002 99 2613 2618 11867749 10.1073/pnas.052707099
Krogsdam AM Nielsen CA Neve S Holst D Helledie T Thomsen B Bendixen C Mandrup S Kristiansen K Nuclear receptor corepressor-dependent repression of peroxisome-proliferator-activated receptor delta-mediated transactivation Biochem J 2002 363 157 165 11903058 10.1042/0264-6021:3630157
Brasaemle DL Barber T Wolins NE Serrero G Blanchette-Mackie EJ Londos C Adipose differentiation-related protein is an ubiquitously expressed lipid storage droplet-associated protein J Lipid Res 1997 38 2249 2263 9392423
Targett-Adams P McElwee MJ Ehrenborg E Gustafsson MC Palmer CN McLauchlan J A PPAR response element regulates transcription of the gene for human adipose differentiation-related protein Biochim Biophys Acta 2005 1728 95 104 15777674
Cheng L Ding G Qin Q Huang Y Lewis W He N Evans RM Schneider MD Brako FA Xiao Y Chen YE Yang Q Cardiomyocyte-restricted peroxisome proliferator-activated receptor-delta deletion perturbs myocardial fatty acid oxidation and leads to cardiomyopathy Nat Med 2004 10 1245 1250 15475963 10.1038/nm1116
Oliver WRJ Shenk JL Snaith MR Russell CS Plunket KD Bodkin NL Lewis MC Winegar DA Sznaidman ML Lambert MH Xu HE Sternbach DD Kliewer SA Hansen BC Willson TM A selective peroxisome proliferator-activated receptor delta agonist promotes reverse cholesterol transport Proc Natl Acad Sci U S A 2001 98 5306 5311 11309497 10.1073/pnas.091021198
Giddings SJ Clarke SE Gibson GG CYP4A1 gene transfection studies and the peroxisome proliferator-activated receptor: development of a high-throughput assay to detect peroxisome proliferators Eur J Drug Metab Pharmacokinet 1997 22 315 319 9512927
Tanaka T Takeno T Watanabe Y Uchiyama Y Murakami T Yamashita H Suzuki A Aoi R Iwanari H Jiang SY Naito M Tachibana K Doi T Shulman AI Mangelsdorf DJ Reiter R Auwerx J Hamakubo T Kodama T The generation of monoclonal antibodies against human peroxisome proliferator-activated receptors (PPARs) J Atheroscler Thromb 2002 9 233 242 12409633
Rozen S Skaletsky H Primer3 on the WWW for general users and for biologist programmers Methods Mol Biol 2000 132 365 386 10547847
Minami T Tachibana K Imanishi T Doi T Both Ets-1 and GATA-1 are essential for positive regulation of platelet factor 4 gene expression Eur J Biochem 1998 258 879 889 9874259 10.1046/j.1432-1327.1998.2580879.x
|
16197558
|
PMC1262764
|
CC BY
|
2021-01-04 16:37:44
|
no
|
Nucl Recept. 2005 Oct 3; 3:3
|
utf-8
|
Nucl Recept
| 2,005 |
10.1186/1478-1336-3-3
|
oa_comm
|
==== Front
Nutr JNutrition Journal1475-2891BioMed Central London 1475-2891-4-271620970610.1186/1475-2891-4-27ResearchEthnic variation in validity of classification of overweight and obesity using self-reported weight and height in American women and men: the Third National Health and Nutrition Examination Survey Gillum RF [email protected] Christopher T [email protected] Centers for Disease Control and Prevention, 3311 Toledo Road, Room 6323, Hyattsville, Maryland, 20782, USA2 National Institutes of Health, 6701 Rockledge Drive, Room 3146, Bethesda, Maryland, 20817, USA2005 6 10 2005 4 27 27 19 7 2005 6 10 2005 Copyright © 2005 Gillum and Sempos; licensee BioMed Central Ltd.2005Gillum and Sempos; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Few data have been published on the validity of classification of overweight and obesity based on self-reported weight in representative samples of Hispanic as compared to other American populations despite the wide use of such data.
Objective
To test the null hypothesis that ethnicity is unrelated to bias of mean body mass index (BMI) and to sensitivity of overweight or obesity (BMI >= 25 kg/m2) derived from self-reported (SR) versus measured weight and height using measured BMI as the gold standard.
Design
Cross-sectional survey of a large national sample, the Third National Health and Nutrition Examination Survey (NHANES III) conducted in 1988–1994.
Participants
American men and women aged 20 years and over (n = 15,025).
Measurements
SR height, weight, cigarette smoking, health status, and socio-demographic variables from home interview and measured weight and height.
Results
In women and Mexican American (MA) men SR BMI underestimated true prevalence rates of overweight or obesity. For other men, no consistent difference was seen. Sensitivity of SR was similar in non-Hispanic European Americans (EA) and non-Hispanic African Americans (AA) but much lower in MA. Prevalence of obesity (BMI >= 30 kg/m2) is consistently underestimated by self-report, the gap being greater for MA than for other women, but similar for MA and other men. The mean difference between self-reported and measured BMI was greater in MA (men -0.37, women -0.76 kg/m2) than in non-Hispanic EA (men -0.22, women -0.62 kg/m2). In a regression model with the difference between self-reported and measured BMI as the dependent variable, MA ethnicity was a significant (p < 0.01) predictor of the difference in men and in women. The effect of MA ethnicity could not be explained by socio-demographic variables, smoking or health status.
Conclusion
Under-estimation of the prevalence of overweight or obesity based on height and weight self-reported at interview varied significantly among ethnic groups independent of other variables.
OverweightObesityHispanicsMexican AmericansBody weightBlacks
==== Body
Introduction
Obesity, a prevalent and well-established risk factor for cardiovascular disease and diabetes, has been rising in prevalence over recent decades in the United States [1-7]. In 1999–2000, 64% of U.S. adults were found to be overweight or obese, an absolute increase of 8 percentage points from 1988–1994 [5]. Hence surveillance of obesity prevalence is of great public health interest. Because of the high costs of examination surveys, much of the surveillance of obesity is based on self-reported height and weight data, particularly at the state and local level. A number of studies have reported the validity of self-reported weight and height and BMI computed from them in populations of European extraction in the US and elsewhere [8-19]. A recent report noted lower validity of self-reported hypertension in Mexican Americans than in non-Hispanics [20]. However, data are lacking on the validity of self-reported obesity data in the growing Hispanic American population.
Therefore a study was done to test the null hypotheses that there is no association of ethnicity with the validity of BMI or "overweight or obesity" determined from self-reported height and weight and that the estimated measure of association of ethnicity with validity of self-reported overweight or obesity is not confounded by age, gender, education, marital status, region, smoking, and health status in the American population. Data on a large, multi-ethnic, national sample of adults from the Third National Health and Nutrition Examination Survey (NHANES III) were examined.
Methods
The Third National Health and Nutrition Examination Survey (NHANES III) was conducted in 1988–1994 on a nationwide multi-stage probability sample of 39,695 persons from the civilian, non-institutionalized population aged 2 months and over of the United States excluding reservation lands of American Indians. Mexican Americans and African Americans were oversampled. Of these, 30,818 (78%) were interviewed and examined. Details of the plan, sampling, operation and response have been published as have procedures used to obtain informed consent and to maintain confidentiality of information obtained [20,21].
Demographic data, years of education completed, medical history including self-assessed health status and behavioral information including smoking history were collected by household interview. Participants chose race and ethnicity categories from a card with categories including Mexican American (MA)[21]. Later in the household interview, interviewers asked, "How tall are you without shoes?" (inches) and "How much do you weigh without clothes or shoes?" (pounds). This occurred before they invited participants to take part in an examination and told them they would be measured and weighed.
Examinations were carried out in a mobile examination center. Technicians measured height to the nearest 0.1 centimeter, weight to the nearest 0.01 kg, as described in detail elsewhere [21-23]. Body mass index was computed (BMI = weight/height2, kg/m2). Obesity was defined as BMI >= 30.0 kg/m2. Overweight or obesity was defined as BMI >= 25.0 kg/m2 [5]. Extensive descriptive data on height, weight, BMI, and obesity prevalence have been published elsewhere and will not be duplicated here [23-26].
Statistical analysis
Of 33,994 interviewed persons, 31,311 were examined (78% of original sample). After exclusion of 14,281 persons under age 20y, 268 pregnant women, and persons with missing height or weight by self report or examination, 15,025 remained for this analysis. Self-reported height and weight were available for over 90% of persons with measured height and weight data. For stratified analysis, a cut-point of 60+ was used because persons 60+ were oversampled. Weighted descriptive statistics and measures of association were computed using the SAS [27-29]. The number of true positives (TP), false positives (FP), true negatives (TN) and false negatives (FN) were used to calculate sensitivity = (TP/(TP+FN))100, positive predictive value (PPV) = (TP/(TP+FP))100, specificity = (TN/(TN+FP))100, and predictive value negative (PVN) = (TN/(TN+FN))100. The BMI difference (error) was computed by subtracting the self-reported value from the measured value. Multivariate linear regression was used for models with BMI difference as a continuous dependent variable. All models controlled for age in years. All statistical testing and variance estimation were performed using the PROC REGRESS procedure for regression models in the SUDAAN system, which takes the complex cluster design of the survey into account [30,31].
Results
Table 1 compares prevalence of overweight or obesity (BMI >= 25 kg/m2) based on BMI computed from self-reported height and weight with that based on BMI computed from measured height and weight. In women, prevalence is consistently and substantially underestimated by self-report data, most markedly among MA both in absolute and relative terms. This was also true of MA men. For other men, no consistent difference was seen. Table 2 shows estimated prevalence of obesity (BMI >= 30 kg/m2) by self-report versus measurement. Prevalence is consistently underestimated by self-report, the gap being greater for MA than for other women, but similar for MA and other men. The gap was greater in women than men but varied little with age.
Table 1 Prevalence of overweight or obesity defined as body mass index >= 25 kg/m2) calculated from self-reported versus measured height and weight by gender, age, and ethnicity: NHANES III.(weighted)
Ethnicity
EA AA MA
Men
20–59y Self-report 59.2 58.4 59.4
Measured 58.7 56.4 63.4
Difference 0.5 0.2 -4.0
60+y Self-report 61.0 57.6 60.2
Measured 67.6 58.7 68.9
Difference -6.6 -1.1 -8.7
Women
20–59y Self-report 39.8 58.9 49.0
Measured 43.5 64.0 65.1
Difference -3.7 -5.1 -16.1
60+y Self-report 47.0 64.6 50.7
Measured 57.9 75.6 72.1
-10.9 -11.0 -21.4
N 6846 4431 3748
EA, Non-Hispanic European American, AA, Non-Hispanic African American, MA, Mexican American
Table 2 Prevalence (percent) of obesity defined as body mass index >= 30 kg/m2) calculated from self-reported versus measured height and weight by gender, age, and ethnicity: NHANES III.(weighted)
Ethnicity
EA AA MA
Men
20–59y Self-report 16.2 18.2 18.5
Measured 19.4 20.5 20.4
Difference -3.2 -2.3 -1.9
60+y Self-report 15.0 17.1 18.7
Measured 21.5 22.4 22.2
Difference -6.5 -5.3 -3.5
Women
20–59y Self-report 17.6 30.0 21.6
Measured 22.1 36.1 33.3
Difference -4.5 -6.1 -11.7
60+y Self-report 16.4 31.2 20.8
Measured 24.1 40.0 34.4
Difference -7.7 -8.8 -13.6
N 6846 4431 3748
EA, Non-Hispanic European American, AA, Non-Hispanic African American, MA, Mexican American WT VOB123003
Tables 3 and 4 show test characteristics for the classification of overweight or obesity based on BMI calculated from self-reported height and weight by age, gender and ethnicity. Sensitivity was similar in non-Hispanic European Americans (EA) and non-Hispanic African Americans (AA) but much lower in MA. Sensitivity was also lower in persons aged 60 years and over than < 60 years and in women than men. The highest sensitivity was in younger EA men (93%) and the lowest in older MA women (69%) (Table 3). PPV was similar across ethnic groups. It was higher in older persons and women. PPV was highest in older EA women (98%) and lowest in younger AA and MA men (89%).
Table 3 Sensitivity and positive predictive value (PPV) of overweight or obesity defined as body mass index >= 25 kg/m2 calculated from self-reported height and weight by gender, age, and ethnicity: NHANES III.(weighted)
Ethnicity
EA AA MA
Men
20–59y Sensitivity 93 92 83
PPV 92 89 89
60+y Sensitivity 87 88 82
PPV 96 90 94
Women
20–59y Sensitivity 88 89 72
PPV 96 97 95
60+y Sensitivity 80 83 69
PPV 98 97 95
N 6846 4431 3748
EA, Non-Hispanic European American, AA, Non-Hispanic African American, MA, Mexican American, PPV, positive predictive value
Table 4 Specificity and predictive value negative (PVN) of overweight or obesity defined as body mass index >= 25 kg/m2) calculated from self-reported height and weight by gender, age, and ethnicity: NHANES III.(weighted)
Ethnicity
EA AA MA
Men
20–59y Specificity 89 85 82
PVN 90 89 74
60+y Specificity 93 86 89
PVN 77 84 69
Women
20–59y Specificity 97 94 93
PVN 91 83 64
60+y Specificity 98 91 91
PVN 78 63 51
N 6846 4431 3748
EA, Non-Hispanic European American, AA, Non-Hispanic African American, MA, Mexican American
Specificity was highest in EA and similar in AA and MA (Table 4). It was higher in women than men and in older than younger men, but varied little with age in women. The highest specificity was in older EA women (98%) and the lowest in younger MA men (82%). In men, PVN was similar in EA and AA but much lower in MA within age groups. It was higher in younger than older men. In women, PVN was highest in EA, intermediate in AA and lowest in MA. It was higher in younger than older women, and lower than in men among AA and MA, but not EA women. The highest PVN was in younger EA women (91%) and the lowest in older MA women (51%).
The effect of smoking status, a possible effect modifier, on sensitivity of overweight or obesity based on self-reported height and weight was examined (Appendix Table A1, see Additional file 1). No consistent effect was seen in men. In women, sensitivity tended to be slightly higher in younger non-smokers than in smokers, but higher in older smokers than non-smokers, the difference being large only among older Mexican American women (smokers 82%, non-smokers 66%).
The percentages not overweight by self report (BMI < 25 kg/m2) with measured BMI >= 25 kg/m2, i.e. false negatives for overweight or obesity, among men and women by age, smoking and ethnicity were examined (Appendix Table 2, see Additional file 1). Rates of false negatives by self-report were slightly higher in non-smokers than in smokers, and in MA than in AA or EA. False negative rates were also higher in older than in younger persons and in less educated than more educated persons (not shown).
Tables 5 and 6 show the test characteristics for self-reported data for classifying persons as obese (BMI >= 30 kg/m2). The pattern is similar to that for overweight or obesity. A few exceptions are the similar sensitivity and PVN but lower PPV in MA men compared to other men.
Table 5 Sensitivity and positive predictive value (PVP) of obesity defined as body mass index >= 30 kg/m2 calculated from self-reported height and weight by gender, age, and ethnicity: NHANES III.(weighted)
Ethnicity
EA AA MA
Men
20–59y Sensitivity 77 78 75
PPV 93 88 83
60+y Sensitivity 65 67 70
PPV 93 88 83
Women
20–59y Sensitivity 76 77 61
PPV 95 93 94
60+y Sensitivity 66 73 57
PPV 98 93 94
N 6846 4431 3748
EA, Non-Hispanic European American, AA, Non-Hispanic African American, MA, Mexican American
Table 6 Specificity and predictive value negative (PVN) of obesity defined as body mass index >= 30 kg/m2) calculated from self-reported height and weight by gender, age, and ethnicity: NHANES III.(weighted)
Ethnicity
EA AA MA
Men
20–59y Specificity 99 97 96
PVN 95 94 94
60+y Specificity 99 97 96
PVN 91 91 92
Women
20–59y Specificity 99 97 98
PVN 93 88 83
60+y Specificity 99 96 98
PVN 90 84 81
N 6846 4431 3748
EA, Non-Hispanic European American, AA, Non-Hispanic African American, MA, Mexican American
The mean difference between self-reported and measured BMI was greater in Mexican Americans (men -0.37, women -0.76 kg/m2) than in non-Hispanic EA (men -0.22, women -0.62 kg/m2). In a regression model with the difference between self-reported and measured BMI as the dependent variable and MA ethnicity as the exposure variable and controlling for demographic variables or demographic variables plus smoking and health status, MA ethnicity was a significant predictor of BMI difference (Table 7).
Table 7 Adjusted regression coefficients (SE) of Mexican American ethnicity (yes/no) as a predictor of difference of measured and self-reported BMI by gender in NHANES III
Variable N Demographic-Adjusted* 95% CI Demographic-and health-adjusted+ 95% CI
Men 7552 1.28** 0.87–1.69 1.20** 0.81–1.59
Women 8142 2.35** 1.68–3.03 2.19** 1.51–2.87
CI, confidence interval
*adjusted for age (y), education < 12 years (yes/no), marital status (married/single), region (South vs other), metropolitan residence (yes/no).
+Adjusted for the above plus smoking status (smoker/nonsmoker) and poor self-reported health (yes/no)
** p < 0.01
An examination of mean differences between measured height and self-reported height indicated that MA men reported themselves to be 0.59 cm, EA men 1.47 cm and AA men 1.16 cm taller than measured. Over-reporting was greater at age 60+y than <60y, e.g. MA 60+ 2.00 cm, 20–59y 0.44 cm. MA men reported themselves to be 0.51 kg, EA men 0.26 kg and AA men 1.09 heavier than measured. Thus underestimates of overweight prevalence in MA men were due to over-reporting of height rather than underreporting of weight.
In women, MA reported themselves to be 0.51 cm, EA 0.26 cm and AA 1.09 cm taller than measured. Over-reporting was greater at age (years) 60+ than <60; e.g. MA 60+ 3.02 cm, 20–59 0.92 cm. MA women reported themselves to be 1.07 kg, EA women 1.21 kg, and AA women 1.58 kg lighter than measured. Under-reporting of weight was greater at age 20–59 than at 60+; e.g. MA 20–59, 1.18 kg; 60+, 0.28 kg. So underestimates of overweight prevalence in each group of women was due primarily to underreporting of weight at younger ages and both to over-reporting height and under-reporting weight at older ages.
Discussion
Overweight/obesity is one of the leading preventable causes of death in the United States and most industrialized countries [1-7,23-26,32-34]. After release of national data from NHANES III (1988–1994) showed striking increases in US obesity prevalence among children and adults compared to earlier surveys, it was widely recognized that an "epidemic" was occurring [23-26]. Therefore monitoring of the prevalence of obesity by interview surveys, such as the National Health Interview Survey (NHIS) and the Behavioral Risk Factor Surveillance System (BRFSS), which rely on self-reported data, have assumed even greater importance than previously [34]. Yet, all current guidelines for the diagnosis and management of obesity are based on measured height and weight.
Over-sampling of the MA population and availability of measured as well as self-reported height and weight data in NHANES III made it possible to evaluate the ethnic variation in validity of self-reported data. NHANES III data show that using self-reported data produces consistent underestimates of the prevalence of overweight and obesity (BMI >= 25 kg/m2) in MA by up to 21 percentage points and of obesity (BMI >= 30 kg/m2) by up to 14 percentage points (Table 1). Criteria for deciding whether SR data are to be considered valid will depend on the application and on a number of factors including statistical test characteristics and expert judgement as discussed at length elsewhere [35]. However, these data suggest that self-reported data for MA are not sufficiently valid for use in producing such estimates and are suspect for etiologic research. Further research is needed determine the applicability of these findings to data from telephone interviews or self-completed questionnaires.
At a mechanistic level, the important underestimate of BMI and hence overweight and obesity prevalence in NHANES in women was due primarily to under-reporting of weight at younger ages and both to over-reporting height and under-reporting weight at older ages. Possible explanations for under-reporting of weight in women have been discussed [8-19]. In MA women and men, these may include recall bias and lack of information because of no recent measurements at home or by health care providers. This is important for both weight and height, since stature declines with age in older persons by up to 2 cm per decade after age 30 [8]. Lack of access to scales and health care and/or lack of utilization of either may be due to barriers such as cost, transportation, language or cultural factors. Poorer validity of self-reported hypertension has been noted in MA, perhaps for similar reasons [20]. Further, if obesity and short stature are seen as socially undesirable, a conscious or unconscious tendency to under-report body weight and over-report height may occur even if true weight and height are known. On the other hand, MA women have been reported to be less likely to perceive themselves as overweight and more satisfied with body size than non-Hispanic EA women in some, but not all studies [36-39]. Specifically in NHANES III, participants were asked, "Do you consider yourself now to be overweight, underweight, or about the right weight?" Hispanic women and men were more likely to under-assess their overweight status than non-Hispanic EA; e.g. 31% of overweight Hispanic women said they were "about the right weight" compared to 14% of non-Hispanic EA [37].
Comparisons with previous reports
NHANES III is one of the largest studies to provide population-based data on the validity of prevalence estimates based on self report and one of the first to provide such data for Mexican Americans. Two previous reports were based on a nationally representative sample of US adults. One used NHANES III data to assess the effect of age on validity of self-reported data [8]. Under-reporting of height and consequently BMI bias was higher above age 60 than below. However, data from all ethnic groups were combined. The present report demonstrates important, consistent ethnic differences between MA and other groups. Patterns of reporting error in weight by gender and age were similar in NHANES II and III [8,9]. A non-significant effect of race ("white"/"black") on reporting bias for both weight and height was observed in NHANES II. In European and Australian adults, height was over-reported and weight under-reported, producing under-estimates of BMI and obesity prevalence [15-19]. An exception was the under-reporting of both weight and height in Scotland leading to BMI and prevalence estimates slightly higher than from measurement [10]. Brazilian adults reported height and weight with small errors that differed by gender [11]. In studies of American adolescents, height was over-reported and weight and consequently BMI and obesity prevalence under-reported [14]. The latter two studies did not describe ethnic variation. One report indicates that correction equations do not eliminate systematic error in self-reported BMI [40]. Comparisons of self-reported and measured BMI in etiologic research was beyond the scope of the present paper [41].
Among these studies, self-reported height and weight were obtained by in-person interview in several [8,9,11,13], by telephone interview [18], and by self-completed questionnaire others [10,14-17]. No study used more than one of these methods for self-reported data. Quantitative comparisons of error magnitude of self-reported data among studies are not possible due to varying designs, analytic methods, and reporting. A large British survey that used a self-completed questionnaire found mean differences between self-reported and measured BMI that ranged for -0.55 in women 35–49 years to -1.13 in men 60+ years, somewhat larger than in NHANES. An Australian telephone survey found differences in prevalence of overweight or obesity (percentage points) of -23 in men and -15 in women, much larger than those see in EA in NHANES. Variation in differences by ethnicity was not examined in other studies.
Limitations and strengths
Several unavoidable limitations of the present study include possible bias arising from survey non-response and from missing values for some variables. Several special studies of NHANES III data have indicated little bias due to non-response [42]. The 2 to 4 week time lapse between ascertainment of self-reported data and measured data is unlikely to affect the difference in this adult population [8]. Compared to interview surveys such as NHIS or BRFSS, error in self-reported height and weight might be underestimated in NHANES III if subjects were aware before the interview that they would be weighed and measured during the subsequent examination and hence might be less likely to misstate height and weight. At the time of the home interview when participants reported their current height and weight, participants had not been invited to the NHANES III examination or told that the examination would include height and weight measurements. However, because of pre-survey media publicity or explicit questions asked of interviewers, some participants might have known or assumed they would later be measured and weighed. It is not possible to determine what effect, if any, this may have had on height and weight reporting.
Confounding by variables not controlled for cannot be excluded. However, given the uncertainty about the existence or nature of the association, it is unclear which other variables should be controlled for as confounders. The number of tests was restricted to those of weighted regression models. The representativeness of the sample and the use of sample weights provide generalizability of the results to United States non-institutionalized population of the same ages, but not necessarily to Mexico or other nations or smaller US ethnic groups such as Cubans.
In order to monitor overweight prevalence in Cuban and Puerto Rican Americans and American Indians using interview survey data, replication of the current findings in these groups is needed. Participants might be asked when their weight and their height were last measured and by whom. The validity of self-report data obtained by in-person interview, telephone interview, and self-completed questionnaire should be compared to establish generalizablity of validity estimates from the former to the latter two. Trends in validity over time should also be sought in MA using more recent surveys and related to possible correlates.
Conclusion
Self reported height, weight and BMI in Mexican Americans may not be of sufficient validity for use by public health agencies and they underestimate obesity prevalence in women of all ethnic groups.
List of Abbreviations
AA, African American
BMI, body mass index
EA, European American
MA, Mexican American
NHANES, National Health and Nutrtion Examination Survey
Supplementary Material
Additional File 1
Appendix Tables 1 and 2
Click here for file
Acknowledgements
We acknowledge the staff and contractors of the Division of Health Examination Statistics of the National Center for Health Statistics, Centers for Disease Control and Prevention, who conducted the survey and prepared the data for analysis and Ms. Catherine Duran, who assisted with computer programming and Mr. J. Lubitz, and Drs. D Makuc and J. Madans who provided helpful comments on the manuscript.
==== Refs
Bray GA An atlas of obesity and weight control 2003 Boca Raton, Parthenon Pub. Group
Racette SB Deusinger SS Deusinger RH Obesity: overview of prevalence, etiology, and treatment Phys Ther 2003 83 276 88 12620091
Flegal KM Carroll MD Ogden CL Johnson CL Prevalence and trends in obesity among US adults, 1999–2000 JAMA 2002 288 1723 7 12365955 10.1001/jama.288.14.1723
Pawson IG Martorelli R Mendoza FE Prevalence of overweight and obesity in US Hispanic populations Am J Clin Nutr 1991 1522S 8S 2031482
National Center for Health Statistics Health United States 2003 with chartbook on trends in the health of Americans 2003 Hyattsville, Maryland 26 31
Wing RR Obesity and related eating and exercise behaviors in women Ann Behav Med 1993 15 124 134
Najjar MF Rowland M Anthropometric reference data and prevalence of overweight, United States, 1976–80 Vital Health Stat 1987 11 1 73
Kuczmarski MF Kuczmarski RJ Najjar M Effects of age on validity of self-reported height, weight, and body mass index: findings from the Third National Health and Nutrition Examination Survey, 1988–1994 J Am Diet Assoc 2001 101 28 34 11209581 10.1016/S0002-8223(01)00008-6
Rowland ML Self-reported weight and height Am J Clin Nutr 1990 52 1125 1133 2239790
Bolton-Smith C Woodward M Tunstall-Pedoe H Morrison C Accuracy of the estimated prevalence of obesity from self reported height and weight in an adult Scottish population J Epidemiol Community Health 2000 54 143 8 10715748 10.1136/jech.54.2.143
Schmidt MI Duncan BB Tavares M Polanczyk CA Pellanda L Zimmer PM Validity of self-reported weight – a study of urban Brazilian adults Rev Saude Publica 1993 27 271 6 Epub 2003 Nov 03 8209159
Kuskowska-Wolk A Karlsson P Stolt M Rossner S The predictive validity of body mass index on self-reported weight and height Int J Obes 1989 13 441 453 2793299
Nieto-Garcia FJ Bush TL Keyl PM Body mass definitions of obesity: sensitivity and specificity using self-reported weight and height Epidemiology 1990 1 146 152 2073502
Brener ND Mcmanus T Galuska DA Lowry R Wechsler H Reliability and validity of self-reported height and weight among high school students J Adolesc Health 2003 32 281 7 12667732 10.1016/S1054-139X(02)00708-5
Roberts RJ Can self-reported data accurately describe the prevalence of overweight? Public Health 1995 109 275 84 7667492
Spencer EA Appleby PN Davey GK Key TJ Validity of self-reported height and weight in 4808 EPIC-Oxford participants Public Health Nutr 2002 5 561 5 12186665 10.1079/PHN2001322
Niedhammer I Bugel I Bonenfant S Goldberg M Leclerc A Validity of self-reported weight and height in the French GAZEL cohort Int J Obes Relat Metab Disord 2000 24 1111 8 11033979 10.1038/sj.ijo.0801375
Flood V Webb K Lazarus R Pang G Use of self-report to monitor overweight and obesity in populations: some issues for consideration Aust N Z J Public Health 2000 24 96 9 10777989
Willet W 2 Nutritional epidemiology 1998 New York, Oxford University Press 244 272
Vargas CM Burt VL Gillum RF Pamuk ER Validity of self-reported hypertension in the National Health and Nutrition Examination Survey III, 1988–1991 Prev Med 1997 26 678 685 9327477 10.1006/pmed.1997.0190
Plan and operation of the Third National Health and Nutrition Examination Survey, 1988–94 National Center for Health Statistics Vital Health Stat 1994 1 27
Westat Inc National Health and Nutrition Examination Survey III. Body Measurements (Anthropometry) National Center for Health Statistics 1996
Flegal KM Carroll MD Kuczmarski RJ Johnson CL Overweight and obesity in the United States: prevalence and trends, 1960–1994 Int J Obes Relat Metab Disord 1998 22 39 47 9481598 10.1038/sj.ijo.0800541
Flegal KM Troiano RP Changes in the distribution of body mass index of adults and children in the US population Int J Obes Relat Metab Disord 2000 24 807 18 10918526 10.1038/sj.ijo.0801232
Kuczmarski RJ Carroll MD Flegal KM Troiano RP Varying body mass index cutoff points to describe overweight prevalence among U.S. adults: NHANES III (1988 to 1994) Obes Res 1997 5 542 8 9449138
Kuczmarski RJ Flegal KM Campbell SM Johnson CL Increasing prevalence of overweight among US adults. The National Health and Nutrition Examination Surveys, 1960 to 1991 JAMA 272 205 11 1994 Jul 20 8022039 10.1001/jama.272.3.205
SAT/STAT software: changes and enhancements, release 8.2 2001 Cary, NC: SAS Institute Inc 1 324
Spector PE SAS programming for researchers and social scientists 2001 2 Thousand Oaks, Calif: Sage Publications 1 231
Freund RJ Rudolf J, Freund, Ramon C Littell SAS system for regression 2000 3 Cary NC: SAS Institute Inc 1 236
Kleinbaum DG Kupfer LL Muller KE Nizam A Applied regression analysis and other multivariate methods 1998 Pacific Grove, California, Brooks/Cole Publishing Company
Shah BV Barnwell BG Bieler GS SUDAAN Software for the Statistical Analysis of Correlated Data User's Manual Release 7.0 1996 Research Triangle Park, North Carolina: Research Triangle Institute
McGinnis JM Foege WH Actual causes of death in the United States JAMA 1993 270 2207 12 8411605 10.1001/jama.270.18.2207
James PT Leach R Kalamara E Shayeghi M The worldwide obesity epidemic Obes Res 2001 9 228S 233S 11707546
Centers for Disease Control and Prevention (CDC) Declining prevalence of no known major risk factors for heart disease and stroke among adults – United States, 1991–2001 MMWR Morb Mortal Wkly Rep 53 4 7 2004 Jan 16 14724558
Wallace RB Wallace RB Screening for early and asymptomatic conditions Public Health and Preventive Medicine 1998 14 New York, Appleton and Lange
Anderson LA Eyler AA Galuska DA Brown DR Brownson RC Relationship of satisfaction with body size and trying to lose weight in a national survey of overweight and obese women aged 40 and older, United States Prev Med 2002 35 390 6 12453717 10.1006/pmed.2002.1079
Kuchler F Variyam JN Mistakes were made: misperception as a barrier to reducing overweight Int J Obes Relat Metab Disord 2003 27 856 61 12821973 10.1038/sj.ijo.0802293
Cachelin FM Rebeck RM Chung GH Pelayo E Does ethnicity influence body-size preference? A comparison of body image and body size Obes Res 2002 10 158 66 11886938
Fitzgibbon ML Blackman LR Avellone ME The relationship between body image discrepancy and body mass index across ethnic groups Obes Res 2000 8 582 9 11156434
Plankey MW Stevens J Flegal KM Rust PF Prediction equations do not eliminate systematic error in self-reported body mass index Obes Res 1997 5 308 14 9285836
Santillan AA Camargo CA Body mass index and asthma among Mexican adults: the effect of using self-reported vs measured weight and height Int J Obes Relat Metab Disord 2003 27 1430 3 14574357 10.1038/sj.ijo.0802395
Mohadjer L Bell B Waksberg J National Health and Nutrition Examination Survey III. Accounting for item nonresponse bias 1996 Rockville, MD, WESTAT, Inc
|
16209706
|
PMC1262765
|
CC BY
|
2021-01-04 16:39:31
|
no
|
Nutr J. 2005 Oct 6; 4:27
|
utf-8
|
Nutr J
| 2,005 |
10.1186/1475-2891-4-27
|
oa_comm
|
==== Front
Nutr Metab (Lond)Nutrition & Metabolism1743-7075BioMed Central London 1743-7075-2-221615330610.1186/1743-7075-2-22ReviewMelatonin, a potent agent in antioxidative defense: Actions as a natural food constituent, gastrointestinal factor, drug and prodrug Hardeland Rüdiger [email protected] SR [email protected] Institute of Zoology and Anthropology, University of Göttingen, Berliner Str. 28, D-37073 Göttingen, Germany2 Comprehensive Center for Sleep Medicine, Department of Pulmonary, Critical Care and Sleep Medicine, Mount Sinai School of Medicine, 1176 - 5th Avenue, New York, NY 10029, USA2005 10 9 2005 2 22 22 7 8 2005 10 9 2005 Copyright © 2005 Hardeland and Pandi-Perumal; licensee BioMed Central Ltd.2005Hardeland and Pandi-Perumal; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Melatonin, originally discovered as a hormone of the pineal gland, is also produced in other organs and represents, additionally, a normal food constituent found in yeast and plant material, which can influence the level in the circulation. Compared to the pineal, the gastrointestinal tract contains several hundred times more melatonin, which can be released into the blood in response to food intake and stimuli by nutrients, especially tryptophan. Apart from its use as a commercial food additive, supraphysiological doses have been applied in medical trials and pure preparations are well tolerated by patients. Owing to its amphiphilicity, melatonin can enter any body fluid, cell or cell compartment. Its properties as an antioxidant agent are based on several, highly diverse effects. Apart from direct radical scavenging, it plays a role in upregulation of antioxidant and downregulation of prooxidant enzymes, and damage by free radicals can be reduced by its antiexcitatory actions, and presumably by contributions to appropriate internal circadian phasing, and by its improvement of mitochondrial metabolism, in terms of avoiding electron leakage and enhancing complex I and complex IV activities. Melatonin was shown to potentiate effects of other antioxidants, such as ascorbate and Trolox. Under physiological conditions, direct radical scavenging may only contribute to a minor extent to overall radical detoxification, although melatonin can eliminate several of them in scavenger cascades and potentiates the efficacy of antioxidant vitamins. Melatonin oxidation seems rather important for the production of other biologically active metabolites such as N1-acetyl-N2-formyl-5-methoxykynuramine (AFMK) and N1-acetyl-5-methoxykynuramine (AMK), which have been shown to also dispose of protective properties. Thus, melatonin may be regarded as a prodrug, too. AMK interacts with reactive oxygen and nitrogen species, conveys protection to mitochondria, inhibits and downregulates cyclooxygenase 2.
==== Body
Introduction
In several countries, melatonin is sold over the counter; in others its free sale is prohibited. The usefulness of melatonin as a food additive continues to be a matter of debate. Meanwhile, countless people have used melatonin for mitigating the symptoms of jet lag, an application which has been tested and is recommended [1-4]; any person we have spoken to has reported positive experiences. Melatonin has been and is being used in several clinical trials with different therapeutic approaches. In some of these studies, in addition to improvements of sleep, the repeatedly demonstrated antioxidant properties [5-10] were the main reason for testing the pineal hormone. This holds especially for the treatment of neurodegenerative disorders, such as Alzheimer's disease [11-13] and amyotrophic lateral sclerosis [14].
In terms of application it seems necessary to thoroughly analyze the mechanisms of antioxidant actions of melatonin and to distinguish between effects observed at pharmacological or physiological concentrations. These considerations must not be restricted to the melatonin released from the pineal gland into the circulation and to the classic hepatic degradation route of 6-hydroxylation followed by conjugation. On the contrary, we would like to lay emphasis on the significance of tissue melatonin and the alternate oxidative pathways of catabolism leading to different, biologically active products. The relationship between melatonin and nutrition will be discussed, with regard to the presence of the compound as a natural food constituent sometimes affecting circulating levels, to the post-prandial release of melatonin from the gastrointestinal tract, and to interactions with other antioxidants present in food. Finally, a model of mitochondrial protection is reviewed.
Melatonin in food and in the gastrointestinal tract
Melatonin is a natural compound of almost ubiquitous occurrence [15-17]. Its presence was demonstrated in all major taxa of organisms, as far as tested, including bacteria, unicellular eukaryotes, macroalgae, plants, fungi and invertebrate animals. Several studies dealt with melatonin in edible plants [8,18-25]. One can conclude that relevant quantities of melatonin are present in most vegetables, fruit, nuts and cereals. However, the precise melatonin contents are sometimes affected by some uncertainties which result from particular methodological problems arising in material from photoautotrophic organisms. First, melatonin can be easily destroyed by oxidants during extraction [26], and, second, false positive and false negative data are easily obtained due to the presence of secondary plant metabolites, either mimicking melatonin or interfering with it in the assays [16,17,21,22]. It is a strict requirement to apply preservative conditions of extraction, to control the yield by determinations of recovery, and to obtain data by two methodologically different procedures. Although this has not been done in any plant tested, the widespread occurrence of melatonin in plants is beyond doubt. To date, the presence of melatonin was demonstrated in more than 20 dicot and monocot families. Usually, the amounts of melatonin reported varied considerably between species and between plant tissues, from the detection threshold to several hundred pg/g fresh weight. One should, however, be aware that these concentrations frequently greatly exceed avian and mammalian blood levels, which rarely attain more than 200 pg/mL during the nocturnal maximum, and can remain below 10 pg/mL during the day. Intestinal resorption of dietary melatonin should not be a particular problem because the amphiphilic molecule can easily cross any membrane. Therefore, an efficient uptake of the indoleamine from food should be expected to influence the blood plasma concentration (see below). Melatonin was observed to be elevated in alpine and mediterranean plants exposed to strong UV irradiation [25], a finding which may be seen in relation to melatonin's antioxidant properties antagonizing damage by light-induced oxidants. It is particularly worth mentioning the very high levels reported for several seeds and medicinal plants [8,15,24,27,28] (Table 1). The high amounts frequently found in seeds may be interpreted in terms of antioxidative protection within a dormant and more or less dry system, in which enzymes are poorly effective and cannot be upregulated, so that low molecular weight antioxidants such as melatonin are of advantage [20]. Moreover, melatonin's amphiphilicity may favor its accumulation especially in oily seeds.
Table 1 Particularly high melatonin levels reported for several edible and medicinal plants (selected examples).
Species Tissue Melatonin [ng/g] References
(A) Edible plants
Lycopersicon esculentum (tomato) fruit 0.5 [18]
Raphanus sativus (red radish) root tuber 0.6 [19]
Brassica campestris (Japanese radish) stem, leaves 0.6 [19]
Brassica nigra (black mustard) seed 129 [24,28]
Brassica hirta (white mustard) seed 189 [24,28]
Prunus cerasus (tart cherry, Montmorency) fruit 15–18 [23,24]
Prunus amygdalus (almond) seed 39 [28]
Pimpinella anisum (anise) seed 7 [24,28]
Foeniculum vulgare (fennel) seed 28 [24,28]
Helianthus annuus (sunflower) seed 29 [24,28]
Oryza sativa (rice) seed 1 [19]
Zea mays (Indian corn) seed 1.3 [19]
Avena sativa (oat) seed 1.8 [19]
Festuca arundinacea (tall fescue) seed 5 [19]
Elettaria cardamomum (green cardamom) seed 15 [24,28]
Zingiber officinale (ginger) tuber 0.5 [19]
Musa paradisiaca (banana) fruit 0.5 [18]
(B) Officinal plants
Melissa officinalis (balm mint) young plant 16 [25]
Scutellaria baicalensis (huang-qin) plant > 2,000 – > 7,000 [24,25,27]
Pimpinella peregrina (-) dried root 38 [25]
Hypericum perforatum (St. Johns wort) leaf 1,750 [27]
Hypericum perforatum (St. Johns wort) flower > 2,400 – > 4,000 [25,27]
Lippia citriodora (lemon verbena) young plant 22 [25]
Tanacetum parthenium (feverfew) leaf (fresh/dried) > 1,300/> 7,000 [24,25,27]
In some of the medicinal plants, interactions or synergisms of melatonin with secondary metabolites may be of importance. In Scutellaria baicalensis, e.g., melatonin is accompanied by acteoside, baicalein, baicalin, wogonin and ganhuangenin, substances with antioxidant, antiinflammatory, sedating and immunomodulatory properties, interfering also with NO synthases and P450 monooxygenases, i.e., functions within the action spectrum of melatonin or affecting melatonin metabolism [17]. Melatonin is also present in fungi and, with regard to nutrition, this may be relevant especially for yeast. In cultures freshly prepared from commercially available cubes of baker's yeast, μmolar concentrations of melatonin were measured, sometimes exceeding 40 μM [29,30].
It is not yet known whether food is the only external source of melatonin in mammals. The presence of melatonin in bacteria including Escherichia coli [31] may suggest a contribution by intestinal bacteria to the high amounts of the indoleamine found in the gut [cf. discussion in ref. [17]]. However, strictly anaerobic bacteria, which predominate in the colon, have not yet been investigated.
The gastrointestinal tract deserves particular attention, not only with regard to melatonin uptake, but, even more, as an extrapineal site of melatonin biosynthesis, where this molecule is present in amounts exceeding those found in the pineal gland by several-hundred-fold, and from where it can be released into the circulation in a post-prandial response, especially under the influence of high tryptophan levels [32-36]. Gastrointestinal melatonin is also released to the lumen and participates in enterohepatic cycling [37-39]. Therefore, nutrition is not only linked to melatonin by uptake, but also by the influences of other food constituents and digestive physiology on melatonin release.
With regard to nutrition, a decisive question is whether the amounts of melatonin present in the food can suffice for changing its level in the blood plasma. This was first indicated by findings of Hattori et al. [19], who observed rises in plasma melatonin after feeding plant material rich in this compound. However, this result allowed a different interpretation, because other substances including its precursor tryptophan might have elicited a post-prandial release of gastrointestinal melatonin. This argument was recently refuted, at least in chicken, because the removal of melatonin from feed caused decreases in plasma levels [40]. The result gave rise to a statement that melatonin may not only be regarded as a hormone and a tissue factor, but also, in a sense, as an antioxidant vitamin.
The redox properties of melatonin may be unfavorable for its preservation in the food. Being an easily oxidizable compound capable of directly detoxifying several free radicals and other oxidants, leads, in turn, to the consequence of non-enzymatic destruction. The experience with melatonin extraction from plant material lets us assume that only a certain fraction of the compound present in food will arrive in the gut and even less in the circulation. Nevertheless, it may be possible that melatonin metabolites, especially substituted kynuramines formed by oxidative pyrrole-ring cleavage, which also possess protective properties and sufficient amphiphilicity [41-43], and/or their derivatives are taken up from the food and will turn out to be beneficial.
Reactions of melatonin with oxidants
With regard to the presence of melatonin in food, in medicinal plants and to the use as a food additive, its antioxidant and other protective properties deserve attention. Since the discovery of melatonin oxidation by photocatalytic mechanisms involving free radicals [15,44,45], scavenging by this indoleamine has become a matter of particular interest. Melatonin was also shown to be oxidized by free radicals formed in the absence of light [46], and its capability of scavenging hydroxyl radicals at high rates [47-51] initiated numerous investigations on radical detoxification and antioxidative protection. Melatonin turned out to be considerably more efficient than the majority of its naturally occurring structural analogs [47,50-52], indicating that the substituents of the indole moiety strongly influenced reactivity and selectivity. Rate constants determined for the reaction with hydroxyl radicals were in the range between 1.2 × 1010 and 7.5 × 1010 M-1 s-1, depending on the methods applied [53-57]. Regardless of differences in the precision of determination, melatonin has been shown, independently by different groups, to be a remarkably good scavenger of this radical species. This property can be crucial for antagonizing oxidative damage under pharmacological and other in vitro conditions. To what extent this may contribute to physiological protection remains, however, a matter of debate.
Meanwhile, melatonin has been shown to react with many other oxidants, such as carbonate radicals [58-60], singlet oxygen [15,34,61-65], ozone [15,34], and several biologically occurring aromatic radicals, such as protoporphyrinyl and substituted anthranilyl radicals [15,59,61,62,66,67]. Reactions with other non-biological radicals were also described [15,34], among which the ABTS cation radical [ABTS = 2, 2'-azino-bis-(3-ethylbenzthiazoline-6-sulfonic acid)] merits special attention because of its analytical value. This extremely long-lived radical which is stable for many days provides a good example for single-electron donation by melatonin [52,68]. This conclusion was unambiguously confirmed by cyclic voltammetry [69]. Single-electron donation is important for several reasons. Free radicals can react with scavengers in different ways, either by abstraction of an electron, or a hydrogen atom, or by addition. In the case of melatonin, radical addition has been observed or predicted theoretically only for interactions with hydroxyl radicals [69-72] and nitric oxide [69,73-75]. Electron/hydrogen abstraction, however, is a common key step for interactions of melatonin with oxidizing free radicals of both high and low reactivity and, therefore, reflects melatonin's property as a broad spectrum antioxidant. Electron abstraction was also concluded to be a primary step of melatonin oxidation in a pseudoenzymatic reaction catalyzed by oxoferrylhemoglobin [76]. Single-electron transfer reactions are also believed to play a role in detoxification of resonance-stabilized free radicals, such as carbonate and aryl radicals, which are frequently underrated in their destructive potential because of their lower reactivity, compared to the hydroxyl radical. However, due to their longer life-time they can reach more distant sites than the extremely short-lived hydroxyl radical, which exists only for nanoseconds. The capability of melatonin of scavenging carbonate and certain aryl radicals may be of much higher significance and protective value than previously thought. Finally, according to a recently proposed model, single-electron exchange is thought to be the basis for interactions of melatonin with the mitochondrial respiratory chain [77,78] which is assumed to require only very small, quasi-catalytic amounts of melatonin and which would convey antioxidative cell protection by radical avoidance rather than detoxification of radicals already formed (see below).
Reactive nitrogen species represent another category of potentially destructive substances, which react with melatonin. Scavenging of nitric oxide by melatonin in a nitrosation reaction is well documented [9,79-81]. Whether this can be regarded as a detoxification reaction keeping NO from forming the more dangerous peroxynitrite is uncertain because nitrosomelatonin easily decomposes, thereby releasing NO [82], an experience also made with other NO adducts from respective scavengers including NO spin traps [83]. Scavenging of peroxynitrite has also been described [9,80,81,84], although it is sometimes difficult to distinguish betweeen direct reactions with peroxynitrite and with hydroxyl radicals formed by decomposition of peroxynitrous acid. What seems more important than direct scavenging of peroxynitrite is the interaction with products from the peroxynitrite-CO2 adduct (ONOOCO2-), namely, carbonate radicals (CO3•-) and •NO2 [79,85]. In the presence of bicarbonate/CO2, this pathway is favored and the primary interaction of melatonin is that with CO3•- [85], a conclusion in agreement with results from other studies on CO3•- scavenging [58-60]. The mixture of CO3•- and •NO2 represents the physiologically most efficient nitration mixture, because of the high availability of CO2 in biological material. It is worth noting that melatonin can, in fact, decrease 3-nitrotyrosine levels, as shown in guinea pig kidney [86].
Another highly interesting aspect of melatonin's antioxidant actions, which may be particularly important from the nutritional aspect, is its interactions with classic antioxidants. In both chemical and cell-free systems, melatonin was repeatedly shown to potentiate the effects of ascorbate, Trolox (a tocopherol analog), reduced glutathione, or NADH [50,68,69,87]. These findings, which can be clearly distinguished from additive effects, surprisingly indicate multiple interactions via redox-based regeneration of antioxidants transiently consumed. This may, in fact, be of practical importance, since melatonin was also shown to prevent decreases in hepatic ascorbate and α-tocopherol levels in vivo, under conditions of long-lasting experimental oxidative stress induced by a high cholesterol diet [88].
Metabolites of melatonin, a scavenger cascade, and melatonin as a prodrug
Reactions of melatonin with free radicals and other oxidants are not only a matter of the toxic reactants eliminated, but also of the products formed. It is highly important to distinguish between metabolites formed under physiological or near-physiological conditions from those produced in chemical systems designed for studying reactions with a single radical species in preparations as pure as possible. Disregard of this point has led to several misinterpretations in the past. We have repeatedly emphasized that studies using reaction systems which preferentially generate hydroxyl radicals mainly lead to hydroxylated adducts or their derivatives such as substituted indolinones, whereas biological material usually contains orders of magnitude more superoxide anions than hydroxyl radicals. Therefore, an entirely different product spectrum is obtained as soon as hydroxyl radicals, or other electron-abstracting radicals, act in the presence of an excess of superoxide anions [60,89]. Radicals derived from melatonin by interaction with a first, reaction-initiating radical likely combine with superoxide anions so that the radical reaction chain is readily terminated [15,49]. The product formed by oxidative pyrrole-ring cleavage is a substituted kynuramine, N1-acetyl-N2-formyl-5-methoxykynuramine (AFMK; Fig. 1). We have investigated numerous reaction systems and in all those containing sufficient quantities of superoxide anions, AFMK was by far the most abundant product [44,46,58-60,66,89]. Interestingly, a profound and sursprising difference exists between melatonin and other structurally related indoleamines. While substituted kynuramines represent only a limited or small fraction of oxidation products from other indolic compounds, AFMK usually greatly exceeds the total of other substances formed. This indicates a significant contribution not only of the 5-methoxy residue, but also of the N-acetylated side chain to the oxidation chemistry of melatonin, a conclusion corroborated by various scavenging assays and chemiluminescence associated with pyrrole-ring cleavage [52]. Moreover, AFMK was the only melatonin metabolite detected in culture media of various aquatic organisms, unicells and small metazoans, whereas several additional products were found in axenic media incubated for extended periods of time [90]. AFMK formation seems to be a favored pathway of melatonin degradation in these species.
Figure 1 The kynuric pathway of melatonin metabolism.
These findings do not represent a peculiarity of non-vertebrates, but rather seem to reflect the non-hepatic melatonin catabolism in vertebrates. Contrary to statements in the earlier literature claiming that almost all melatonin is metabolized in the liver to 6-hydroxymelatonin followed by conjugation and excretion, recent estimations attribute about 30 percent of overall melatonin degradation to pyrrole-ring cleavage [91]. The rate of AFMK formation may be considerably higher in certain tissues, since extrahepatic P450 monooxygenase activities are frequently too low for a high turnover via 6-hydroxylation. The high amounts of gastrointestinal melatonin (see above), as far as they are not released unmetabolized, have to enter a pathway different from monooxygenation. AFMK formation is highly likely.
The significance of pyrrole-ring cleavage in oxidative metabolism of tissue melatonin is particularly illustrated in the central nervous system, where a secondary product, N1-acetyl-5-methoxykynuramine (AMK) derived from AFMK by deformylation, was identified as a main metabolite [92]. When melatonin was injected into the cisterna magna, about 35 percent was recovered as AMK. Under the conditions used, AFMK and AMK were the only products formed from melatonin in the brain and no 6-hydroxymelatonin was detected. In this case, the high turnover in the kynuric pathway of melatonin catabolism is the more remarkable as it cannot be explained on the basis of the enzymes capable of catalyzing the formation of AFMK: (i) indoleamine 2, 3-dioxygenase which uses tryptophan as the main substrate, exhibits sufficiently high activities only after inflammatory stimulation of the microglia [93-95]; (ii) myeloperoxidase, which can also catalyze pyrrole-ring cleavage of melatonin [91,96,97], is again associated with activated phagocytes. To assume free radical reactions as the main cause of kynuric melatonin degradation in the brain is, therefore, highly suggestive. Non-enzymatic AFMK formation in other tissues will be a matter for future research.
It is a remarkable fact that AFMK is formed by many different mechanisms [summarized in refs. [15,41,59,66,89]]. Apart from the enzymes mentioned, pseudoenzymatic catalysis by oxyferrylhemoglobin or by hemin, interactions with free radicals, e.g., combinations of •OH and O2•-, or CO3•- and O2•-, or organic cation radicals and O2•-, oxidation by singlet oxygen, by ozone, or by O2 under photoexcitation of melatonin all lead to AFMK. Even another product formed from melatonin by interactions with free radicals, cyclic 3-hydroxymelatonin [70], can be further metabolized by free radicals to AFMK [68]. All these findings indicate that AFMK is a central metabolite of melatonin oxidation especially in non-hepatic tissues.
As already mentioned, AFMK is easily deformylated to AMK. To date two enzymes capable of catalyzing this reaction have been identified, arylamine formamidase and hemoperoxidase [49,89,98]. The two methoxylated kynuramines, AFMK and AMK, are of particular interest because of their own radical-scavenging and protective properties. In any case, kynuramines, a separate class of biogenic amines, exhibit various biological activities [99], which are, however, rarely investigated. With regard to antioxidative protection, AFMK was shown to reduce 8-hydroxy-2-deoxyguanosine formation [42] and lipid peroxidation, and to rescue hippocampal neurons from oxidotoxic cell death [41]. Although AFMK interacts, not surprisingly, with the highly reactive hydroxyl radicals, it is otherwise relatively inert towards radicals of lower or intermediate reactivity [43,89]. This is convincingly explained by its preference for two-electron transfer reactions as demonstrated by cyclic voltammetry [41].
The deformylated product AMK, easily formed from AFMK [92], appears to be a highly interesting substance, for several reasons: first, it is a radical scavenger of considerably higher reactivity than AFMK because it easily undergoes single-electron transfer reactions [43,89,100] and, second, it acts as a cyclooxygenase (COX) inhibitor that is much more potent than acetylsalicylic acid [101] and has relative specificity for COX-2 (B Poeggeler, pers. commun.). Moreover, AMK was recently shown to downregulate COX-2 expression in macrophages [102]. AMK might, therefore, contribute to the attenuation of oxidative stress both directly and indirectly by interference with inflammatory responses. A third, mitochondrial effect will be discussed below. Unfortunately, the precise tissue levels of AMK are still unknown, partially because of a lack of specific assays, partially due to its high reactivity which readily leads to other products. Since AMK can be recovered from the urine after a melatonin load [92], sufficient amounts may be present in the tissues, at least after administration of pharmacological doses. Therefore, melatonin seems to act not only directly, but, additionally, as a prodrug of AMK.
It is a remarkable fact that the kynuric pathway of melatonin metabolism includes a series of radical scavengers, which may be regarded as a scavenger cascade [68], with a possible sequence of melatonin → cyclic 3-hydroxymelatonin → AFMK → AMK, where melatonin can be alternately converted to AFMK directly. From melatonin to AFMK, up to 4 free radicals can be consumed [68]; recent determinations [Rosen J, Hardeland R, unpubl. data] have shown that even higher numbers of free radicals can be eliminated, and other, previously unknown products are being characterized. The potent scavenger AMK consumes further radicals in primary and secondary reactions. Interestingly, AMK not only interacts with reactive oxygen but also with reactive nitrogen species and several products have been structurally characterized in Göttingen [[103]; manuscript in preparation]. Neither the end of the kynuric pathway of melatonin nor that of the scavenger cascade is in sight.
Multiple levels of antioxidative protection by melatonin
Antioxidative protection by melatonin is not just a matter of direct radical scavenging (Fig. 2), as becomes immediately evident from stoichiometry. Although tissue levels of melatonin can be considerably higher than those in the circulation, the quantities of free radicals generated in its metabolism would still be too high for the available amounts of the indoleamine. Our understanding is that direct scavenging by physiological concentrations of melatonin by a non-enzymatic contribution to the kynuric pathway and the subsequent actions of the metabolites formed becomes important. Signaling effects of melatonin, however, are always possible at physiological levels.
Figure 2 Overview of the pleiotropic actions of melatonin and some of its metabolites in antioxidative protection.
Melatonin upregulates several antioxidant enzymes. Most frequently, this has been demonstrated for glutathione peroxidase [7,77,88,104-118] and sometimes glutathione reductase [7,108,112,119], presumably indirectly via GSSG. In some tissues Cu, Zn- and/or Mn-superoxide dismutases [7,108-112,117-123] and, rarely, catalase [112,118,123,124] are upregulated. Stimulation of glutathione peroxidase seems to be widely distributed among tissues and is observed quite regularly in both mammalian and avian brain; upregulations in other organs were more variable. The action of melatonin on glutathione metabolism seems to exceed the effects mentioned. Stimulation of glucose-6-phosphate dehydrogenase [108] and γ-glutamylcysteine synthase [10,112] indirectly supports the action of glutathione peroxidase by providing reducing equivalents (NADPH) for the action of glutathione reductase and by increasing the rate of glutathione synthesis, respectively.
Contrary its effect on the enzymes of glutathione metabolism, the effect of melatonin on superoxide dismutase subforms and catalase strongly depends on organs and species. Stimulation was observed in some tissues, but not in others; in some cases, even decreases were reported. This may not only be a matter of differences in responsiveness of cell types. The complexity in the regulation of the respective enzymes has to be considered. Frequently, they exhibit compensatory rises in response to oxidative stress. When melatonin is counteracting experimentally induced stress, the result may be a normalization of enzyme activity, i.e., lower values, compared to animals treated with oxidotoxins, rather than inductions. Such normalizations were, in fact, described [114,125]. However, in cases of stronger oxidative stress, active centers of enzymes may be destroyed by the free radicals generated and normalization of enzyme activities by melatonin administration appears as an increase [110,124,125].
An additional aspect of melatonin's actions on antioxidant enzymes deserves future attention: In two neuronal cell lines, physiological concentrations of melatonin not only induced glutathione peroxidase and superoxide dismutases at the mRNA level, but concomitantly increased the life-time of these mRNAs [117].
Melatonin also contributes to the avoidance of radical formation in several independent ways. It downregulates prooxidant enzymes, in particular 5- and 12-lipoxygenases [112,126-128] and NO synthases [9,34,77,108,112,129-134]. The widely observed attenuation of NO formation is particularly important in terms of limiting rise in the strongly prooxidant metabolite peroxynitrite and of the free radicals derived from this compound, namely, •NO2, carbonate (CO3•-) and hydroxyl (•OH) radicals. Suppressions of both lipoxygenase and NO synthase may additionally set limits to inflammatory responses, although the immunomodulatory actions of melatonin are certainly more complex and may involve additional effects of melatonin and AMK, too.
Another widely unexplored but potentially important signaling effect of melatonin in antioxidative protection concerns quinone reductase 2 [77,135,136]. This enzyme, which is implicated in the detoxification of potentially prooxidant quinones, binds melatonin at upper physiological concentrations, so that it had originally been presumed to represent a melatonin receptor. Although its precise function under the influence of melatonin is not yet fully understood, the relationship to the indoleamine may become of future interest from the standpoint of nutrition, since quinones are taken up with food, especially, vegetables.
Although less relevant from a nutritional point of view, melatonin also contributes indirectly to radical avoidance, e.g., by its antiexcitatory effects in the central nervous system, and as an endogenous regulator molecule controlling rhythmic time structures. This last action may be particularly important for well-timed alimentary melatonin supplementation in the elderly, who exhibit a strongly reduced amplitude in the circadian melatonin rhythm. The significance of appropriate timing for maintaining low levels of oxidative damage has been overlooked for quite some time. However, temporal perturbations as occurring in short-period or arrhythmic circadian clock mutants lead to enhanced oxidative damage, effects observed in organisms as different as Drosophila and the Syrian Hamster [77,137,138].
In the last few years, mitochondrial effects of melatonin have been discovered which may turn out to be even more important than the protective actions described above. Mitochondria are the main source of free radicals in the majority of animal cells and are implicated in aging processes. The importance of mitochondrial diseases is increasingly perceived. Mitochondria play a key role in apoptosis. Notably, several of the mitochondrial effects of melatonin were obtained at low pharmacological doses in drinking water [116,139,140] or even at near-physiological concentrations down to 1 nM [141].
Several studies of mitochondrial effects revealed attenuation of mitochondrial lipid peroxidation, prevention of oxidative protein and DNA modifications, preservation of ultrastructure, resistance against toxins etc., findings which were widely in line with previous concepts of protection [10,113,142-146]. Moreover, melatonin was shown to affect redox-active compounds in mitochondria, in particular, to decrease NO [143,147] and to restore normal levels of reduced glutathione [113,144] and coenzyme Q10 [148].
More importantly, beyond these rather conventional findings, with few exceptions, melatonin was found to increase mitochondrial respiration and ATP synthesis, in conjunction with rises in complex I and IV activities [112,141-143,146,147,149,150]. Complex I and IV activities were also found to be increased by melatonin in hepatic mitochondria of senescence-accelerated mice [116,140,151]. Moreover, melatonin was found to enhance gene expression of complex IV components [147].
The improvements of ATP formation and O2 consumption are presumably not decisive for protection, but can serve as good indicators for the reduction of electron leakage from the respiratory chain. Electron transfer to molecular oxygen at the matrix side, largely at iron-sulfur cluster N2 of complex I [152], is a major source of free radicals. This process also diminishes electron flux rates and, therefore, the ATP-generating proton potential. Processes affecting the mitochondrial membrane potential such as calcium overload, either due to overexcitation, to protein misfolding or to damage by free radicals, are antagonized by melatonin. In cardiomyocytes, astrocytes and striatal neurons, melatonin prevented calcium overload [153,154], counteracted the collapse of the mitochondrial membrane potential induced by H2O2 [153], doxorubicin [155] or oxygen/glucose deprivation [154], and also inhibited the opening of the mitochondrial permeability transition pore (mtPTP), thereby rescuing cells from apoptosis. In addition to the antioxidant actions, melatonin directly diminished mtPTP currents, with an IC50 of 0.8 μM [154], a concentration which would require mitochondrial accumulation of melatonin, something which is possible again due to the amphiphilicity of melatonin.
The effects of melatonin on the respiratory chain open new perspectives for diminishing radical formation, instead of seeking only antioxidant effects for the elimination of radicals already formed. We have proposed a model of radical avoidance (Fig. 3) in which electron leakage is reduced by single-electron exchange reactions betwen melatonin and components of the electron transport chain [77,78]. In fact, mitochondrial H2O2 formation was found to be reduced by melatonin [156]. The basic idea of the model is that of a cycle of electron donation to the respiratory chain, eventually to cytochrome c [78], followed by reduction of the formed melatonyl cation radical by electron transfer from N2 of complex I. The cation radical is assumed to act as alternate electron acceptor competing with molecular oxygen, thereby decreasing the rate of O2•- formation. In addition to the electrons being largely recycled, most of the melatonin is also. Therefore, such a mechanism would only require very low, quasi-catalytic amounts of melatonin, in accordance with the effects demonstrated with nanomolar concentrations. Because the recycled electrons are not lost for the respiratory chain, this would also lead to improvements in complex IV activity, oxygen consumption and ATP production. Alternately, the melatonin metabolite AMK, which is also highly reactive and can undergo single-electron transfer reactions [43], may act in the same way. The prediction of our model of mitochondrial protection by AMK was confirmed by other investigators [147]: AMK was shown to exert effects on electron flux through the respiratory chain and ATP synthesis very similar to those observed with melatonin.
Figure 3 A model of mitochondrial radical avoidance and support of electron flux by melatonin and its metabolite AMK. The potent electron donors melatonin and AMK are thought to feed electrons into the respiratory chain, thereby forming resonance-stabilized cation radicals which may efficiently compete with molecular oxygen for electrons leaking from iron-sulfur cluster N2 or from ubisemiquinone. The competition reduces superoxide anion formation and, thereby, the generation of secondary radicals; at the same time, electrons re-fed to the electron transport chain contribute to the maintenance of the proton potential and, thus, to ATP synthesis. The model is partially hypothetical, but might explain observations of reductions in electron leakage and oxidant formation as well as an enhancement of ATP formation.
A highly attractive aspect of mitochondrial protection results from the small quantities required: experimentally induced mitochondrial damage in rat fetuses was even prevented by maternally administered melatonin [146]. The mechanism as outlined, requiring only low amounts of melatonin or its metabolite AMK, would make these compounds even more interesting from a nutritional point of view. The amounts present in selected food, such as some vegetables, but even more nuts and cereals, could suffice for maintaining tissue levels of the indoleamine capable of safeguarding mitochondrial function, particularly in elderly persons whose nocturnal melatonin maxima in pineal gland and circulation have substantially declined with age. Transient moderate rises in blood melatonin during the day resulting from direct uptake or postprandrial release from the gastrointestinal tract should not be regarded as a problem in terms of circadian timing. The circadian system responds to melatonin according to a phase response curve [157,158]: the so-called silent zone, during which no substantial phase shifts are induced, extends throughout the largest part of the day.
Safety of melatonin
Can all these findings on antioxidant and radical-avoiding actions of melatonin justify its intake as a food additive or as a medication? The idea of substitution therapy may seem especially attractive for the elderly who have more or less lost the nocturnal peak of circulating melatonin. Nevertheless, the use as a food additive is still a matter of controversy. The argument for a naturally occurring compound, which is a normal food constituent, cannot suffice alone, since commerical preparations would always lead to at least transient pharmacological concentrations in the blood, and the immunomodulatory actions of melatonin may not be desired in every case. Therefore, experience will have to answer the question of its usefulness. Without any doubt, melatonin is remarkably well tolerated. Of course, one can find in any large statistical sample of melatonin users some individuals who complain about side effects, scientifically understandable or not. In a currently running study on ALS, patients receiving daily very high doses of melatonin (30 or even 60 mg per day), we did not see any harmful side effects [14] and have not to date. In patients with rheumatoid arthritis, some symptoms were suspected to be associated with immunomodulatory actions of melatonin [159], so that caution is due in this group of individuals. More research will be required on melatonin in different diseases and disorders, but there is no good reason to assume that melatonin, at moderate or even at high doses, is dangerous to a healthy person or to patients with types of oxidative stress phenomena not caused by (auto-)-immune responses. One might also suspect that melatonin could exert unfavorable effects by increasing the blood pressure, due to downregulation of NO synthase and NO scavenging by the indoleamine itself or by AMK. Melatonin was tested in clinical trials on hypertension and was reported to decrease blood pressure in one study [160], but to interfere with nifedipine [161], whereas a combination of lacidipine with melatonin was recommended in another investigation [162]. Therefore, interaction with other medication has to be considered.
Problems of dosage and side effects may also arise from impurities in the melatonin preparations sold over the counter. Contaminants have repeatedly been detected in such material, including our own experience of that kind. As long as the contaminant is only AFMK, this may be less serious, but one should be aware that the pharmacology of kynuramines is only partially known. Moreover, manufacturers must consider that an easily oxidizable compound like melatonin can undergo reactions under air exposure. On large surfaces, such as silica gels, we see this every day in the laboratory.
Another important aspect for the use of melatonin as a food additive is timing. As soon as the substance is given as a pill or as a preparation from a medicinal plant causing relatively high pharmacological blood levels, the situation is entirely different from the uptake with normal food or from the postprandial gastrointestinal release. Since circulating melatonin peaks at night, pharmaceutical preparations should be strictly given at the same time of day in the evening. The usual recommendation „at bed time" may be insufficient since this could mean in practice different hours of the day. Here, one has to consider the chronobiological functions of melatonin. When given during the day, a high dose of melatonin would cause mild narcotic effects, drowsiness etc. and the practice is not recommended for this reason. It would not shift the circadian oscillator much, because of the silent zone of the phase response curve for melatonin, in which phase shifts are negligibly small. This is the same reason that a postprandial release of gastrointestinal melatonin does not shift the circadian oscillator. Advance shifts of the endogenous clock by melatonin are much larger at late afternoon and early night [157,158]. Therefore, melatonin should be given relatively precisely at the same hour, to avoid phase shifts differing in extent and pushing of the circadian oscillator back and forth. As mentioned above, pertubations of the internal time structure can also cause oxidative stress [77].
Conclusion
In terms of nutrition, melatonin is interesting both as a natural constituent of food, and as a food additive. Its use for the latter purpose can be recommended only with some caution, given the present state of our knowledge, although the risks by melatonin appear remarkably low, compared to other medications and food additives. Melatonin's antioxidant capacity is based not only on direct radical detoxification, but comprises manifold effects. Some of the most promising areas, modulation of mitochondrial metabolism by melatonin and actions of its kynuric metabolites, deserve particular attention in the future and may change our view of the value of these compounds profoundly.
List of abbreviations
ABTS: 2, 2'-azino-bis-(3-ethylbenzthiazoline-6-sulfonic acid)
AFMK: N1-acetyl-N2-formyl-5-methoxykynuramine
ALS: amyotrophic lateral sclerosis
AMK: N1-acetyl-5-methoxykynuramine
c3OHM: cyclic 3-hydroxymelatonin
COX-2: cyclooxygenase 2
GSSG: oxidized glutathione
Competing interests
Authors declare that they have no competing interests concerning the use of melatonin or melatonin-containing preparations as a food additive.
Authors' contributions
This review was initiated by SRP-P; a first version by RH was jointly revised.
==== Refs
Arendt J Deacon S Treatment of circadian rhythm disorders – melatonin Chronobiol Int 1997 14 185 204 9095378
Arendt J Jet-lag Lancet 1998 351 293 294 9457125 10.1016/S0140-6736(05)78236-2
Herxheimer A Waterhouse J The prevention and treatment of jet lag Br Med J 2003 326 296 297 12574022
Oxenkrug GF Requintina PJ Melatonin and jet lag syndrome: experimental model and clinical applications CNS Spectr 2003 8 139 148 12612500
Reiter RJ Oxidative damage in the central nervous system: Protection by melatonin Prog Neurobiol 1998 56 359 384 9770244 10.1016/S0301-0082(98)00052-5
Reiter RJ Tan D-X Cabrera J D'Arpa D Sainz RM Mayo JC Ramos S The oxidation/antioxidant network: role of melatonin Biol Signals Recept 1999 8 56 63 10085463 10.1159/000014569
Reiter RJ Tan D-X Osuna C Gitto E Actions of melatonin in the reduction of oxidative stress. A review J Biomed Sci 2000 7 444 458 11060493 10.1159/000025480
Reiter RJ Tan D-X Burkhardt S Manchester LC Melatonin in plants Nutr Rev 2001 59 286 290 11570431
Reiter RJ Tan D-X Burkhardt S Reactive oxygen and nitrogen species and cellular and organismal decline: amelioration with melatonin Mech Ageing Dev 2002 123 1007 1019 12044950 10.1016/S0047-6374(01)00384-0
Reiter RJ Tan D-X Mayo JC Sainz RM Leon J Czarnocki Z Melatonin as an antioxidant: biochemical mechanisms and pathophysiological implications in humans Acta Biochim Pol 2003 50 1129 1146 14740000
Pappolla MA Chyan Y-J Poeggeler B Frangione B Wilson G Ghiso J Reiter RJ An assessment of the antioxidant and antiamyloidogenic properties of melatonin: implications for Alzheimer's disease J Neural Transm 2000 107 203 231 10847561 10.1007/s007020050018
Cardinali DP Brusco LI Liberczuk C Furio AM The use of melatonin in Alzheimer's disease Neuroendocrinol Lett 2002 23 20 23 12019347
Singer C Tractenberg RE Kaye J Schafer K Gamst A Grundman M Thomas R Thal LJ A multicenter, placebo-controlled trial of melatonin for sleep disturbance in Alzheimer's disease Sleep 2003 26 893 901 14655926
Jacob S Poeggeler B Weishaupt JH Sirén A-L Hardeland R Bähr M Ehrenreich H Melatonin as a candidate compound for neuroprotection in amyotrophic lateral sclerosis (ALS): High tolerability of daily oral melatonin administration in ALS patients J Pineal Res 2002 33 186 187 12220335 10.1034/j.1600-079X.2002.02943.x
Hardeland R Fuhrberg B Ubiquitous melatonin – Presence and effects in unicells, plants and animals Trends Comp Biochem Physiol 1996 2 25 45
Hardeland R Melatonin and 5-methoxytryptamine in non-metazoans Reprod Nutr Dev 1999 39 399 408 10420441
Hardeland R Poeggeler B Non-vertebrate melatonin J Pineal Res 2003 34 233 241 12662344 10.1034/j.1600-079X.2003.02941.x
Dubbels R Reiter RJ Klenke E Goebel A Schnakenberg E Ehlers C Schiwara HP Schloot W Melatonin in edible plants identified by radioimmunoasay and by high performance liquid-chromatography-mass spectrometry J Pineal Res 1995 18 28 31 7776176
Hattori A Migitaka H Iigo M Itoh M Yamamoto K Ohtani-Kaneko R Hara M Suzuki T Reiter RJ Identification of melatonin in plants and its effects on plasma melatonin levels and binding to melatonin receptors in vertebrates Biochem Mol Biol Int 1995 35 627 634 7773197
Balzer I Hardeland R Melatonin in algae and higher plants – Possible new roles as a phytohormone and antioxidant Bot Acta 1996 109 180 183
Van Tassel DL O'Neill SD Putative regulatory molecules in plants: evaluating melatonin J Pineal Res 2001 31 1 7 11484999 10.1034/j.1600-079X.2001.310101.x
Van Tassel DL Roberts N Lewy A O'Neill SD Melatonin in plant organs J Pineal Res 2001 31 8 15 11485009 10.1034/j.1600-079X.2001.310102.x
Burkhardt S Tan DX Manchester LC Hardeland R Reiter RJ Detection and quantification of the antioxidant melatonin in Montmorency and Balaton tart cherries (Prunus cerasus) J Agric Food Chem 2001 49 4898 4902 11600041 10.1021/jf010321+
Reiter RJ Tan D-X Melatonin: an antioxidant in edible plants Ann NY Acad Sci 2002 957 341 344 12074994
Conti A Tettamanti C Singaravel M Haldar C Pandi-Perumal RS Maestroni GJM Haldar C, Singaravel M, Maitra SK Melatonin: An ubiquitous and evolutionary hormone Treatise on Pineal Gland and Melatonin 2002 Enfield, NH: Science Publishers 105 143
Poeggeler B Hardeland R Detection and quantification of melatonin in a dinoflagellate, Gonyaulax polyedra. Solutions to the problem of methoxyindole destruction in non-vertebrate material J Pineal Res 1994 17 1 10 7853137
Murch SJ Simmons CB Saxena PK Melatonin in fever few and other medicinal plants Lancet 1997 350 1598 1599 9393344 10.1016/S0140-6736(05)64014-7
Manchester LC Tan D-X Reiter RJ Park W Monis K Qi W High levels of melatonin in the seeds of edible plants. Possible function in germ tissue protection Life Sci 2000 67 3023 3029 11125839 10.1016/S0024-3205(00)00896-1
Sprenger J Hardeland R Fuhrberg B Hardeland R Methoxyindoles in yeast: Variability and effects of growth conditions Biological Rhythms and Antioxidative Protection 1997 Göttingen: Cuvillier 116 118
Sprenger J Hardeland R Fuhrberg B Han S-Z Melatonin and other 5-methoxylated indoles in yeast: Presence in high concentrations and dependence on tryptophan availability Cytologia 1999 64 209 213
Balzer I Höcker B Kapp H Bartolomaeus B Vanden Driessche T, Guisset J-L, Petiau-de Vries GM Occurrence and comparative physiology of melatonin in evolutionary diverse organisms The Redox State and Circadian Rhythms 2000 Boston – London: Kluwer 95 119
Huether G Poeggeler B Reimer A George A Effect of tryptophan administration on circulating melatonin levels in chicks and rats: evidence for stimulation of melatonin synthesis and release in the gastrointestinal tract Life Sci 1992 51 945 53 1518369 10.1016/0024-3205(92)90402-B
Huether G The contribution of extrapineal sites of melatonin synthesis to circulating melatonin levels in higher vertebrates Experientia 1993 49 665 670 8359272 10.1007/BF01923948
Hardeland R Altmeyer P, Hoffmann K, Stücker M Melatonin: Multiple functions in signaling and protection Skin Cancer and UV Radiation 1997 Berlin – Heidelberg: Springer 186 198
Bubenik GA Localization, physiological significance and possible clinical implication of gastrointestinal melatonin Biol Signals Recept 2001 10 350 366 11721091 10.1159/000046903
Bubenik GA Gastrointestinal melatonin: localization, function, and clinical relevance Dig Dis Sci 2002 47 2336 2348 12395907 10.1023/A:1020107915919
Messner M Hardeland R Rodenbeck A Huether G Tissue retention and subcellular distribution of continuously infused melatonin in rats under near physiological conditions J Pineal Res 1998 25 251 259 9885995
Bubenik GA Hacker RR Brown GM Bartos L Melatonin concentrations in the luminal fluid, mucosa and muscularis of the bovine and porcine gastrointestinal tract J Pineal Res 1999 26 56 63 10102761
Tan D-X Manchester LC Reiter RJ Qi W Hanes MA Farley NJ High physiological levels of melatonin in the bile of mammals Life Sci 1999 65 2523 2529 10622237 10.1016/S0024-3205(99)00519-6
Tan D-X Manchester LC Hardeland R Lopez-Burillo S Mayo JC Sainz RM Reiter RJ Melatonin – a hormone, a tissue factor, an autocoid, a paracoid, and an antioxidant vitamin J Pineal Res 2003 34 75 78 12485375 10.1034/j.1600-079X.2003.02111.x
Tan D-X Manchester LC Burkhardt S Sainz RM Mayo JC Kohen R Shohami E Huo Y-S Hardeland R Reiter RJ N 1-Acetyl-N2-formyl-5-methoxykynuramine, a biogenic amine and melatonin metabolite, functions as a potent antioxidant FASEB J 2001 15 2294 2296 11511530
Burkhardt S Reiter RJ Tan D-X Hardeland R Cabrera J Karbownik M DNA oxidatively damaged by chromium(III) and H2O2 is protected by melatonin, N 1-acetyl-N2-formyl-5-methoxykynuramine, resveratrol and uric acid Int J Biochem Cell Biol 2001 33 775 783 11404181 10.1016/S1357-2725(01)00052-8
Ressmeyer A-R Mayo JC Zelosko V Sáinz RM Tan D-X Poeggeler B Antolín I Zsizsik BK Reiter RJ Hardeland R Antioxidant properties of the melatonin metabolite N 1-acetyl-5-methoxykynuramine (AMK): scavenging of free radicals and prevention of protein destruction Redox Rep 2003 8 205 213 14599344 10.1179/135100003225002709
Hardeland R Poeggeler B Balzer B Behrmann G Common basis of photoperiodism in phylogenetically distant organisms and its possible origins J Interdiscipl Cycle Res 1991 22 122 123
Hardeland R Poeggeler B Balzer I Behrmann G A hypothesis on the evolutionary origins of photoperiodism based on circadian rhythmicity of melatonin in phylogenetically distant organisms Chronobiology & Chronomedicine 1993 Gutenbrunner C, Hildebrandt G, Moog R. Frankfurt: Lang 113 120
Hardeland R Fuhrberg B Behrmann G Balzer I Sleep-latency reducing pineal hormone melatonin as a scavenger of free radicals: hemin-catalysed formation of N1-acetyl-N2-formyl-5-methoxykynuramine Sleep Res 1993 22 621
Tan D-X Chen L-D Poeggeler B Manchester LC Reiter RJ Melatonin: a potent, endogenous hydroxyl radical scavenger Endocr J 1993 1 57 60
Reiter RJ Poeggeler B Tan D-X Chen L-D Manchester LC Guerrero JM Antioxidant capacity of melatonin: A novel action not requiring a receptor Neuroendocrinol Lett 1993 15 103 116
Hardeland R Reiter RJ Poeggeler B Tan D-X The significance of the metabolism of the neurohormone melatonin: antioxidative protection and formation of bioactive substances Neurosci Biobehav Rev 1993 17 347 357 8272286
Poeggeler B Reiter RJ Hardeland R Sewerynek E Melchiorri D Barlow-Walden LR Melatonin, a mediator of electron transfer and repair reactions, acts synergistically with the chain-breaking antioxidants ascorbate, trolox and glutathione Neuroendocrinol Lett 1995 17 87 92
Poeggeler B Reiter RJ Hardeland R Tan D-X Barlow-Walden LR Melatonin and structurally-related, endogenous indoles act as potent electron donors and radical scavengers in vitro Redox Rep 1996 2 179 184
Poeggeler B Thuermann S Dose A Schoenke M Burkhardt S Hardeland R Melatonin's unique radical scavenging properties – Roles of its functional substituents as revealed by a comparison with its structural analogs J Pineal Res 2002 33 20 30 12121482 10.1034/j.1600-079X.2002.01873.x
Poeggeler B Saarela S Reiter RJ Tan D-X Chen L-D Manchester LC Barlow-Walden LR Melatonin – a highly potent endogenous radical scavenger and electron donor: new aspects of the oxidation chemistry of this indole accessed in vitro Ann NY Acad Sci 1994 738 419 420 7832450
Matuszak Z Reszka KJ Chignell CF Reaction of melatonin and related indoles with hydroxyl radicals: EPR and spin trapping investigations Free Radic Biol Med 1997 23 367 372 9214572 10.1016/S0891-5849(96)00614-4
Stasica P Ulanski P Rosiak JM Reactions of melatonin with radicals in deoxygenated aqueous solution J Radioanal Nucl Chem 1998 232 107 113
Chyan Y-J Poeggeler B Omar RA Chain DG Frangione B Ghiso J Pappolla MA Potent neuroprotective properties against the Alzheimer β-amyloid by an endogenous melatonin-related indole structure, indole-3-propionic acid J Biol Chem 1999 274 21937 21942 10419516 10.1074/jbc.274.31.21937
Mahal HS Sharma HS Mukherjee T Antioxidant properties of melatonin: a pulse radiolysis study Free Radic Biol Med 1999 26 557 565 10218644 10.1016/S0891-5849(98)00226-3
Hardeland R Niebergall R Schoenke M Poeggeler B Hardeland R Carbonate radicals as initiators of melatonin oxidation: Chemiluminescence and formation of oxidation products Actions and Redox Properties of Melatonin and Other Aromatic Amino Acid Metabolites 2001 Göttingen: Cuvillier 49 55
Hardeland R Poeggeler B Burkhardt S Zelosko V Niebergall R Tomás-Zapico C Thuermann S Schoenke M Dose A Gruetzner T Coto-Montes A Haldar C, Singh SS Oxidation chemistry of melatonin: New aspects of radical reactions Neuroendocrine System and Pineal Gland with Special Reference to Lifestock 2003 Varanasi: Banaras Hindu University 27 35
Hardeland R Poeggeler B Niebergall R Zelosko V Oxidation of melatonin by carbonate radicals and chemiluminescence emitted during pyrrole ring cleavage J Pineal Res 2003 34 17 25 12485367 10.1034/j.1600-079X.2003.02941.x
Uría H Fuhrberg B Poeggeler B Hardeland R Menéndez-Peláez A Hardeland R Photooxidation of melatonin Cell Biological Problems in Chronobiology 1994 Göttingen: University of Göttingen 89 99
Hardeland R Balzer I Poeggeler B Fuhrberg B Uría H Behrmann G Wolf R Meyer TJ Reiter RJ On the primary functions of melatonin in evolution: Mediation of photoperiodic signals in a unicell, photooxidation and scavenging of free radicals J Pineal Res 1995 18 104 111 7629689
Cagnoli CM Atabay C Kharlamova E Manev H Melatonin protects neurons from singlet oxygen-induced apoptosis J Pineal Res 1995 18 222 226 8531053
Matuszak Z Bilska MA Reszka KJ Chignell CF Bilski P Interaction of singlet molecular oxygen with melatonin and related indoles Photochem Photobiol 2003 78 449 455 14653575 10.1562/0031-8655(2003)078<0449:IOSMOW>2.0.CO;2
de Almeida EA Martinez GR Klitzke CF de Medeiros MHG Di Mascio P Oxidation of melatonin by singlet molecular oxygen (O2(1Δg)) produces N 1-acetyl-N2-formyl-5-methoxykynurenine J Pineal Res 2003 35 131 137 Remark: according to formula and mechanism, authors obviously mean N1-acetyl-N2-formyl-5-methoxykynuramine = AFMK, not the corresponding kynurenine. 12887657 10.1034/j.1600-079X.2003.00066.x
Burkhardt S Poeggeler B Tan D-X Rosner C Gruetzner T Nitzki F Schoenke M Thuermann S Reiter RJ Hardeland R Hardeland R Oxidation products formed from melatonin in various radical-generating systems Actions and Redox Properties of Melatonin and Other Aromatic Amino Acid Metabolites 2001 Göttingen: Cuvillier 9 22
Tomás-Zapico C Hardeland R Poeggeler B Coto-Montes A Hardeland R 3-Hydroxyanthranilic acid and 3-hydroxykynurenine as photooxidants: Catalysis of N 1-acetyl-N2-formyl-5-methoxykynuramine formation from melatonin Actions and Redox Properties of Melatonin and Other Aromatic Amino Acid Metabolites 2001 Göttingen: Cuvillier 136 141
Tan D-X Hardeland R Manchester LC Poeggeler B Lopez-Burillo S Mayo JC Sainz RM Reiter RJ Mechanistic and comparative studies of melatonin and classic antioxidants in terms of their interactions with the ABTS cation radical J Pineal Res 2003 34 249 259 12662346 10.1034/j.1600-079X.2003.02111.x
Tan D-X Reiter RJ Manchester LC Yan MT El-Sawi M Sainz RM Mayo JC Kohen R Allegra M Hardeland R Chemical and physical properties and potential mechanisms: melatonin as a broad spectrum antioxidant and free radical scavenger Curr Top Med Chem 2002 2 181 197 11899100 10.2174/1568026023394443
Tan D-X Manchester LC Reiter RJ Plummer BF Hardies LJ Weintraub ST Vijayalaxmi Shepherd AMM A novel melatonin metabolite, cyclic 3-hydroxymelatonin: A biomarker of in vivo hydroxyl radical generation Biochem Biophys Res Commun 1998 253 614 620 9918777 10.1006/bbrc.1998.9826
Stasica P Paneth P Rosiak JM Hydroxyl radical reaction with melatonin molecule: a computational study J Pineal Res 2000 29 125 127 10981826 10.1034/j.1600-079X.2000.290209.x
Agozzino P Avellone G Bongiorno D Ceraulo L Filizzola F Natoli MC Livrea MA Tesoriere L Melatonin: structural characterization of its non-enzymatic mono-oxygenate metabolite J Pineal Res 2003 35 269 275 14521633 10.1034/j.1600-079X.2003.00086.x
Turjanski AG Chaia Z Rosenstein RE Estrin DA Doctorovich F Piro O N-nitrosomelatonin Acta Crystallogr 2000 C56 682 683
Turjanski AG Leonik F Rosenstein RE Estrin DA Doctorovich F Scavenging of NO by melatonin J Am Chem Soc 2000 122 10468 10469 10.1021/ja002006u
Turjanski AG Saenz DA Doctorovich F Estrin DA Rosenstein RE Nitrosation of melatonin by nitric oxide: a computational study J Pineal Res 2001 31 97 101 11555163 10.1034/j.1600-079x.2001.310201.x
Tesoriere L Avellone G Ceraulo L D'Arpa D Allegra M Livrea MA Oxidation of melatonin by oxoferryl hemoglobin: a mechanistic study Free Radic Res 2001 35 633 642 11811517
Hardeland R Coto-Montes A Poeggeler B Circadian rhythms, oxidative stress and antioxidative defense mechanisms Chronobiol Int 2003 20 921 962 14680136 10.1081/CBI-120025245
Hardeland R Poeggeler B Haldar C, Singaravel M, Pandi-Perumal SR, Cardinali DP Melatonin: New aspects of its protective actions and novel metabolites Experimental Endocrinology and Reproductive Biology 2005 Enfield, NH: Science Publishers
Blanchard B Pompon D Ducrocq C Nitrosation of melatonin by nitric oxide and peroxynitrite J Pineal Res 2000 29 184 192 11034116 10.1034/j.1600-079X.2000.290308.x
Reiter RJ Tan D-X Manchester LC Qi W Biochemical reactivity of melatonin with reactive oxygen and nitrogen species: a review of the evidence Cell Biochem Biophys 2001 34 237 256 11898866 10.1385/CBB:34:2:237
Reiter RJ Tan D-X Manchester LC Lopez-Burillo S Sainz RM Mayo JC Melatonin: detoxification of oxygen and nitrogen-based toxic reactants Adv Exp Med Biol 2003 527 53 548
Blanchard-Fillion B Servy C Ducrocq C 1-Nitrosomelatonin is a spontaneous NO-releasing compound Free Radic Res 2001 35 857 866 11811536
Hardeland R Saverschek N Hardeland R The NO adduct of spin trap TMA-PTIO [= 2-(trimethylammonio-phenyl)-4, 4, 5, 5-tetramethyl-imidazoline-1-oxyl-3-oxide] acts as an NO donor and stimulator of nitrosation reactions Actions and Redox Properties of Melatonin and Other Aromatic Amino Acid Metabolites 2001 Göttingen, Germany: Cuvillier 131 135
Zhang H Squadrito GL Uppu R Pryor WA Reaction of peroxynitrite with melatonin: a mechanistic study Chem Res Toxicol 1999 12 626 534
Zhang H Squadrito GL Pryor WA The reaction of melatonin with peroxynitrite: formation of melatonin radical cation and absence of stable nitrated products Biochem Biophys Res Commun 1998 251 83 87 9790911 10.1006/bbrc.1998.9426
Cimen B Turkozkan N Unlu A Erbil MK Effects of melatonin on 3-nitrotyrosine formation and energy charge ratio in guinea pig kidney in LPS-induced stress Cell Biochem Funct 2005 23 273 277 15515121 10.1002/cbf.1151
Gitto E Tan D-X Reiter RJ Karbownik M Manchester LC Cuzzocrea S Fulia F Berberi I Individual and synergistic antioxidative actions of melatonin: studies with vitamin E, vitamin C, glutathione and desferrioxamine (desferoxamine) in rat liver homogenates J Pharm Pharmacol 2001 53 1393 1401 11697548 10.1211/0022357011777747
Balkan J Sener G Cevikbas U Keyser-Uysal M Uysal M Melatonin improved the disturbances in hepatic prooxidant and antioxidant balance and hepatotoxicity induced by a cholesterol diet in C57BL/6J mice Int J Vitam Nutr Res 2004 74 349 254 15628673
Hardeland R Ressmeyer A-R Zelosko V Burkhardt S Poeggeler B Haldar C, SS Singh SS Metabolites of melatonin: Formation and properties of the methoxylated kynuramines AFMK and AMK Recent Advances in Endocrinology and Reproduction: Evolutionary, Biotechnological and Clinical Applications 2004 Varanasi: Banaras Hindu University 21 38
Poeggeler B Hardeland R Hardeland R Observations on melatonin oxidation and metabolite release by unicellular organisms and small aquatic metazoans Actions and Redox Properties of Melatonin and Other Aromatic Amino Acid Metabolites 2001 Göttingen: Cuvillier 66 69
Ferry G Ubeaud C Lambert PH Bertin S Cogé F Chomarat P Delagrange P Serkiz B Bouchet JP Truscott RJ Boutin JA Molecular evidence that melatonin is enzymatically oxidized in a different manner than tryptophan. Investigation on both indoleamine-2, 3-dioxygenase and myeloperoxidase Biochem J 2005 388 205 215 15636586
Hirata F Hayaishi O Tokuyama T Senoh S In vitro and in vivo formation of two new metabolites of melatonin J Biol Chem 1974 249 1311 1313 4814344
Alberati-Giani D Ricciardi-Castagnoli P Köhler C Cesura AM Regulation of the kynurenine pathway by IFN-γ in murine cloned macrophages and microglial cells Adv Exp Med Biol 1966 398 171 175 8906262
Lestage J Verrier D Palin K Dantzer R The enzyme indoleamine 2, 3-dioxygenase is induced in the mouse brain in response to peripheral administration of lipopolysaccharide and superantigen Brain Behav Immun 2002 16 596 601 12401474 10.1016/S0889-1591(02)00014-4
Kwidzinski E Bunse J Kovac AD Ullrich O Zipp F Nitsch R Bechmann U IDO (indolamine 2, 3-dioxygenase) expression and function in the CNS Adv Exp Med Biol 2003 527 113 118 15206723
Ximenes VF Catalani LH Campa A Oxidation of melatonin and tryptophan by an HRP cycle involving compound III Biochem Biophys Res Commun 2001 287 130 134 11549265 10.1006/bbrc.2001.5557
Rodrigues MR Rodriguez D Catalani LH Russo M Campa A Interferon-gamma independent oxidation of melatonin by macrophages J Pineal Res 2003 34 69 74 12485374 10.1034/j.1600-079X.2003.02944.x
Tan D-X Manchester LC Reiter RJ Qi W-B Karbownik M Calvo JR Significance of melatonin in antioxidative defense system: Reactions and products Biol Signals Recept 2000 9 137 159 10899700 10.1159/000014635
Balzer I Hardeland R Action of kynuramine in a dinoflagellate: stimulation of bioluminescence in Gonyaulax polyedra Comp Biochem Physiol 1989 94C 129 132
Behrends A Hardeland R Ness H Grube S Poeggeler B Haldar C Photocatalytic actions of the pesticide metabolite 2-hydroxyquinoxaline: destruction of antioxidant vitamines and biogenic amines – implications of organic redox cycling Redox Rep 2004 9 279 288 15606981 10.1179/135100004225006759
Kelly RW Amato F Seamark RF N-Acetyl-5-methoxy kynurenamine, a brain metabolite of melatonin, is a potent inhibitor of prostaglandin biosynthesis Biochem Biophys Res Commun 1984 121 372 379 6428407
Mayo JC Sainz RM Tan D-X Hardeland R Leon J Rodriguez C Reiter RJ Anti-inflammatory actions of melatonin and its metabolites, N1-acetyl-N2-formyl-5-methoxykynuramine (AFMK) and N1-acetyl-5-methoxykynuramine (AMK), in macrophages J Neuroimmunol 2005 165 139 149 15975667 10.1016/j.jneuroim.2005.05.002
Guenther AL Schmidt SI Laatsch H Fotso S Ness H Ressmeyer A-R Poeggeler B Hardeland R Reactions of the melatonin metabolite AMK (N 1-acetyl-5-methoxykynuramine) with reactive nitrogen species: Formation of novel compounds, 3-acetamidomethyl-6-methoxycinnolinone and 3-nitro-AMK J Pineal Res 2005 online ahead publ.: DOI 10.1111/j.1600-079X.2005.00242.x 16150105
Barlow-Walden LR Reiter RJ Abe M Pablos M Menendez-Pelaez A Chen LD Poeggeler B Melatonin stimulates brain glutathione peroxidase activity Neurochem Int 1995 26 497 502 7492947 10.1016/0197-0186(94)00154-M
Reiter RJ Melchiorri D Sewerynek E Poeggeler B Barlow-Walden L Chuang J Ortiz GG Acuña-Castroviejo D A review of the evidence supporting melatonin's role as an antioxidant J Pineal Res 1995 18 1 11 7776173
Pablos MI Agapito MT Gutierrez R Recio JM Reiter RJ Barlow-Walden L Acuña-Castroviejo D Menendez-Pelaez A Melatonin stimulates the activity of the detoxifying enzyme glutathione peroxidase in several tissues of chicks J Pineal Res 1995 19 111 115 8750343
Pablos MI Chuang J Reiter RJ Ortiz GG Daniels WM Sewerynek E Melchiorri D Poeggeler B Time course of the melatonin-induced increase in glutathione peroxidase activity in chick tissues Biol Signals 1995 4 325 30 8688911
Reiter RJ Garcia JJ Pie J Oxidative toxicity in models of neurodegeneration: responses to melatonin Restor Neurol Neurosci 1998 12 135 42 12671308
Kotler M Rodríguez C Sáinz RM Antolín I Menéndez-Peláez A Melatonin increases gene expression for antioxidant enzymes in rat brain cortex J Pineal Res 1998 24 83 89 9510432
Ustundag B Kazez A Demirbag M Canatan H Halifeoglu I Ozercan IH Protective effect of melatonin on antioxidative system in experimental ischemia-reperfusion of rat small intestine Cell Physiol Biochem 2000 10 229 236 11093033 10.1159/000016354
Okatani Y Wakatsuki A Kaneda C Melatonin increases activities of glutathione peroxidase and superoxide dismutase in fetal rat brain J Pineal Res 2000 28 89 96 10709970 10.1034/j.1600-079X.2001.280204.x
Reiter RJ Acuña-Castroviejo D Tan DX Burkhardt S Free radical-mediated molecular damage. Mechanisms for the protective actions of melatonin in the central nervous system Ann NY Acad Sci 2001 939 200 215 11462772
Wakatsuki A Okatani Y Shinohara K Ikenoue N Kaneda C Fukaya T Melatonin protects fetal rat brain against oxidative mitochondrial damage J Pineal Res 2001 30 22 28 11168903 10.1034/j.1600-079X.2001.300103.x
Gürdöl F Genç S Öner-Ýyidogan Y Süzme R Coadministration of melatonin and estradiol in rats: effects on oxidant status Horm Metab Res 2001 33 608 611 11607881 10.1055/s-2001-17908
Okatani Y Wakatsuki A Shinohara K Kaneda C Fukaya T Melatonin stimulates glutathione peroxidase activity in human chorion J Pineal Res 2001 30 199 205 11339508 10.1034/j.1600-079X.2001.300402.x
Okatani Y Wakatsuki A Reiter RJ Melatonin protects hepatic mitochondrial respiratory chain activity in senescence-accelerated mice J Pineal Res 2002 32 143 148 12074097 10.1034/j.1600-079x.2002.1o106.x
Mayo JC Sainz RM Antolin I Herrera F Martin V Rodriguez C Melatonin regulation of antioxidant enzyme gene expression Cell Mol Life Sci 2002 59 1706 1713 12475181
Gómez M Esparza JL Nogués MR Giralt M Cabré M Domingo JL Pro-oxidant activity of aluminum in the rat hippocampus: gene expression of antioxidant enzymes after melatonin administration Free Radic Biol Med 2005 38 104 111 15589378 10.1016/j.freeradbiomed.2004.10.009
Liu F Ng TB Effect of pineal indoles on activities of the antioxidant defense enzymes superoxide dismutase, catalase, and glutathione reductase, and levels of reduced and oxidized glutathione in rat tissues Biochem Cell Biol 2000 78 447 453 11012083 10.1139/bcb-78-4-447
Antolín I Rodríguez C Sáinz RM Mayo JC Uría H Kotler ML Rodríguez-Colunga MJ Tolivia D Menéndez-Peláez A Neurohormone melatonin prevents cell damage: effect on gene expression for antioxidant enzymes FASEB J 1996 10 882 890 8666165
Öztürk G Coşkun S Erbaş D Hasanoğlu E The effect of melatonin on liver superoxide dismutase activity, serum nitrate and thyroid hormone levels Jap J Physiol 2000 50 149 153 10866707 10.2170/jjphysiol.50.149
Wang H Wei W Shen YX Dong C Zhang LL Wang NP Yue L Xu SY Protective effect of melatonin against liver injury in mice induced by Bacillus Calmette-Guerin plus lipopolysaccharide World J Gastroenterol 2004 10 2690 2696 15309720
Rodriguez C Mayo JC Sainz RM Antolín I Herrera F Martín V Reiter RJ Regulation of antioxidant enzymes: a significant role for melatonin J Pineal Res 2004 36 1 9 14675124 10.1046/j.1600-079X.2003.00092.x
Tomás-Zapico C Coto-Montes A Martínez-Fraga J Rodríguez-Colunga MJ Hardeland R Tolivia D Effects of δ-aminolevulinic acid and melatonin in the Harderian gland of female Syrian hamster Free Radic Biol Med 2002 32 1197 1204 12031903 10.1016/S0891-5849(02)00812-2
Erten SF Kocak A Ozdemir I Aydemir S Colak A Reeder BS Protective effect of melatonin on experimental spinal cord ischemia Spinal Cord 2003 41 533 538 14504608 10.1038/sj.sc.3101508
Uz T Manev H Circadian expression of pineal 5-lipoxygenase mRNA Neuroreport 1998 9 783 786 9579665
Manev H Uz T Qu T Early upregulation of hippocampal 5-lipoxygenase following systemic administration of kainate to rats Restor Neurol Neurosci 1998 12 81 85 12671301
Zhang H Akbar M Kim HY Melatonin: an endogenous negative modulator of 12-lipoxygenation in the rat pineal gland Biochem J 1999 344 487 493 10567232 10.1042/0264-6021:3440487
Pozo D Reiter RJ Calvo JR Guerrero JM Physiological concentrations of melatonin inhibit nitric oxide synthase in rat cerebellum Life Sci 1994 55 PL455 PL460 7527477 10.1016/0024-3205(94)00532-X
Bettahi I Pozo D Osuna C Reiter RJ Acuña-Castroviejo D Guerrero JM Melatonin reduces nitric oxide synthase activity in rat hypothalamus J Pineal Res 1996 20 205 210 8836954
Gilad E Wong HR Zingarelli B Virag L O'Connor M Salzman AL Szabo C Melatonin inhibits expression of the inducible isoform of nitric oxide synthase in murine macrophages: role of inhibitions of NFκB activation FASEB J 1998 12 685 693 9619447
Chang HM Ling EA Chen CF Lue H Wen CY Shieh JY Melatonin attenuates the neuronal NADPH-d/NOS expression in the nodose ganglion of acute hypoxic rats J Pineal Res 2002 32 65 73 12071470 10.1034/j.1600-079x.2002.1816.x
Rogério F de Souza Queiroz L Teixeira SA Oliveira AL de Nucci G Langone F Neuroprotective action of melatonin on neonatal rat motoneurons after sciatic nerve transection Brain Res 2002 926 33 41 11814404 10.1016/S0006-8993(01)03286-3
Storr M Koppitz P Sibaev A Saur D Kurjak M Franck H Schusdziarra V Allescher HD Melatonin reduces non-adrenergic, non-cholinergic relaxant neurotransmission by inhibition of nitric oxide synthase activity in the gastrointestinal tract of rodents in vitro J Pineal Res 2002 33 101 108 12153444 10.1034/j.1600-079X.2002.02909.x
Nosjean O Ferro M Cogé F Beauverger P Henlin J-M Lefoulon F Fauchère J-L Delagrange P Canet E Boutin JA Identification of the melatonin-binding site MT3 as the quinone reductase 2 J Biol Chem 2000 275 31311 31317 10913150 10.1074/jbc.M005141200
Nosjean O Nicolas JP Klupsch F Delagrange P Canet E Boutin JA Comparative parmacological studies of melatonin receptors: MT1, MT2 and MT3/QR2. Tissue distribution of MT3/QR2 Biochem Pharmacol 2001 61 1369 1379 11331072 10.1016/S0006-2952(01)00615-3
Coto-Montes A Hardeland R Diurnal rhythm of protein carbonyl as an indicator of oxidative damage in Drosophila melanogaster : Influence of clock gene alleles and deficiencies in the formation of free-radical scavengers Biol Rhythm Res 1999 30 383 391 10.1076/brhm.30.4.383.1414
Coto-Montes A Tomás-Zapico C Rodríguez-Colunga MJ Tolivia-Cadrecha D Martínez-Fraga J Hardeland R Tolivia D Effects of the circadian mutation 'tau' on the Harderian glands of Syrian hamsters J Cell Biochem 2001 83 426 434 11596111 10.1002/jcb.1240
Sharman EH Bondy SC Effects of age and dietary antioxidants on cerebral electron transport chain activity Neurobiol Aging 2001 22 629 634 11445263 10.1016/S0197-4580(01)00226-3
Okatani Y Wakatsuki A Reiter RJ Miyahara Y Hepatic mitochondrial dysfunction in senescence-accelerated mice: correction by long-term, orally administered physiological levels of melatonin J Pineal Res 2002 33 127 133 12220325 10.1034/j.1600-079X.2002.02109.x
Martín M Macías M León J Escames G Khaldy H Acuña-Castroviejo D Melatonin increases the activity of the oxidative phosphorylation enzymes and the production of ATP in rat brain and liver mitochondria Int J Biochem Cell Biol 2002 34 348 357 11854034 10.1016/S1357-2725(01)00138-8
Acuña-Castroviejo D Martin M Macias M Escames G Leon J Khaldy H Reiter RJ Melatonin, mitochondria, and cellular bioenergetics J Pineal Res 2001 30 65 74 11270481 10.1034/j.1600-079X.2001.300201.x
Escames G León J Macías M Khaldy H Acuña-Castroviejo D Melatonin counteracts lipopolysaccharide-induced expression and activity of mitochondrial nitric oxide synthase in rats FASEB J 2003 17 932 934 12670878
Okatani Y Wakatsuki A Reiter RJ Enzan H Miyahara Y Protective effect of melatonin against mitochondrial injury induced by ischemia and reperfusion of rat liver Eur J Pharmacol 2003 469 145 152 12782196 10.1016/S0014-2999(03)01643-1
Watanabe K Wakatsuki A Shinohara K Ikenoue N Yokota K Fukaya T Maternally administered melatonin protects against ischemia and reperfusion-induced oxidative mitochondrial damage in premature fetal rat brain J Pineal Res 2004 37 276 280 15485554 10.1111/j.1600-079X.2004.00167.x
León J Acuña-Castroviejo D Escames G Tan D-X Reiter RJ Melatonin mitigates mitochondrial malfunction J Pineal Res 2005 38 1 9 15617531 10.1111/j.1600-079X.2004.00181.x
Acuña-Castroviejo D Escames G León J Carazo A Khaldy H Mitochondrial regulation by melatonin and its metabolites Adv Exp Med Biol 2003 527 549 557 15206773
Yalcin A Kilinc E Kocturk S Resmi H Sozmen EY Effect of melatonin cotreatment against kainic acid on coenzyme Q10, lipid peroxidation and Trx mRNA in rat hippocampus Int J Neurosci 2004 114 1085 1097 15370175 10.1080/00207450490475535
Martín M Macías M Escames G Reiter RJ Agapito MT Ortiz GG Acuña-Castroviejo D Melatonin-induced increased activity of the respiratory chain complexes I and IV can prevent mitochondrial damage induced by ruthenium red in vivo J Pineal Res 2000 28 242 248 10831160 10.1034/j.1600-079X.2000.280407.x
León J Acuña-Castroviejo D Sáinz RM Mayo JC Tan D-X Reiter RJ Melatonin and mitochondrial function Life Sci 2004 75 765 790 15183071 10.1016/j.lfs.2004.03.003
Okatani Y Wakatsuki A Reiter RJ Miyahara Y Acutely administered melatonin restores hepatic mitochondrial physiology in old mice Int J Biochem Cell Biol 2003 35 367 375 12531250 10.1016/S1357-2725(02)00260-1
Genova ML Merlo Pich M Bernacchia A Bianchi C Biondi A Bovina C Falasca AI Formiggini G Parenti Castelli G Lenaz G The mitochondrial production of reactive oxygen species in relation to aging and pathology Ann NY Acad Sci 2004 1011 86 100 15126287 10.1196/annals.1293.010
Jou M-J Peng T-I Reiter RJ Jou S-B Wu H-Y Wen S-T Visualization of the antioxidative effects of melatonin at the mitochondrial level during oxidative stress-induced apoptosis of rat brain astrocytes J Pineal Res 2004 37 55 70 15230869 10.1111/j.1600-079X.2004.00140.x
Andrabi SA Sayeed I Siemen D Wolf G Horn TF Direct inhibition of the mitochondrial permeability transition pore: a possible mechanism responsible for anti-apoptotic effects of melatonin FASEB J 2004 18 869 871 15033929
Xu M Ashraf M Melatonin protection against lethal myocyte injury induced by doxorubicin as reflected by effects on mitochondrial membrane potential J Mol Cell Cardiol 2002 34 75 79 11812166 10.1006/jmcc.2001.1485
Poeggeler B Neuroprotection by indole and nitrone compounds acting as mitochondrial metabolism modifiers with potent antioxidant activity [abstract] 9th International Conference on Alzheimer's Disease and Related Disorders, 17-22 July 2004; Philadelphia P4 398
Lewy AJ Ahmed S Jackson JM Sack RL Melatonin shifts human circadian rhythms according to a phase-response curve Chronobiol Int 1992 9 380 392 1394610
Lewy AJ Ahmed S Sack RL Phase shifting the human circadian clock using melatonin Behav Brain Res 1996 73 131 134 8788490 10.1016/0166-4328(96)00084-8
Sulli A Maestroni GJ Villaggio B Hertens E Craviotto C Pizzorni C Briata M Seriolo B Cutolo M Melatonin serum levels in rheumatoid arthritis Ann NY Acad Sci 2002 966 276 283 12114284
Scheer FA Van Montfrans GA van Someren EJ Mairuhu G Buijs RM Daily nighttime melatonin reduces blood pressure in male patients with essential hypertension Hypertension 2004 43 192 197 14732734 10.1161/01.HYP.0000113293.15186.3b
Lusardi P Piazza E Fogari R Cardiovascular effects of melatonin in hypertensive patients well controlled by nifedipine: a 24-hour study Br J Clin Pharmacol 2000 49 423 427 10792199 10.1046/j.1365-2125.2000.00195.x
Escames G Khaldy H León J González L Acuña-Castroviejo D Changes in iNOS activity, oxidative stress and melatonin levels in hypertensive patients treated with lacidipine J Hypertens 2004 22 629 635 15076170 10.1097/00004872-200403000-00027
|
16153306
|
PMC1262766
|
CC BY
|
2021-01-04 16:37:46
|
no
|
Nutr Metab (Lond). 2005 Sep 10; 2:22
|
utf-8
|
Nutr Metab (Lond)
| 2,005 |
10.1186/1743-7075-2-22
|
oa_comm
|
==== Front
Nutr Metab (Lond)Nutrition & Metabolism1743-7075BioMed Central London 1743-7075-2-251617429210.1186/1743-7075-2-25ReviewDietary protein intake and renal function Martin William F [email protected] Lawrence E [email protected] Nancy R [email protected] Department of Nutritional Sciences, University of Connecticut, Storrs, CT, USA2 Department of Kinesiology, University of Connecticut, Storrs, CT, USA2005 20 9 2005 2 25 25 8 3 2005 20 9 2005 Copyright © 2005 Martin et al; licensee BioMed Central Ltd.2005Martin et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Recent trends in weight loss diets have led to a substantial increase in protein intake by individuals. As a result, the safety of habitually consuming dietary protein in excess of recommended intakes has been questioned. In particular, there is concern that high protein intake may promote renal damage by chronically increasing glomerular pressure and hyperfiltration. There is, however, a serious question as to whether there is significant evidence to support this relationship in healthy individuals. In fact, some studies suggest that hyperfiltration, the purported mechanism for renal damage, is a normal adaptative mechanism that occurs in response to several physiological conditions. This paper reviews the available evidence that increased dietary protein intake is a health concern in terms of the potential to initiate or promote renal disease. While protein restriction may be appropriate for treatment of existing kidney disease, we find no significant evidence for a detrimental effect of high protein intakes on kidney function in healthy persons after centuries of a high protein Western diet.
==== Body
Dietary protein intake and renal function
Dietary protein intake can modulate renal function [1] and its role in renal disease has spawned an ongoing debate in the literature. At the center of the controversy is the concern that habitual consumption of dietary protein in excess of recommended amounts promotes chronic renal disease through increased glomerular pressure and hyperfiltration [2,3]. Media releases often conclude that, "too much protein stresses the kidney" [4]. The real question, however, is whether research in healthy individuals supports this notion. In fact, studies suggest that hyperfiltration in response to various physiological stimuli is a normal adaptative mechanism [5-10].
The purpose of this paper is to review the available evidence regarding the effects of protein intake on renal function with particular emphasis on renal disease. This review will consider research regarding the role of dietary protein in chronic kidney disease, normal renal function and kidney stone formation and evaluate the collective body of literature to ascertain whether habitual consumption of dietary protein in excess of what is recommended warrants a health concern in terms of the initiation and promotion of renal disease. In the following review, high protein (HP) diets will be defined as a daily consumption of greater than or equal to 1.5 g/kg/day, which is almost twice the current Recommended Dietary Allowance but within the range of current Dietary Reference Intakes (DRIs) for protein [11]. The Institute of Medicine DRI report concluded that there was insufficient scientific evidence for recommendations of an upper limit of protein intake but suggested an acceptable macronutrient distribution range of 10–35% of total energy for protein intake [11].
While the optimal ratio of macronutrient intake for adults has typically focused on fat and carbohydrate [12], contemporary discussions include the role of dietary protein [13-15]. This is particularly true given the recent popularity of high protein diets in weight management [16]. Although the efficacy of these diets with regard to weight loss is still subject to debate, several studies have demonstrated favorable physiological effects [12,16-24]. This has led to a substantial increase in protein intake by individuals adhering to contemporary weight loss plans. As a result, the safety of habitually consuming dietary protein in excess of the Recommended Daily Allowance (RDA) has been questioned.
An overview of chronic kidney disease
Chronic Kidney Disease (CKD) is defined as either kidney damage or a decline in renal function as determined by decreased glomerular filtration rate (GFR) for three or more months [25]. It is estimated that 1 in 9 adults in the United States meet this criteria, while an additional 1 in 9 adults are at increased risk for CKD [26]. In the general population, a decline in renal function is considered an independent risk factor for both cardiovascular disease and all-cause mortality [27]. However, the extent to which a mild diminution in renal function influences this risk is not known [28].
According to the National Kidney Foundation guidelines, CKD is classified into five stages, each of which directly correlates with the severity of the disease [25]. As one progresses from stage 1 to 5 there is a concomitant decline in GFR and thus renal function. The final stage, known as end stage renal disease, represents the most severe manifestation of CKD [29]. This classification system provides a universal standard for application of clinical treatment guidelines.
Hypertension is the second leading cause of CKD and accounts for approximately 30% of all cases in the U.S. [30,31]. In one study, hypertension was associated with a premature decline in renal function in men with normal kidney function [32]. Although, initial estimates of CKD prevalence in hypertensive individuals were about 2%, recent evidence suggests that prevalence rates may be significantly higher [33]. Blood pressure control is of particular importance in hypertensive individuals with CKD. This point has been demonstrated in several trials in which antihypertensive therapy slowed the progression of CKD [34-36].
Race, gender, age and family history are four risk factors for CKD [37-40]. Recent findings suggest that modifiable lifestyle risk factors (i.e., physical inactivity, smoking, obesity) are also associated with CKD. Limited data exist regarding the role of dietary protein intake as an independent risk factor for either the initiation or progression of renal disease but population studies have consistently demonstrated an inverse relationship between dietary protein intake and systemic blood pressure [41,42]. In a randomized control trial [43], dietary protein and fiber had additive effects in lowering 24-hour and awake systolic blood pressure in a group of 36 hypertensives. While these findings suggest that high protein diets may be beneficial to hypertensive individuals, additional research is warranted since increased protein intakes often result in increased consumption of certain micronutrients known to impact blood pressure (e.g., potassium, magnesium, calcium) [44].
Dietary protein and renal function
The relationship between dietary protein and renal function has been studied for over half a century [1]. In 1923, Addis and Drury [45] were among the first to observe a relationship between level of dietary protein and rates of urea excretion. Soon after, it was established that increased protein intake elevated rates of creatinine and urea excretion in the dog model [46]. The common mechanism underlying increased excretion rates was eventually attributed to changes in GFR [47,48] and Van Slyke et al. [49] demonstrated that renal blood flow was the basis for GFR mediated changes in clearance rates in response to increased protein intake. Clearly dietary protein effects GFR [50], with both acute and chronic increases in protein consumption elevating GFR [50,51].
Dietary protein and the progression of renal disease
Observational data from epidemiological studies provide evidence that dietary protein intake may be related to the progression of renal disease [52]. In the Nurses' Health Study, protein intake, assessed with a semi-quantitative food frequency questionnaire, was compared to the change in estimated GFR over an 11-year span in individuals with pre-existing renal disease [53]. Regression analysis showed an association between increased consumption of animal protein and a decline in renal function suggesting that high total protein intake may accelerate renal disease leading to a progressive loss of renal capacity. However, no association between protein intake and change in GFR was found in a different cohort of 1,135 women with normal renal function (Figure 1.). The latter finding led the authors to conclude that there were no adverse effects of high protein intakes on kidney function in healthy women with normal renal status.
Figure 1 This figure is a plot of multivariate linear regression for change in estimated GFR according to quintile of total protein intake* in participants with normal renal function (n = 1135). Data are taken from Knight et al., Ann Intern Med 2003 Mar 18;138(6):460-7 [53].
Research by Johnson et al. [54], showed protein intake as a possible risk factor for progressive loss of remaining renal function in dialysis patients. Indeed, dietary protein restriction is a common treatment modality for patients with renal disease [55,56] and practice guidelines exist regarding reduced dietary protein intakes for individuals with chronic renal disease in which proteinuria is present [57]. The National Kidney Foundation (NKF) has extensive recommendations with regard to protein intake, which are a byproduct of the Dialysis Outcome Quality Initiative [58]. Again, it is important to note that these recommendations are not indicated for individuals with normal renal function nor are they intended to serve as a prevention strategy to avoid developing CKD. Despite the clarity of these guidelines, their mere existence has resulted in concern regarding the role of dietary protein in the onset or progression of renal disease in the general population [59].
Dietary protein and renal disease
Allen and Cope's observation that increased dietary protein induced renal hypertrophy in dogs [60] led to speculation that dietary protein intake may have deleterious effects on the kidney. Later research in the rat model produced evidence supporting earlier observations from canine research [61-63]. Recently, Hammond and Janes [64] demonstrated an independent effect of increased protein intake on renal hypertrophy in mice. In this study, changes in renal function (i.e., increased glomerular filtration rate and renal hypertrophy) were observed.
Currently, a combination of hormonal interactions and renal processes are thought to explain protein-induced hyperfiltration [65]. Increased glucagon secretion in response to protein administration induced hyperfiltration [66] subsequent to a cascade of events referred to as the"pancreato-hepatorenal cascade" [67]. It has been hypothesized that cAMP works in concert with glucagon to mediate GFR [68]. To date, however, this hypothesis has not been tested and other competing hypotheses suggest other novel mechanisms of protein-induced hyperfiltration [69].
While the effect of hyperfiltration on renal function in those individuals with pre-existing renal disease is well documented [52], the application of these observations to healthy persons with normal renal function is not appropriate. To date, scientific data linking protein-induced renal hypertrophy or hyperfiltration to the initiation or progression of renal disease in healthy individuals is lacking. The possibility that protein-induced changes in renal function are a normal physiological adaptation to nitrogen load and increased demands for renal clearance is supported by changes noted in renal structure and function during pregnancy [70]. GFR increases by as much as 65% in healthy women [8] during pregnancy, typically returning to nonpregnant levels by three months postpartum [7]. Despite these changes in renal function, pregnancy is not a risk factor for developing CKD [6].
The renal hypertrophy and accompanying improvements in renal function in the contralateral kidney that occur subsequent to unilateral nephrectomy also suggest these processes are an adaptive, and possibly beneficial, response [5]. Studies show, despite prolonged hyperfiltration, remnant kidney function remained normal and did not deteriorate during long-term (> 20 yrs) follow-up in nephrectomized patients [9,10]. Thus, compensatory hyperfiltration appears to be a biological adaptation to a variety of renal challenges that is not associated with increased risk of chronic kidney disease in healthy individuals.
The Brenner Hypothesis
Perhaps the most consistently cited reference with regard to the potentially harmful effects of dietary protein intake on renal function is that of Brenner et al. [3]. In brief, the Brenner Hypothesis states that situations associated with increased glomerular filtration and glomerular pressure cause renal injury, ultimately compromise renal function, and potentially increase the risk for or progression of renal disease. Brenner proposed that habitual consumption of excessive dietary protein negatively impacted kidney function by a sustained increased in glomerular pressure and renal hyperfiltration [3]. Since the majority of scientific evidence cited by the authors was generated from animal models and patients with co-existing renal disease, extension of this relationship to healthy individuals with normal renal function is inappropriate. Indeed, a relationship between increased glomerular pressure or hyperfiltration and the onset or progression of renal disease in healthy individuals has not been clearly documented in the scientific literature. Rather, findings from individuals with compensatory hyperfiltration during pregnancy and following unilateral nephrectomy suggest otherwise [9].
The Modification of Diet in Renal Disease (MDRD) study was the largest randomized multicenter, controlled trial undertaken to evaluate the effect of dietary protein restriction on the progression of renal disease [71]. Several variables, including GFR, were measured in patients with chronic renal disease at baseline and throughout the approximately 2 year follow-up period. Patients with renal disease randomized to the very low-protein diet group had slightly slower decline in GFR decline compared with patients randomized to the low-protein diet group. Further data analyses showed patients with lower total protein intake would have a longer time to renal failure and suggested that a lower protein intake postponed the progression of advanced renal disease. Using meta-analysis to assess the efficacy of dietary protein restriction in previously published studies of diabetic and nondiabetic renal diseases, including the MDRD Study, Pedrini et al. concluded that the progression of both nondiabetic and diabetic renal disease could be effectively delayed with restriction of dietary protein [56]. Indeed, current clinical guidelines for the management of patients with renal disease continue to be based on the premise that protein intake greater than that recommended or which results in a renal solute load in excess of the kidney's excretory capabilities will contribute to progressive renal failure in persons with compromised renal function. However, of significance to this review, is the fact that imposing these guidelines on healthy individuals with normal renal function is overzealous given the current status of the scientific literature in this area.
Dietary protein and renal strain
Concerns about level of dietary protein and renal function are often presented in public health guidelines [59]. In addition to the claims that high protein intake causes renal disease, some studies have suggested that renal function may be negatively affected by routine consumption of high protein diets [72-75]. Although high protein diets cause changes in renal function (i.e., increased GFR) and several related endocrine factors [1,76,77] that may be harmful to individuals with renal disease [52,53], there is not sufficient research to extend these findings to healthy individuals with normal renal function at this time.
The lay public is often told that high protein diets "overwork" the kidney and may negatively impact renal function over time [78]. In addition, a number of highly regarded organizations appear to support this line of reasoning [79] given the physiological processes required for excretion of protein-related metabolic waste products to maintain homeostasis following consumption of protein at levels in excess of recommended amounts. Increased consumption of dietary protein is linearly related to the production of urea [80] and urea excretion is controlled by the kidney. These processes are of significant energetic cost to the kidney and represent the physiological "strain" associated with increased protein intake [81].
The word "strain" is misleading given its negative connotation. In a press release [82], one group asserted that increased dietary protein "strains" the kidney via increased urea production, and causes dehydration and accumulation of blood urea nitrogen. This press release also suggested that these events synergistically overwork the kidney and predispose humans to CKD. Scientific research is often misrepresented in this context. Research from our laboratory [83] which is cited in the press release, does not support these contentions. Rather, we found that habitual consumption of a high protein diet minimally affected hydration indices. Changes in total body water and renal function were not measured.
The concept that increased dietary protein leads to dehydration may have originated from an unsubstantiated extension of a 1954 review of the nitrogen balance literature [84]. This review focused on the design of survival rations for military operations in the desert or at sea, when water supply and energy intake are limited. Since the excretion of 1 gram of urea nitrogen requires 40 – 60 mL of additional water, increased protein intakes in the study translated into an increased water requirement (i.e., +250 mL water per 6 grams of dietary nitrogen in a 500 Kcal diet) for excretion of urea nitrogen. This increased fluid requirement is situation specific and is not necessarily applicable to individuals whose calorie and water intakes are adequate. Presently, we know of no studies executed in healthy individuals with normal renal function which demonstrate a clear relation between increased dietary protein intake and dehydration or a detrimental "strain" on the kidney. Therefore, claims that a high protein diet promotes dehydration or adversely "strains" the kidney remain speculative.
Evidence in healthy individuals
Although the efficacy of high protein diets for weight loss has been evaluated, there have been no reports of protein-induced diminutions in renal function despite subject populations that are generally at risk for kidney disease (e.g., dyslipidemia, obesity, hypertension) [14,15,22,85-87]. A randomized comparison of the effects of high and low protein diets on renal function in obese individuals suggested that high protein diets did not present a health concern with regard to renal function their study population [65]. In this study, 65 overweight, but otherwise healthy, subjects adhered to a low or high protein diet for six months. In the high protein group, both kidney size and GFR were significantly increased from that measured at baseline. No changes in albumin excretion were noted for either group and the authors concluded that, despite acute changes in renal function and size, high protein intake did not have detrimental effects on renal function in healthy individuals. Similar findings were recently reported by Boden et al. [88] in a study of 10 subjects who consumed their typical diet for 7 days followed by strict adherence to a high protein diet for 14 days. No significant changes were noted in serum or urinary creatinine and albumin excretion, suggesting no ill-effects of a high protein diet on renal function.
Athletes, particularly in sports requiring strength and power, consume high levels of dietary protein [89,90]. In fact, many athletes habitually consume protein in excess of 2.0 g/kg/day [91]. Supplementation with amino acids will further increase dietary protein levels in these individuals [92]. Yet there is no evidence that this population is at greater risk for kidney disease or losses in renal function [90]. Poortsmans and Dellalieux [93] found that protein intakes in the range of ~1.4–1.9 g/kg/day or 170–243% of the recommended dietary allowance did not impair renal function in a group of 37 athletes. We found no data in the scientific literature to link high protein intakes to increased risk for impaired kidney function in healthy, physically active men and women.
Dietary protein and renal function in animal models
Although there is limited research regarding the long-term effects of high protein intakes on renal function in humans, animal models have provided insight into this quandary. Mammals fed acute and chronic high protein diets exhibit increases in GFR and renal blood flow [94]. These changes, which are comparable to those observed in humans, led to the hypothesis that high protein intakes are associated with progressive glomerulosclerosis in the rat. Recently, Lacroix et al. [95] studied the effects of a diet containing 50% protein on renal function in Wistar rats and noted no abnormalities in renal function or pathology. Collins et al. [96] also reported no adverse effects of long-term consumption of high protein diets on renal function when two years of a diet containing 60% protein failed to evoke changes in the percentage of sclerotic glomeruli in rats. Robertson et al., [97] studied the effect of increased protein intake on hyperperfusion and the progression of glomerulosclerosis in dogs that were 75% nephrectomized. After four years of feeding diets that were either 56, 27 or 19% protein, no association between diet and structural changes in the kidney were observed.
To the best of our knowledge, there has been only one report of a potentially toxic effect of excessive protein intake on renal function in the rat. Stonard et al. [98] found a diet containing 33% protein produced tubular damage in a specific strain of female rats. However, findings from this study are limited by the fact that damage was induced by a bacterial single-cell protein (Pruteen).
In summary, studies documenting high protein intake as a cause of renal disease in any animal model have not been done. Rather, studies have typically focused on the interaction between protein intake and renal function in the diseased state. As a result, findings from these investigations should not be used as a basis for dietary recommendations for humans. Studies designed to characterize the effects of dietary protein intake on renal function in healthy subjects are warranted.
Dietary protein and kidney stones
The role of high protein diets in kidney stone formation has received considerable attention. Excessive protein intake increases excretion of potentially lithogenic substances such as calcium and uric acid [99,100]. Reddy et al. [101] noted that consumption of a high protein diet for six weeks was associated aciduria and urinary calcium and claimed that this constituted increased risk of stone formation in ten healthy subjects although none of the ten subjects developed renal stones. The severe carbohydrate restriction imposed in this study may have increased keto-acid production thereby contributing acid formation. Since consumption of fruits and vegetables usually produces a marked base load [102], restriction of these foods subsequent to the diet intervention may have also contributed to the net acid load.
Studies that claim an increased propensity for stone formation as a result of increased protein intake should be taken at face value because propensity is a surrogate marker and does not represent actual stone formation. Further, randomized control trials have not been done to test whether an increased tendency for stone formation is enhanced with consumption of a high protein diet.
Epidemiological studies provide conflicting evidence with regard to the association between protein intake and the predisposition for kidney stone formation. In a prospective study of over 45,000 men, researchers found a direct correlation between animal protein intake and risk of stone formation [103]. However, findings in women are difficult to interpret due to conflicting reports in the literature. While some studies have shown a direct relationship between animal protein intake and risk of stone formation in women [104,105], other work suggests an inverse relationship exists [106].
Conflicting findings regarding the role of dietary protein in kidney stone formation limit the development of universal guidelines with regard to a recommended protein intake for individuals at increased risk for stone formation [107]. It is not likely that diet alone causes kidney stone formation [108]. Rather, metabolic abnormalities are typically the underlying cause [109]. For example, Nguyen et al. [110] found that high intakes of animal protein adversely affected markers of stone formation in those afflicted with a stone causing disorder, while no changes were observed in healthy individuals. It has been suggested that one must have a preexisting metabolic dysfunction before dietary protein can exert an effect relative to stone formation [108]. This notion has been coined the "powderkeg and tinderbox" theory of renal stone disease by Jaeger [111]. This theory asserts that dietary excesses, such as high protein intake, serve as a tinderbox which, only in tandem with a metabolic abnormality (the powderkeg), can bring about stone formation. At the present time, however, evidence showing that a high protein intake is an inherent cause of this renal abnormality or is consistently associated with increased kidney stone formation does not exist.
Conclusion
Although excessive protein intake remains a health concern in individuals with pre-existing renal disease, the literature lacks significant research demonstrating a link between protein intake and the initiation or progression of renal disease in healthy individuals. More importantly, evidence suggests that protein-induced changes in renal function are likely a normal adaptative mechanism well within the functional limits of a healthy kidney. Without question, long-term studies are needed to clarify the scant evidence currently available regarding this relationship. At present, there is not sufficient proof to warrant public health directives aimed at restricting dietary protein intake in healthy adults for the purpose of preserving renal function.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
WFM conducted literature search, prepared the manuscript and assisted in presentation of final draft, LEA and NRR conceived the idea, organized contents and participated in preparation of final manuscript.
==== Refs
King AJ Levey AS Dietary protein and renal function J Am Soc Nephrol 1993 3 1723 1737 8329667
Metges CC Barth CA Metabolic consequences of a high dietary-protein intake in adulthood: assessment of the available evidence J Nutr 2000 130 886 889 10736347
Brenner BM Meyer TW Hostetter TH Dietary protein intake and the progressive nature of kidney disease: the role of hemodynamically mediated glomerular injury in the pathogenesis of progressive glomerular sclerosis in aging, renal ablation, and intrinsic renal disease N Engl J Med 1982 307 652 659 7050706
The University of Pennsylvania Health System Media Review: mouth to mouth 1999 Janurary, 1999
Sugaya K Ogawa Y Hatano T Koyama Y Miyazato T Naito A Yonou H Kagawa H Compensatory renal hypertrophy and changes of renal function following nephrectomy Hinyokika Kiyo 2000 46 235 240 10845153
Calderon JL Zadshir A Norris K A survey of kidney disease and risk-factor information on the World Wide Web MedGenMed 2004 6 3 15775830
Lindheimer MD Katz AI Seldin DW, Giebisch G Physiology and Pathophysiology Renal physiology and disease in pregnancy 1992 2nd New York , Raven Press 3371–3431
Conrad KP Mechanisms of renal vasodilation and hyperfiltration during pregnancy J Soc Gynecol Investig 2004 11 438 448 15458740 10.1016/j.jsgi.2004.05.002
Higashihara E Horie S Takeuchi T Nutahara K Aso Y Long-term consequence of nephrectomy J Urol 1990 143 239 243 2299710
Regazzoni BM Genton N Pelet J Drukker A Guignard JP Long-term followup of renal functional reserve capacity after unilateral nephrectomy in childhood J Urol 1998 160 844 848 9720572 10.1097/00005392-199809010-00073
Food and Nutrition Board Institute of Medicine Macronutrient and Healthful Diets Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids (Macronutrients) 2002 Washington, D.C. , The National Academies Press 609 696
Layman DK Boileau RA Erickson DJ Painter JE Shiue H Sather C Christou DD A Reduced Ratio of Dietary Carbohydrate to Protein Improves Body Composition and Blood Lipid Profiles during Weight Loss in Adult Women J Nutr 2003 133 411 417 12566476
Piatti PM Monti F Fermo I Baruffaldi L Nasser R Santambrogio G Librenti MC Galli-Kienle M Pontiroli AE Pozza G Hypocaloric high-protein diet improves glucose oxidation and spares lean body mass: comparison to hypocaloric high-carbohydrate diet Metabolism 1994 43 1481 1487 7990700 10.1016/0026-0495(94)90005-1
Luscombe ND Clifton PM Noakes M Farnsworth E Wittert G Effect of a high-protein, energy-restricted diet on weight loss and energy expenditure after weight stabilization in hyperinsulinemic subjects Int J Obes Relat Metab Disord 2003 27 582 590 12704402 10.1038/sj.ijo.0802270
Brinkworth GD Noakes M Keogh JB Luscombe ND Wittert GA Clifton PM Long-term effects of a high-protein, low-carbohydrate diet on weight control and cardiovascular risk markers in obese hyperinsulinemic subjects Int J Obes Relat Metab Disord 2004 28 661 670 15007396 10.1038/sj.ijo.0802617
Fine EJ Feinman RD Thermodynamics of weight loss diets Nutr Metab (Lond) 2004 1 15 15588283 10.1186/1743-7075-1-15
Parker B Noakes M Luscombe N Clifton P Effect of a High-Protein, High-Monounsaturated Fat Weight Loss Diet on Glycemic Control and Lipid Levels in Type 2 Diabetes Diabetes Care 2002 25 425 430 11874925
Layman DK Protein quantity and quality at levels above the RDA improves adult weight loss J Am Coll Nutr 2004 23 631S 636S 15640518
Wolfe RR Chinkes D Baba H Rosenblatt J Zhang XJ Response of phosphoenolpyruvate cycle activity to fasting and to hyperinsulinemia in human subjects Am J Physiol Endocrinol Metab 1996 271 E159 176
Sharman MJ Volek JS Weight loss leads to reductions in inflammatory biomarkers after a very-low-carbohydrate diet and a low-fat diet in overweight men Clin Sci (Lond) 2004 107 365 369 15265001
Yancy WSJ Olsen MK Guyton JR Bakst RP Westman EC A low-carbohydrate, ketogenic diet versus a low-fat diet to treat obesity and hyperlipidemia: a randomized, controlled trial Ann Intern Med 2004 140 769 777 15148063
Johnston CS Tjonn SL Swan PD High-Protein, Low-Fat Diets Are Effective for Weight Loss and Favorably Alter Biomarkers in Healthy Adults J Nutr 2004 134 586 591 14988451
Foster GD Wyatt HR Hill JO McGuckin BG Brill C Mohammed BS Szapary PO Rader DJ Edman JS Klein S A randomized trial of a low-carbohydrate diet for obesity N Engl J Med 2003 348 2082 2090 12761365 10.1056/NEJMoa022207
Skov AR Toubro S Ronn B Holm L Astrup A Randomized trial on protein vs carbohydrate in ad libitum fat reduced diet for the treatment of obesity Int J Obes Relat Metab Disord 1999 23 528 536 10375057 10.1038/sj.ijo.0800867
Levey AS Coresh J Balk E Kausz AT Levin A Steffes MW Hogg RJ Perrone RD Lau J Eknoyan G National Kidney Foundation practice guidelines for chronic kidney disease: evaluation, classification, and stratification Ann Intern Med 2003 139 137 147 12859163
Coresh J Byrd-Holt D Astor BC Briggs JP Eggers PW Lacher DA Hostetter TH Chronic Kidney Disease Awareness, Prevalence, and Trends among U.S. Adults, 1999 to 2000 J Am Soc Nephrol 2005 16 180 188 15563563 10.1681/ASN.2004070539
Muntner P He J Hamm L Loria C Whelton PK Renal insufficiency and subsequent death resulting from cardiovascular disease in the United States J Am Soc Nephrol 2002 13 745 753 11856780
Coresh J Astor BC Greene T Eknoyan G Levey AS Prevalence of chronic kidney disease and decreased kidney function in the adult US population: Third National Health and Nutrition Examination Survey Am J Kidney Dis 2003 41 1 12 12500213 10.1053/ajkd.2003.50007
Johnson CA Creating practice guidelines for chronic kidney disease: an insider's view Am Fam Physician 2004 70 823 824 15368721
Palmer BF Disturbances in renal autoregulation and the susceptibility to hypertension-induced chronic kidney disease Am J Med Sci 2004 328 330 343 15599329
National Kidney Foundation: Chronic Kidney Diseases.
Vupputuri S Batuman V Muntner P Bazzano LA Lefante JJ Whelton PK He J Effect of blood pressure on early decline in kidney function among hypertensive men Hypertension 2003 42 1144 1149 14597644 10.1161/01.HYP.0000101695.56635.31
Segura J Campo C Gil P Roldan C Vigil L Rodicio JL Ruilope LM Development of chronic kidney disease and cardiovascular prognosis in essential hypertensive patients J Am Soc Nephrol 2004 15 1616 1622 15153573 10.1097/01.ASN.0000127045.14709.75
Wright JTJ Bakris G Greene T Agodoa LY Appel LJ Charleston J Cheek D Douglas-Baltimore JG Gassman J Glassock R Hebert L Jamerson K Lewis J Phillips RA Toto RD Middleton JP Rostand SG Effect of blood pressure lowering and antihypertensive drug class on progression of hypertensive kidney disease: results from the AASK trial Jama 2002 288 2421 2431 12435255 10.1001/jama.288.19.2421
Kasiske BL Kalil RS Ma JZ Liao M Keane WF Effect of antihypertensive therapy on the kidney in patients with diabetes: a meta-regression analysis Ann Intern Med 1993 118 129 138 8416309
Peterson JC Adler S Burkart JM Greene T Hebert LA Hunsicker LG King AJ Klahr S Massry SG Seifter JL Blood pressure control, proteinuria, and the progression of renal disease. The Modification of Diet in Renal Disease Study Ann Intern Med 1995 123 754 762 7574193
Tarver-Carr ME Powe NR Eberhardt MS LaVeist TA Kington RS Coresh J Brancati FL Excess risk of chronic kidney disease among African-American versus white subjects in the United States: a population-based study of potential explanatory factors J Am Soc Nephrol 2002 13 2363 2370 12191981 10.1097/01.ASN.0000026493.18542.6A
Zheng F Plati AR Banerjee A Elliot S Striker LJ Striker GE The molecular basis of age-related kidney disease Sci Aging Knowledge Environ 2003 2003 PE20 12878780
Brown WW Sandberg K Introduction: gender and kidney disease Adv Ren Replace Ther 2003 10 1 2 12616457 10.1053/jarr.2003.50009
Satko SG Freedman BI The importance of family history on the development of renal disease Curr Opin Nephrol Hypertens 2004 13 337 341 15073494
Zhou BF Wu XG Tao SQ Yang J Cao TX Zheng RP Tian XZ Lu CQ Miao HY Ye FM Dietary patterns in 10 groups and the relationship with blood pressure. Collaborative Study Group for Cardiovascular Diseases and Their Risk Factors Chin Med J (Engl) 1989 102 257 261 2507238
He J Klag MJ Whelton PK Chen JY Qian MC He GQ Dietary macronutrients and blood pressure in southwestern China J Hypertens 1995 13 1267 1274 8984124
Burke V Hodgson JM Beilin LJ Giangiulioi N Rogers P Puddey IB Dietary Protein and Soluble Fiber Reduce Ambulatory Blood Pressure in Treated Hypertensives Hypertension 2001 38 821 826 11641293
St. Jeor ST Howard BV Prewitt TE Bovee V Bazzarre T Eckel RH Dietary Protein and Weight Reduction: A statement for healthcare professionals from the nutrition committee of the council on nutrition, physical Activity, and metabolism of the american heart association Circulation 2001 104 1869 1874 11591629
Addis T Drury DR The Rate of Urea Excretion. VII. The effect of various other factors than blood urea concentration on the rate of urea excretion J Biol Chem 1923 55 629 638
Jolliffe N Smith HW The Excretion of Urine In The Dog: I. The Urea and Creatinine Clearances on a Mixed Diet Am J Physiol 1931 98 572 577
Shannon JA Jolliffe N Smith HW The Excretion of Urine in The Dog: VI. The Filtration and Secretion of Exogenous Creatinine Am J Physiol 1932 102 534 550
Herrin RC Rabin A Feinstein RN The influence of diet upon urea clearance in dogs Am J Physiol 1937 119 87 92
Van Slyke DD Rhoads CP Hiller A Alving A The relationship of the urea clearance to the renal blood flow Am J Physiol 1934 110 387 391
Tuttle KR Puhlman ME Cooney SK Short RA Effects of amino acids and glucagon on renal hemodynamics in type 1 diabetes Am J Physiol Renal Physiol 2002 282 F103 12 11739118
Bilo HJ Schaap GH Blaak E Gans RO Oe PL Donker AJ Effects of chronic and acute protein administration on renal function in patients with chronic renal insufficiency Nephron 1989 53 181 187 2797339
Lentine K Wrone EM New insights into protein intake and progression of renal disease Curr Opin Nephrol Hypertens 2004 13 333 336 15073493
Knight EL Stampfer MJ Hankinson SE Spiegelman D Curhan GC The Impact of Protein Intake on Renal Function Decline in Women with Normal Renal Function or Mild Renal Insufficiency Ann Intern Med 2003 138 460 467 12639078
Johnson DW Mudge DW Sturtevant JM Hawley CM Campbell SB Isbel NM Hollett P Predictors of decline of residual renal function in new peritoneal dialysis patients Perit Dial Int 2003 23 276 283 12938830
Meloni C Tatangelo P Cipriani S Rossi V Suraci C Tozzo C Rossini B Cecilia A Di Franco D Straccialano E Casciani CU Adequate protein dietary restriction in diabetic and nondiabetic patients with chronic renal failure J Ren Nutr 2004 14 208 213 15483780
Pedrini MT Levey AS Lau J Chalmers TC Wang PH The effect of dietary protein restriction on the progression of diabetic and nondiabetic renal diseases: a meta-analysis Ann Intern Med 1996 124 627 632 8607590
Franz MJ Wheeler ML Nutrition therapy for diabetic nephropathy Curr Diab Rep 2003 3 412 417 12975032
Beto JA Bansal VK Medical nutrition therapy in chronic kidney failure: integrating clinical practice guidelines Journal of the American Dietetic Association 2004 104 404 409 14993863 10.1016/j.jada.2003.12.028
Krauss RM Eckel RH Howard B Appel LJ Daniels SR Deckelbaum RJ Erdman JWJ Kris-Etherton P Goldberg IJ Kotchen TA Lichtenstein AH Mitch WE Mullis R Robinson K Wylie-Rosett J St Jeor S Suttie J Tribble DL Bazzarre TL AHA Dietary Guidelines: revision 2000: A statement for healthcare professionals from the Nutrition Committee of the American Heart Association Stroke 2000 31 2751 2766 11062305
Allen FM Cope OM Influence of Diet on Blood Pressure and Kidney Size in Dogs J Urol 1942 47 751
Osborne TB Mendel LB Park EA Winternitz MC With the cooperation of Helen C. Cannon and Deborah Jackson Physiological effectgs of diets unusually rich in protein or inorganic salts J Biol Chem 1927 71 317 350
Wilson HE An Investigation of the Cause of Renal Hypertrophy in Rats Fed on a High Protein Diet Biochem J 1933 27 1348 16745239
Addis T MacKay EM MacKay LL The effect on the kidney of the long continued administration of diets containing an excess of certain food elements. II. Excess of acid and alkali J Biol Chem 1926 71 157 166
Hammond KA Janes DN The effects of increased protein intake on kidney size and function J Exp Biol 1998 201 ( Pt 13) 2081 2090 9622580
Skov AR Toubro S Bulow J Krabbe K Parving HH Astrup A Changes in renal function during weight loss induced by high vs low-protein low-fat diets in overweight subjects Int J Obes Relat Metab Disord 1999 23 1170 1177 10578207 10.1038/sj.ijo.0801048
Gorin E Dickbuch S Release of cyclic AMP from chicken erythrocytes Horm Metab Res 1980 12 120 124 6245999
Bankir L Martin H Dechaux M Ahloulay M Plasma cAMP: a hepatorenal link influencing proximal reabsorption and renal hemodynamics? Kidney Int Suppl 1997 Suppl 59 S50 6 9185105
Bankir L Ahloulay M Devreotes PN Parent CA Extracellular cAMP inhibits proximal reabsorption: are plasma membrane cAMP receptors involved? Am J Physiol Renal Physiol 2002 282 F376 392 11832418
Slomowitz LA Gabbai FB Khang S Satriano J Thareau S Deng A Thomson SC Blantz RC Munger KA Protein intake regulates the vasodilatory function of the kidney and the NMDA receptor expression Am J Physiol Regul Integr Comp Physiol 2004 R1184 9 15242825
Conrad KP Novak J Danielson LA Kerchner LJ Jeyabalan A Mechanisms of renal vasodilation and hyperfiltration during pregnancy: current perspectives and potential implications for preeclampsia Endothelium 2005 12 57 62 16036316
Klahr S The modification of diet in renal disease study N Engl J Med 1989 320 864 866 2494456
Consumer Reports on Health Feature Report; Is your diet up-to-date? (New recommendations have changed the standard advice) 2003 4 6
Time How to Eat Smarter; In a world that is raining food, making healthy choices about what and how to eat is not easy. Here are some rules to live by 2003 48
CNBC Diet Dangers; Debate on the dangers of high protein-low carb diets after Physicians Committee for Responsible Medicine came to warn about health risks. 2003 Capital Report , Video Monitoring Services of America
Fox News Channel Diet; People who do the low carb diet will get hurt. 2004 The O'Reilly Factor , Video Monitoring Services of America
Schaffer SW Lombardini JB Azuma J Interaction between the actions of taurine and angiotensin II Amino Acids 2000 18 305 318 10949914
Brandle E Sieberth HG Hautmann RE Effect of chronic dietary protein intake on the renal function in healthy subjects Eur J Clin Nutr 1996 50 734 740 8933120
USA Today High-protein diets gaining support 1999
Taubes G What if It's All Been a Big Fat Lie? The New York Times Magazine 2002
Young VR El-Khoury AE Raguso CA Forslund AH Hambraeus L Rates of urea production and hydrolysis and leucine oxidation change linearly over widely varying protein intakes in healthy adults J Nutr 2000 130 761 766 10736327
Bankir L Bouby N Trinh-Trang-Tan MM Ahloulay M Promeneur D Direct and indirect cost of urea excretion Kidney Int 1996 49 1598 1607 8743462
AtkinsExposed.org
Martin WF Bolster DR Gaine PC Hanley LJ Pikosky MA Bennett BT Maresh CM Armstrong LE Rodriguez NR Increased Dietary Protein Affects Hydration Indices in Runners [in press] J Am Diet Assoc 2006 106
Calloway DH Spector H Nitrogen balance as related to caloric and protein intake in active young men Am J Clin Nutr 1954 2 405 412 13207036
Layman DK Baum JI Dietary Protein Impact on Glycemic Control during Weight Loss J Nutr 2004 Suppl 4 968S 973 15051856
Due A Toubro S Skov AR Astrup A Effect of normal-fat diets, either medium or high in protein, on body weight in overweight subjects: a randomised 1-year trial Int J Obes Relat Metab Disord 2004 28 1283 1290 15303109 10.1038/sj.ijo.0802767
Stern L Iqbal N Seshadri P Chicano KL Daily DA McGrory J Williams M Gracely EJ Samaha FF The effects of low-carbohydrate versus conventional weight loss diets in severely obese adults: one-year follow-up of a randomized trial Ann Intern Med 2004 140 778 785 15148064
Boden G Sargrad K Homko C Mozzoli M Stein TP Effect of a Low-Carbohydrate Diet on Appetite, Blood Glucose Levels, and Insulin Resistance in Obese Patients with Type 2 Diabetes Ann Intern Med 2005 142 403 411 15767618
Fern EB Bielinski RN Schutz Y Effects of exaggerated amino acid and protein supply in man Experientia 1991 47 168 172 2001721 10.1007/BF01945420
Lemon PW Is increased dietary protein necessary or beneficial for individuals with a physically active lifestyle? Nutr Rev 1996 54 S169 75 8700446
Chen JD Wang JF Li KJ Zhao YW Wang SW Jiao Y Hou XY Nutritional problems and measures in elite and amateur athletes Am J Clin Nutr 1989 49 1084 1089 2541605
Chromiak JA Antonio J Use of amino acids as growth hormone-releasing agents by athletes Nutrition 2002 18 657 661 12093449 10.1016/S0899-9007(02)00807-9
Poortmans JR Dellalieux O Do regular high protein diets have potential health risks on kidney function in athletes? Int J Sport Nutr Exerc Metab 2000 10 28 38 10722779
Singer MA Dietary protein-induced changes in excretory function: a general animal design feature Comparative Biochemistry and Physiology Part B: Biochemistry and Molecular Biology 2003 136 785 801 10.1016/j.cbpc.2003.08.012
Lacroix M Gaudichon C Martin A Morens C Mathe V Tome D Huneau JF A long-term high-protein diet markedly reduces adipose tissue without major side effects in Wistar male rats Am J Physiol Regul Integr Comp Physiol 2004 287 R934 42 15155276
Collins DMCTRJBKPRTMCPEK Chronic high protein feeding does not produce glomerulosclerosis or renal insufficiency in the normal rat J Am Soc Nephrol 1990 1 624
Robertson JL Goldschmidt M Kronfeld DS Tomaszewski JE Hill GS Bovee KC Long-term renal responses to high dietary protein in dogs with 75% nephrectomy Kidney Int 1986 29 511 519 3702209
Stonard MD Samuels DM Lock EA The pathogenesis and effect on renal function of nephrocalcinosis induced by different diets in female rats Food Chem Toxicol 1984 22 139 146 6538163 10.1016/0278-6915(84)90094-2
Wasserstein AG Stolley PD Soper KA Goldfarb S Maislin G Agus Z Case-control study of risk factors for idiopathic calcium nephrolithiasis Miner Electrolyte Metab 1987 13 85 95 3696092
Robertson WG Heyburn PJ Peacock M Hanes FA Swaminathan R The effect of high animal protein intake on the risk of calcium stone-formation in the urinary tract Clin Sci (Lond) 1979 57 285 288 573189
Reddy ST Wang CY Sakhaee K Brinkley L Pak CY Effect of low-carbohydrate high-protein diets on acid-base balance, stone-forming propensity, and calcium metabolism Am J Kidney Dis 2002 40 265 274 12148098 10.1053/ajkd.2002.34504
Cordain L Eaton SB Sebastian A Mann N Lindeberg S Watkins BA O'Keefe JH Brand-Miller J Origins and evolution of the Western diet: health implications for the 21st century Am J Clin Nutr 2005 81 341 354 15699220
Curhan GC Willett WC Rimm EB Stampfer MJ A prospective study of dietary calcium and other nutrients and the risk of symptomatic kidney stones N Engl J Med 1993 328 833 838 8441427 10.1056/NEJM199303253281203
Breslau NA Brinkley L Hill KD Pak CY Relationship of animal protein-rich diet to kidney stone formation and calcium metabolism J Clin Endocrinol Metab 1988 66 140 146 2826524
Curhan GC Willett WC Speizer FE Spiegelman D Stampfer MJ Comparison of dietary calcium with supplemental calcium and other nutrients as factors affecting the risk for kidney stones in women Ann Intern Med 1997 126 497 504 9092314
Curhan GC Willett WC Knight EL Stampfer MJ Dietary factors and the risk of incident kidney stones in younger women: Nurses' Health Study II Arch Intern Med 2004 164 885 891 15111375 10.1001/archinte.164.8.885
Meschi T Schianchi T Ridolo E Adorni G Allegri F Guerra A Novarini A Borghi L Body weight, diet and water intake in preventing stone disease Urol Int 2004 Suppl 1 29 33 15133330 10.1159/000076588
Hess B Nutritional aspects of stone disease Endocrinol Metab Clin North Am 2002 31 1017 30, ix-x 12474643 10.1016/S0889-8529(02)00029-4
Raj GV Auge BK Assimos D Preminger GM Metabolic abnormalities associated with renal calculi in patients with horseshoe kidneys J Endourol 2004 18 157 161 15072623 10.1089/089277904322959798
Nguyen QV Kalin A Drouve U Casez JP Jaeger P Sensitivity to meat protein intake and hyperoxaluria in idiopathic calcium stone formers Kidney Int 2001 59 2273 2281 11380831
Jaeger P Renal stone disease in the 1990s: the powder keg and tinderbox theory Curr Opin Nephrol Hypertens 1992 1 141 148 1366343
|
16174292
|
PMC1262767
|
CC BY
|
2021-01-04 16:37:45
|
no
|
Nutr Metab (Lond). 2005 Sep 20; 2:25
|
utf-8
|
Nutr Metab (Lond)
| 2,005 |
10.1186/1743-7075-2-25
|
oa_comm
|
==== Front
Part Fibre ToxicolParticle and Fibre Toxicology1743-8977BioMed Central London 1743-8977-2-51620213710.1186/1743-8977-2-5ResearchExposure and risks from wearing asbestos mitts Cherrie John W [email protected] Matthew [email protected] Hilary [email protected] Institute of Occupational Medicine, Research Park North, Riccarton, Edinburgh, EH14 4AP, UK2 University of Aberdeen, Department of Environmental and Occupational Medicine, Foresterhill Road, Aberdeen, AB25 2ZP, UK3 Rilmac (Insulation) Ltd, Crofton Drive, Lincoln, LN3 4NJ, UK2005 3 10 2005 2 5 5 11 5 2005 3 10 2005 Copyright © 2005 Cherrie et al; licensee BioMed Central Ltd.2005Cherrie et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Very high fibre inhalation exposure has been measured while people were wearing personal protective equipment manufactured from chrysotile asbestos. However, there is little data that relates specifically to wearing asbestos gloves or mitts, particularly when used in hot environments such as those found in glass manufacturing. The aim of this study was to assess the likely personal exposure to asbestos fibres when asbestos mitts were used.
Results
Three types of work activity were simulated in a small test room with unused mitts and artificially aged mitts. Neither pair of mitts were treated to suppress the dust emission. The measured respirable fibre exposure levels ranged from <0.06 to 0.55 fibres/ml, with no significant difference in fibre exposure between aged and unused mitts. The use of high localised ventilation to simulate convective airflows from a furnace reduced exposure levels by about a factor of five. Differences between tasks were statistically significant, with simulated "rowing" of molten glass lowest and replacement of side seals on the furnace highest. Estimated lifetime cancer risk from 20 years exposure at the upper end of the exposure range measured during the study is less than 22 per 100,000.
Conclusion
People who wore asbestos mitts were likely to have been exposed to relatively low levels of airborne chrysotile asbestos fibres, certainly much lower than the standards that were accepted in the 1960's and 70's. The cancer risks from this type of use are likely to be very low.
==== Body
Background
Asbestos protective clothing was widely used in "hot" industries such as foundries, steel plants and glassworks, and by fire fighters. Undoubtedly the use of such clothing has saved many lives and made the working conditions of others more bearable. Use of asbestos protective clothing was considered acceptable in the UK until the late 1970s and, for example, it was not until 1976 that Scottish health civil servants advised the fire service of the possible hazards posed by asbestos equipment used by fire fighters [1]. At that time it was concluded that although the risks to health were minimal, fire brigades should phase out their use and find replacement gloves made from alternative materials.
Bamber and Butterworth [2] first published data on the airborne fibre exposure while wearing asbestos protective clothing. They carried out a laboratory study where personal monitoring was undertaken on a subject wearing an asbestos apron and gauntlets while carrying objects and doing bench work. The laboratory was well ventilated with approximately 10 air changes per hour extracted at ceiling level. The airborne fibre concentrations measured in their six tests ranged from 2.4 to 4.2 fibres/ml, with a mean airborne fibre concentration of 3.5 fibres/ml.
In a later study by Gibbs [3], personal airborne fibre exposure from wearing asbestos safety coats, hoods, gloves and leggings was measured for workers at two ore reduction plants. The men working on furnaces prepared channels along which molten iron flowed, tapped the furnaces and kept the channels clear of slag during casting. They wore asbestos safety coats for the duration of the castings (0.5 to 1.25 hours), which were repeated at approximately 4-hour intervals. Asbestos gloves and hoods were also worn at times during the work. A second survey was undertaken at a small plant manufacturing elemental phosphorous where four men carried out work similar to that at the steel works. Asbestos safety coats and leggings were worn throughout the slagging operations but hoods and gloves were worn only when necessary. Personal monitoring at both plants was only carried out from the start to finish of the slagging operations (15 – 47 minutes).
At the steel works, the mean airborne fibre concentration measured during the 39 personal monitoring tests was 2.0 fibres/ml, with a range from 0.3 to 5.0 fibres/ml, based on a mean sampling period of 52 minutes. The analysis of these data suggested that the fibre release increased with age of the garments up to 8 weeks, although the number of measurements was too small and the correlation coefficient too low to reliably predict fibre release from garments of different ages. At the phosphorous manufacturing plant, the mean personal airborne fibre concentration measured by personal sampling was 14 fibres/ml. This was based on 6 tests and a mean exposure period of 35 minutes. The measured airborne fibre concentrations ranged from 9.9 to 26 fibres/ml. The reason for the considerable differences in measured airborne fibre concentrations between the two plants was not known, although Gibbs suggests the higher levels were because the coats and mittens in the phosphorous plant were untreated (i.e., not aluminized outside or dust suppressed) and leggings were also quite badly damaged.
One other possible contributor to the differences in measured airborne fibre concentrations between the two plants was the room volume. Although the exact volume of each workroom is not given, the steel plant is described as large in comparison with the small phosphorous manufacturing plant. Damaged asbestos clothing could have contaminated the workplace and workers may then have disturbed this contamination. The lower airborne fibre concentrations were measured in the larger plant, which may be explained by the dilution effect of general ventilation being greater in large rooms in comparison with smaller ones [4].
Riediger [5] describes a combined controlled laboratory test and associated factory study of fibre release from asbestos clothing. He showed that in the laboratory tests asbestos cloth impregnated with a binder could produce airborne fibre concentrations that were approximately four times lower than those generated by untreated cloth. However, only three of the five treated asbestos cloths were effective and the other two samples produced higher fibre levels than any of the untreated materials. Heating impregnated asbestos cloth at 200°C reduced the effectiveness of the binder.
A study of occupational exposure to airborne fibres from the use of asbestos gloves was published by Samimi and Williams [6]. They investigated fibre exposure during the simulation of laboratory procedures in an unventilated isolation chamber and in a biology preparation room, as well as during actual work carried out by laboratory staff in two separate situations. The laboratory tasks where asbestos gloves were used comprised routine sterilization and the drying of laboratory glassware, both of which required the workers to put their hands in a hot autoclave or oven. The asbestos gloves were classified into four categories based on structural integrity and apparent surface cleanliness: well-worn & clean, well-worn & lightly soiled, well-worn & heavily soiled and brand-new. All gloves were of the same type. The experimenter carried out the simulation of work activities inside the isolation chamber after inserting his arms through two portholes in the front panel. The same sterilization operation was also simulated in a well-ventilated biology preparation room that had five air changes per hour. The interval between consecutive operations was either 30 minutes to represent the normal workload, or 10 minutes to represent a heavy workload. In the studies performed on workers in their actual workplaces, air samples were collected from the breathing zone of each worker and 75 cm above the tabletop where the gloves were laid or tossed.
The mean time weighted average (TWA) concentrations of airborne fibres from the 176 measurements in the isolation chamber, ranged from 0.95 to 12 fibres/ml. The minimum TWA concentration measured was 0.61 fibres/ml for well-worn and heavily soiled gloves. The maximum TWA concentration measured was 16 fibres/ml for well-worn and clean gloves. The results showed that clean well-worn gloves emitted significantly more fibres than did brand-new gloves, but fibre emission decreased with increased surface soiling.
Eighty air samples were collected during a simulation carried out in the well-ventilated laboratory. The range of mean TWA airborne fibre concentrations was 0.07 to 0.99 fibres/ml for the personal samples, and 0.06 to 0.60 fibres/ml for the static samples. These airborne concentrations were considerably lower than those obtained in the isolation chamber. This was due to the dispersion of fibres within the larger volume of the room, as well as the fact that the room was well ventilated when compared to the unventilated isolation chamber.
Thirteen samples were collected by Samimi and Williams in actual workplaces. The maximum and minimum TWA airborne fibre concentrations ranged from 0.07 to 2.93 fibres/ml for personal samples and from 0.04 to 0.74 fibres/ml for static samples (sampling over an 8-hour shift). With this limited number of samples, it was found that exposure levels depended more on the particular laboratory than on glove condition or workload, which were the main influencing factors under the controlled conditions of the simulation experiments. For example, when comparing the fibre exposures from the same task carried out with the same glove type, but at different laboratories, one was found to be 29 times higher than the other. This was explained by the presence of an efficient exhaust system over the row of five autoclaves in the laboratory where the lower exposures were measured.
As is clearly demonstrated by these studies, there is considerable room for debate over the level of fibre exposure from wearing asbestos protective clothing. Gibbs [3] in his study recorded a maximum airborne fibre concentration of 26 fibres/ml from personal sampling during slagging operations in a phosphorous manufacturing plant, whilst Bamber and Butterworth [2] and Samimi and Williams [6] in their studies measured airborne fibre levels which were generally between 1 and 5 fibres/ml.
Tougher legislation and greater awareness of the risks of working with asbestos have ensured that most organisations in Europe no longer use products containing asbestos, and many are in the process of eliminating all sources of asbestos from their work environments. However, as well as organisations becoming more aware of the risks of asbestos exposure, workers' knowledge of such risks has also increased. This increased awareness has resulted in more civil claims for compensation being made against employers for previous asbestos use. In many situations where asbestos exposure took place there is limited data on which to assess the likely airborne fibre concentrations from past working conditions. One approach to obtain more reliable information is to simulate work activities undertaken in the past and measure fibre exposure.
The aim of this study was to assess the personal exposure to airborne fibres arising from the use of chrysotile asbestos mitts worn in a glass manufacturing plant. This information was then used to assess the likely health risks to workers who had worn asbestos mitts.
Results
The airborne fibre sampling results from 33 personal samples collected during the simulation of the three different tasks under the various work conditions are summarised in Figure 1. In the figure each point represents the average fibre concentration measured during the activity. The mean personal airborne fibre concentrations for each test condition ranged from 0.03 to 0.35 fibre/ml for rowing, 0.05 to 0.48 fibres/ml for glass window repair and 0.09 to 0.47 fibres/ml for side seal replacement. The lowest average personal fibre concentrations were all obtained when high localised ventilation was used, whereas the maximum mean concentrations were measured for all three tasks when no ventilation was used. This suggests that the presence of localised ventilation substantially reduced the airborne fibre concentrations. This trend was also shown by both the 1 metre and 3 metre static area samples (data not presented in this paper). The measured airborne fibre concentrations for both the glass window repair and side seal replacement tasks were also generally higher than those for the rowing.
Figure 1 Fibre exposure levels during simulated work with asbestos mitts.
An analysis of variance was carried out on the data to investigate the differences between the three factors: type of gloves (unused versus aged), task and whether ventilation was used. The results showed the decrease in exposure levels when ventilation was used was highly statistically significant. Differences observed between the mean airborne fibre concentrations for the three simulation tasks were also highly significant (p < 0.01). Differences between mean airborne fibre concentrations for aged gloves and unused gloves were not significant. There were no significant interaction effects for the three factors (e.g. relative differences between tasks were similar whether or not ventilation was used).
The mean heart rate for each of the three simulation tasks ranged from 94 to 133 beats per minute. These data showed that the most strenuous task was the side seal replacement, which would be classed as "very heavy work" (estimated breathing rate 37.5 – 50 l/min). Both rowing and glass window repair were classed as "moderate" activities in terms of severity of workload (breathing rate 12.5 – 25 l/min). Glass window repair was the least strenuous activity with the lowest mean, maximum and minimum heart rate values.
Discussion
There was no difference between the measured airborne fibre concentration when unused or aged asbestos mitts were used to carry out the various tasks. This is contrary to what was observed in other studies and the most likely explanation is a combination of ineffective artificial ageing of the mitts and the abrasive nature of the tasks carried out in this study. Observations made during the tests showed that it was the abrasion of the mitts on sharp metal edges that resulted in obvious release of airborne dust and this would have applied equally to both types of glove.
We have shown that the use of a high level of localised ventilation significantly reduced the measured airborne fibre concentration when compared with the results from the same simulations carried out without any ventilation. The ventilation was designed to simulate the upwards flow of air produced by thermal convection from a hot glass float bath and the results show that the presence of hot work equipment would probably have reduced the workers exposure. From the present data the exposures for those working next to a glass float bath were likely to have been about a fifth of what they would have otherwise have been.
Examination of the results from the three different tasks shows that the side seal replacement and glass window repair tasks generally created similar airborne fibre concentrations (0.05 to 0.48 fibres/ml), whilst the rowing task produced lower airborne fibre concentrations (0.03 to 0.35 fibres/ml). These differences may be explained by the nature of the work. Both the side seal replacement and glass window repair required the subject to grip the object along narrow edges, some of which were sharp and this was seen to generate visible dust emission from the glove. The rowing task did not involve handling any sharp abrasive edges. The 2 m long steel pole used for rowing was cylindrical and had a smooth surface. Nevertheless, there was slight abrasion on the surface of the mitts when carrying out this task, as the hands had to rotate the steel pole in the hand whilst moving the pole backwards and forwards through the window.
There have been a few studies published on asbestos fibre exposure where asbestos protective clothing or mitts were the only source of exposure [2,3,5,6]. In previous research it was not been possible to determine the relative contribution of each item of asbestos protective clothing to the overall airborne fibre exposure. This makes it difficult to directly compare the results from these studies with the results from our research involving only asbestos mitts. However, it seems likely that much of the difference between these earlier studies and our simulations arises from the poor condition of the clothing worn in the workplace studies.
The only directly comparable study to ours is that published by Samimi and Williams [6]. This study investigated airborne fibre exposure in a biology preparation room and during work at two other laboratories. The range of mean airborne fibre concentrations was 0.07 to 0.99 fibres/ml for the personal samples. Thirteen samples were collected at the other workplaces where the minimum and maximum personal airborne fibre concentrations were 0.07 and 2.93 fibres/ml.
It is unclear why personal fibre exposures measured in the Samimi and Williams study were higher than those measured by us, as the work activities carried out in this earlier study were probably less damaging on the integrity of the gloves, when compared to the tasks carried out in our study. One reason may have been that the gloves used by Samimi and Williams show were generally in a poorer state in terms of structural integrity than the mitts that we used. Another reason that affects all of the historic studies is the poorer standard of quality assurance employed in studies carried out in the past compared with that used routinely today. This was highlighted in a paper published by Gibbs et al [7], which showed large intra-laboratory differences in fibre counting results. This lower standard of quality assurance in earlier work could possibly have resulted in either the over-estimation or under-estimation of fibre exposure in studies published at that time.
The results from our simulation study clearly show that tasks undertaken by glass furnace workers whilst wearing asbestos mitts would have resulted in asbestos fibres being released into the air. However, the contribution of wearing asbestos mitts to overall personal exposure to airborne asbestos fibre was probably quite low. There are a number of reasons for this assumption. Firstly, the simulation conditions that most accurately represented the actual workplace were those where high localised ventilation was used. This ventilation mimicked the upwards flow of air created by thermal convection next to a bath of molten glass, and if anything, was probably less than would normally be encountered. The results of personal monitoring for all three tasks under these conditions were very low, with the measured airborne fibre concentrations below the analytical detection limit of the method, for both the glass window repair and rowing, with the results for side seal replacement ranging from 0.09 to 0.14 fibres/ml.
Secondly, for the purposes of our study, the subject carried out each simulation over a 30-minute period, continuously repeating during that period one of the three tasks being investigated. This was necessary to generate an airborne fibre concentration that could be accurately evaluated by the analytical method. However, this regime of work was not employed in the past in the workplace. The furnace man would typically carry out one of the three tasks on an intermittent and random basis, whereas in the simulation the tasks were repeated about 100 times during the experiment. Although justified for the purposes of the study, the method has probably generated airborne fibre concentrations that were higher than those likely to be produced by wearing asbestos mitts in the workplace.
Thirdly, the relatively small volume of the experimental enclosure may have increased exposures over what would have occurred in the past. The dilution effect of general ventilation is usually greater in large areas in comparison with smaller areas and this could result in a lower measured airborne fibre concentration in the larger area next to the glass float bath, when all other influencing factors are the same in both the large and small areas. The experiments do not take account of exposure from residual asbestos contamination in a workplace resulting from fibre release from damaged clothing, but this type of secondary source would generally be small in comparison with direct emission. Other sources might however predominate in some circumstances, e.g. where other asbestos-containing materials were disturbed.
It seems unlikely that glass float furnace men wearing chrysotile asbestos mitts would have been exposed to respirable concentrations of asbestos above the present UK control limit for chrysotile. During the 1960's and 70's the standards that were accepted by the scientific community were higher than currently applied, although they mostly reflected concerns about non-malignant disease. In 1968 the British Occupational Hygiene Society published an internationally recognised hygiene standard for chrysotile asbestos dust [8,9]. This standard implied that the risk of having early clinical signs of asbestos-related disease would be less than one percent for 50 years exposure at 2 fibres/ml. Even working continuously carrying out the dustiest activity, workers wearing asbestos mitts in a glass manufacturing plant could never have received such an exposure.
Other workers who wore asbestos gloves or mitts, e.g. firemen or laboratory workers, would probably have had higher exposure than the glass workers because they would not necessarily have been working close to high convective airflows. In these situations the measurements we have made without ventilation might provide the best estimate of exposure level from wearing asbestos mitts, i.e. about 0.5 fibres/ml while the gloves were worn (see Figure 1).
Hodgson and Darnton from the UK Health and Safety Executive have carried out an extensive review of epidemiological studies that inform the quantitative link between cancer risks and asbestos exposure [10]. They provide mathematical models linking cumulative exposure to asbestos with both lung cancer and mesothelioma; in both cases the models are non-linear functions dependent on the cumulative inhalation exposure to asbestos, although for mesothelioma the risk is calculated separately for pleural and peritoneal tumours. The risk of mesothelioma increases as the time since first exposure increases and Hodgson and Darnton allow for this by using age-related adjustment factors.
We have used these models to estimate the risks for a glass worker aged 20 when first employed (in 1955) who worked for 20 years using chrysotile mitts. Assuming he was exposed for 10 minutes at the estimated 90th percentile for each task each day his annual average exposure would have been 0.012 fibres/ml. In this calculation we have weighted the side seal replacement task three times more than the others to account for the higher breathing rate during this work. Based on these assumptions the best estimate of his lifetime risk of mesothelioma is around 3 in 100,000 and the risk of lung cancer is less than one per 100,000. The "highest arguable" risks (as defined by Hodgson and Darnton) were 16 per 100,000 for mesothelioma and 6 per 100,000 for lung cancer, which would equate to a total annual risk of about 3.7 per million. Even these highest estimates are around the risks that most would consider trivial, i.e. around 1 in a million per year.
The estimates are prone to uncertainty because of the processes involved in estimating the actual exposure of someone wearing asbestos mitts, from the analysis used to quantify the association between exposure and cancer risk and from the necessity to extrapolate this relationship to low exposures, certainly lower than most asbestos workers would have experienced in the past. However, despite this we believe that our measurements show that wearing asbestos mitts would have given rise to relatively low cumulative exposures to chrysotile asbestos, and taking account of the possible uncertainties in the process the risk of death from cancer from such exposures must be low; we believe trivially low.
Conclusion
In the past protective mittens made from chrysotile asbestos were commonly used in glass manufacturing and fibres were released from asbestos mitts while they were being worn. During simulated work activities the airborne concentration in the workers breathing zone did not exceed 0.5 fibres/ml. Lower concentrations were measured in environmental conditions designed to reproduce high localised convective airflows found in glass plants. The lifetime risk of a worker contracting mesothioma or lung cancer from 20 years of past use of asbestos mitts in the glass industry was estimated to be 22 per 100,000, which is very low.
Methods
The tests were carried out in an asbestos enclosure designed and constructed to the standard recommended in the Health and Safety Executive [11]. The size of this enclosure was approximately 45 m3, with dimensions 5 m × 3 m × 3 m. Extract ventilation was provided by a fan and a high-efficiency particle arrester (HEPA) filter. Air was extracted from the enclosure at ceiling level through a canopy, located directly above the workstation, and carried via flexible ducting to the extraction unit where the air was filtered. Again using flexible ducting, the filtered air was carried back into the enclosure and discharged upwards from an elevated 1 m × 1 m platform, upon which the test subject stood.
The workstation in the enclosure was to be used to simulate activities carried out close to a hot glass tank where upward convective airflows are found. The extraction system was designed to provide an upward air velocity of approximately 3 to 4 m/s. It was not possible during the course of this study to measure the upward convective airflow next to a hot glass tank in a glass works, although it was assumed that such airflow could be quite high.
This study investigated the three most common tasks where asbestos mittens were reportedly used in glass production plants. These tasks were:
1. "Rowing" of molten glass. This task was simulated using a window taken from a float bath and an approximately 2 m long steel pole that was used to move glass in a bath. Half of the steel pole was placed through the window situated at chest height, with the other half being gripped firmly and moved in a 'rowing' motion.
2. Removal and replacement of a glass window in a float bath. This task required the loosening of clamps holding the window in its frame, removing the window, setting it aside and then replacing it.
3. Removal and replacement of a side seal in a float bath. This task was simulated by lifting the side seal from the floor next to the workstation, pushing it into an orifice at chest height, and then removing it and placing it back on the floor.
Each of the three tasks was simulated for a period of 30 minutes. Throughout the test the subject's heart rate was monitored continuously using a "Polar Sport Tester" heart rate monitor, which recorded the subjects heart rate at 60 second intervals. This data enabled the subjects breathing rate to be estimated [12]. These data were used to make an assessment of breathing rate during the tasks.
A glass company provided two pairs of chrysotile asbestos gloves, made in the 1970's, for the exercise. One pair were unused and still in the original packaging, whilst the second pair had either been unused or had had very light usage. The tests were undertaken separately with the unused mitts and with the second pair artificially aged by heating the gloves for 20 hours at 100°C followed by hammering them in sealed packaging for five minutes inside an enclosed glove box.
The measurement of both personal fibre exposure and airborne fibre concentrations within the test room were made [13]. Two personal samples were collected for each simulation exercise. The sampling heads were positioned in the test subject's breathing zone, i.e. within approximately 200 mm of the nose and mouth, one on each side of the head. Prior to sampling the flow rate was set at 2.0 l/min, and this was checked both during and after the sampling period using a calibrated flow meter. In addition to the personal sampling, two static room samples were collected during each simulation exercise, one in the test subject's near-field (i.e. within 1 m of the breathing zone) and another in the far-field (i.e. approximately 3 m distant from the workstation). Both static samples were collected as described for the personal samples, except the sampling flow rate was set at 8.0 l/min. Both the 1 m and 3 m sampling heads were situated approximately 1.5 m from the ground. The results from the static samples are not presented in this paper.
Twelve separate tests were undertaken during the simulation exercise involving various combinations of the three tasks, the two types of mitt (used and unused) and two ventilation conditions (none and high). In addition, five of the tests were carried out twice to assess the repeatability of the measurements. During all of the tests the subject wore a high efficiency positive pressure respirator and protective clothing.
After each simulation the test enclosure was thoroughly cleaned using a high efficiency vacuum cleaner. Wet wipes were used to remove all traces of asbestos dust or debris produced during the simulation exercise. Air monitoring was undertaken after the cleaning of the test enclosure to ensure that the airborne fibre concentration was below 0.010 fibres/ml.
All the membrane filters were analysed using procedures complying with the United Kingdom Accreditation Service (UKAS) by an experienced analyst using the HSE method MDHS 39/4 [13].
Statistical analysis of the data was undertaken using a general linear model to assess the significance of the three different factors which were expected to influence the measured airborne fibre concentrations: glove type (i.e. unused or aged), ventilation condition (i.e. none or high localised ventilation) and simulation task (i.e. Rowing, Glass window repair or Side seal replacement). This approach was used because only five of the twelve tests were repeated. This produced an unbalanced data set that did not allow a simple analysis. The three factors were analysed using an Analysis of Variance test, to test all the factors both independently from each other, as well as for any interactions that were taking place between them.
Cancer risks from asbestos exposure were estimated using the method described by Hodgson and Darnton [10]. The percent excess mortality from mesothelioma (Pm) was estimated using equation 1.
Pm = AplXr + AprXt (equation 1)
where Apl and Apr are constants of proportionality for pleural and peritoneal risks, X is the cumulative exposure (in fibres/ml.years) and r and t are the slopes of the exposure-response on log-log scales.
Two sets of coefficients were used: one for the "best estimate" of risk and the other for the "highest arguable" risk estimates. An adjustment factor was used to allow for the age at which the person was first exposed, as described by Hodgson and Darnton.
Lung cancer excess percent excess mortality was similarly estimated using equation 2.
PL = ALXr (equation 2)
where AL is the lung cancer constant and r is the slope of the exposure response on log-log scales (note the coefficient r in this equation is different from that in equation 1).
These estimates are based on British male mortality in 1997 when 9.5% of deaths were from lung cancer. The predictions therefore represent the past smoking prevalence of older men. Non-smokers would have substantially lower predicted risks and lifetime smokers would have lung cancer risks about double those quoted.
Calculations were undertaken using an EXCEL spreadsheet supplied by the authors (Hodgson, personal communication).
Competing interests
JWC and HC prepare reports in connection with various asbestos civil litigation cases, some of which involve asbestos gloves or mitts.
Authors' contributions
MT undertook the experimental work and assisted in the preparation of the manuscript. HC carried out the statistical analysis and the risk assessments. JWC conceived the project, supervised the work and prepared the manuscript. All authors read and approved the final text of the manuscript.
Acknowledgements
We are grateful to Drs Robert Aitken and Sean Semple for comments on our paper, and to the many colleagues at the Institute of Occupational Medicine who provided advice and assistance during the study. This work was partly supported by funding from Pilkington UK Ltd.
==== Refs
Harrison D Hazards posed by asbestos Fire Service Circular 31/1976 1976 Edinburgh: Scottish Home and Health Department
Bamber HA Butterworth R Asbestos hazard from protective clothing Ann Occup Hyg 1970 13 77 79 5412916
Gibbs GW Fibre release from asbestos garments Ann Occup Hyg 1975 18 143 149 1190649
Cherrie JW The effect of room size and general ventilation on the relationship between near and far-field concentrations App Env Occup Hyg 1999 14 539 546 10.1080/104732299302530
Riediger G Vetements de protection en amiante. Degagement de poussieres lors de la fabrication et de l'emploi Cashiers de Notes Documentaires 1979 96 425 433 HSE Translation 12620D
Samimi BS Williams AM Occupational exposure to asbestos fibres resulting from use of asbestos gloves Am Ind Hyg Assoc J 1981 42 870 875 7315743
Gibbs GW Baron P Beckett ST Dillen R Du Toit RSJ Koponen M Robock K A summary of asbestos fibre counting experience in seven countries Ann Occup Hyg 1977 20 321 332 610506
Lane RE Gilson JC Roach SA Smith S Addingley CG Holmes S Hunt R Knox JF King E Hygiene standard for chrysotile asbestos Ann Occup Hyg 1968 11 47 69 5654330
Ogden TL Commentary: The 1968 BOHS Chrysotile Asbestos Standard Ann Occup Hyg 2003 47 3 6 12505901 10.1093/annhyg/meg011
Hodgson JT Darnton A The quantitative risks of mesothelioma and lung cancer in relation to asbestos exposure Ann Occup Hyg 2000 44 565 601 11108782 10.1016/S0003-4878(00)00045-4
Health and Safety Executive Guidance Note EH 51 – Enclosures provided for work with asbestos insulation, coatings and insulating board 1989 London: HMSO
Rodahl K The Physiology of Work 1989 London: Taylor & Francis
Health and Safety Executive Methods for the Determination of Hazardous Substances (MDHS) 39/4, Asbestos fibres in air – Sampling and evaluation by Phase Contrast Microscopy (PCM) under the Control of Asbestos at Work Regulations 1995 London: HMSO
|
16202137
|
PMC1262768
|
CC BY
|
2021-01-04 16:38:30
|
no
|
Part Fibre Toxicol. 2005 Oct 3; 2:5
|
utf-8
|
Part Fibre Toxicol
| 2,005 |
10.1186/1743-8977-2-5
|
oa_comm
|
==== Front
Part Fibre ToxicolParticle and Fibre Toxicology1743-8977BioMed Central London 1743-8977-2-61620215410.1186/1743-8977-2-6ResearchCombustion of dried animal dung as biofuel results in the generation of highly redox active fine particulates Mudway Ian S [email protected] Sean T [email protected] Chandra [email protected] Gazala [email protected] Frank J [email protected] Jonathan [email protected] Lung Biology: Pharmaceutical Science Research Division, School of Biomedical & Health Sciences, King's College London, Franklin-Wilkins Building, 150 Stamford Street, London, SE1 9NH, UK2 Department of Chemical Engineering, Indian Institute of Technology, Bombay, Powai, Mumbai-400 076, India3 Division of Child Health, Department of Immunology, Infection and Immunity, University of Leicester, PO Box 65, Leicester2005 4 10 2005 2 6 6 7 2 2005 4 10 2005 Copyright © 2005 Mudway et al; licensee BioMed Central Ltd.2005Mudway et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The burning of biomass in the developing world for heating and cooking results in high indoor particle concentrations. Long-term exposure to airborne particulate matter (PM) has been associated with increased rates of acute respiratory infections, chronic obstructive lung disease and cancer. In this study we determined the oxidative activity of combustion particles derived from the biomass fuel dung cake by examining their capacity to deplete antioxidants from a model human respiratory tract lining fluid (RTLF). For comparison, the observed oxidative activity was compared with that of particles derived from industrial and vehicular sources.
Results
Incubation of the dung cake particle suspensions in the RTLF for 4 h resulted in a mean loss of ascorbate of 72.1 ± 0.7 and 89.7 ± 2.5% at 50 and 100 μg/ml, respectively. Reduced glutathione was depleted by 49.6 ± 4.3 and 63.5 ± 22.4% under the same conditions. The capacity of these samples to deplete ascorbate was in excess of that observed with diesel or gasoline particles, but comparable to that seen with residual oil fly ash and considerably in excess of all three control particles in terms of glutathione depletion. Co-incubation with the metal chelator diethylenetriaminepentaacetate inhibited these losses, whilst minimal inhibition was seen with superoxide dismutase and catalase treatment. The majority of the activity observed appeared to be contained within aqueous particle extracts.
Conclusion
These data demonstrate that biomass derived particles have considerable oxidative activity, largely attributable to their transition metal content.
==== Body
Background
Approximately two billion people in the developing world use biomass fuels such as wood, crop-waste and dried animal dung, as their major source of domestic energy [1,2]. The burning of this material in open fires and stoves results in high concentrations of particulate matter (PM), carbon monoxide, nitrogen dioxide, as well as volatile and semi-volatile organic species in the indoor environment [3,4]. In homes where biomass burning occurs airborne particle concentrations are far in excess of those found in homes where it is not used as a heating/cooking source, with 24 h average PM10 concentrations ranging between 200 and 5000 μg/m3 depending on the fuel type, stove and ventilation [5,6]. These concentrations are far in excess of the level considered safe for PM10 in outdoor air – 150 μg/m3 24 h average [7].
An increasing body of evidence has linked exposure to indoor pollutants with increased rates of respiratory morbidity and mortality. Women and children exposed to high indoor PM concentrations have significantly increased rates of acute respiratory infections (ARI) [5,8-10], and women cooking over biomass fires for extensive periods have an enhanced risk of chronic obstructive pulmonary disease (COPD) and lung cancer [11,12]. In a recent study we found that lower airway cells from women and children exposed to biomass smoke contained significantly more carbonaceous material than age-matched subjects exposed to fossil fuel derived PM10 [13]. Whilst enhanced particle deposition in the airways is clearly important, the chemical and physical characteristics of airborne particles that contribute to their toxicity have not been firmly established, though particle size and surface area [14], as well as acidity and composition [15,16] have all been mooted as important determinants in this regard.
The capacity of inhaled PM to elicit damaging oxidation reactions in the lung and systemic circulation may account for many of the toxic responses observed in exposed individuals [17]. Certain transition metal components of PM are capable of catalysing oxidation reactions [18], as are quinones [19] resulting in the production of reactive oxygen species. An excessive production of these species can activate redox sensitive transcription factors regulating the expression of pro-inflammatory cytokines, leading to inflammation and tissue injury [20]. Measuring the capacity of PM to oxidise physiologically relevant molecules can therefore be used to provide a biologically relevant index of activity, integrating PM size, surface area and composition.
In this study we evaluated the oxidative activity of PM samples derived from the controlled burning of cow-dung cake using a traditional Indian cooking stove. These assessments were based on the capacity of dung cake particulate suspensions to deplete physiologically relevant antioxidants, ascorbate (AA), urate (UA) and reduced glutathione (GSH) from a synthetic model of human respiratory tract lining fluid. Previously we have used this model to investigate the nature of gaseous and particulate pollutant-antioxidant interactions at the air lung interface [21,22]. In addition we examined the extent to which any observed activity could be attributed to the metal or organic content of the dung cake particles.
Results
Three Teflon filters were supplied containing 0.86 (DC1), 1.20 (DC2) and 1.14 mg (DC3) of PM2.5. Particle suspensions prepared from each of the three filters depleted AA from the synthetic RTLF in a dose dependent manner with a mean loss of 72.1 ± 0.72 and 89.7 ± 3.7% AA at 50 and 100 μg/m3 respectively relative to the 4 h particle free control concentration. This loss of AA was significantly greater than that associated with an equal mass (50 μg/m3) of diesel (6.4%), gasoline (5.2%) or residual oil fly ash (ROFA) (53.9%) particles (Figure 1). All of the tested PM samples were PM2.5, apart from the diesel and gasoline samples that were collected using a high volume cascade impactor as PM0.1–2.5. Of these particles ROFA had the smallest median diameter, 0.07–0.08 μm [23], with diesel and gasoline PM samples in the range 0.1–0.2 μm (Thomas Sandström, personal communication), with the dung cake samples displaying the largest median aerodynamic diameter, 0.6–0.8 μm [4]. These data further emphasise the reactivity of the DC samples, taking into account their lower surface area per unit mass, compared with the control PM samples. These AA losses were completely inhibited when the dung cake particles were co-incubated with 100 μM DTPA (Figure 1).
Figure 1 AA remaining in synthetic RTLF following a 4 h incubation with three separate dung cake (DC) samples at 50 and 100 μg/ml (grey bars)The impact of co-incubation with the transition metal chelator DTPA (200 μM) is illustrated by the white bars. Pre (0 hr) and post (4 hr) AA concentrations in the particle-free controls are illustrated in the black filled bars. The losses observed with 50 μg/ml doses of fresh diesel and gasoline PM0.1–2.5, as well as ROFA are also illustrated. Data represent the mean (SD) of 3 separate experiments: 'a' – indicates that AA concentrations after the 4 h incubation were significantly lower (P < 0.05) than the 4 h particle free control values; 'b' illustrates that the losses observed at 100 μg/ml are significantly greater than those seen at the lower 50 μg/ml concentration. The '*' illustrates a significant loss of AA in the particle free control over the 4 h incubation period.
GSH was also depleted in a dose dependent manner by the dung cake PM extracts, 49.6 ± 4.3 and 63.5 ± 22.4% at 50 and 100 μg/m3 respectively. These losses were significantly greater than those observed with ROFA, which actually demonstrated a reduced loss of GSH compared with the particle free control, or the traffic derived samples, which depleted GSH 10.2% (diesel) and 15.7% (gasoline) (Figure 2). Notably, no loss in total glutathione was noted during the incubations suggesting that adsorption to the particle surface was not occurring to any great extent. Similar to the situation with AA, co-incubation with DTPA completely abolished the loss of GSH. Incubation with DTPA also prevented the background losses of both ascorbate (-20.8 ± 6.2 μM – Figure 1) and glutathione (-47.4 ± 6.1 μM – Figure 2) from the particle-free controls over the 4 h incubation. These losses were attributable to the presence of contaminating metal ions in the synthetic RTLF following Chelex-resin treatment. Urate was not depleted from the RTLF model by any of the particle types examined in line with our previous observations [21] (data not shown).
Figure 2 Total (GSx) and reduced glutathione (GSH) remaining in synthetic RTLF following a 4 h incubation with three separate dung cake (DC) samples at 50 and 100 μg/ml. Reduced glutathione concentrations after 4 h are illustrated by dark grey bars, whilst the corresponding total glutathione concentration (GSx) are shown in light grey beneath the corresponding GSH data. The impact of co-incubation with the transition metal chelator DTPA (200 μM) is illustrated by the white bars, otherwise the figure is formatted as outlined in the legend to figure 1 with the following amendments. Notably DTPA incubation prevented not only the PM-induced loss of GSH but the background auto-oxidation seen in the particle-free controls. 'a" indicates that the concentration of GSH remaining after 4 h following co-incubation with DTPA was significantly greater than that in the 4 h particle-free control.
Further characterisation of the metal dependence of the oxidative reactions observed was performed using an AA only RTLF model. Individual AA consumption rates at 50 μg/mL for each of the three dung cake extracts were similar: 13.7 ± 0.15 (DC1), 13.5 ± 0.49 (DC2) and 13.3 ± 0.19 nM/s (DC3). Co-incubation with the metal chelator DTPA (100 μM) significantly reduced the rate of AA oxidation: 6.5 ± 0.36 versus 13.5 ± 0.27 nM/s for all three samples (Figure 3). Notably, in this simplified model in the absence of GSH and UA, the protection afforded by DTPA was only approximately half of that seen using the synthetic RTLF. Increasing the DTPA concentration to 200 μM completely blocked AA depletion in this model (Figure 4). Limited, though statistically significant protection was also seen using the antioxidant enzymes superoxide dismutase and catalase (11.9%), whilst no decrease in AA oxidation was seen with catalase alone or heat inactivated superoxide dismutase and catalase (Figure 3).
Figure 3 Rate of ascorbate consumption observed from a simple ascorbate-only solution with and without a range of metal chelation and free radical scavenger treatments. All data represent the mean (SD) of three separate experiments: AOX = particle free control; DC = dung cake samples (1–3 at 50 μg/ml); SOD/CAT = superoxide dismutase (50 U/mL) and catalase (150 U/mL); hiSOD/CAT = heat inactivated antioxidant enzymes; CAT = catalase only (150 U/mL), and DTPA (100 μM). 'a' indicates that the rate of ascorbate consumption was significantly reduced (P < 0.05) after SOD/CAT and DTPA treatment; 'b' that the DTPA treatment reduced the rate significantly more than the antioxidant enzyme treatment.
Figure 4 Time dependent loss of ascorbate associated with incubation of DC at 50 μg/ml at varying concentrations of DTPA. Each trace represents the mean of three separate experiments. The SDs on each mean value are not illustrated for graphical simplicity but were less than 5% of the mean values in all cases.
When the dung cake aqueous and organic extracts were separated the vast majority (80.6%) of the oxidative activity was found to be associated with water-soluble components (Figure 5). The residual activity associated with the hexane extract could be completely inhibited through the co-incubation of 100 μM DTPA indicating that this activity was not due to organic radicals. Consistent with these findings, dung cake aqueous extracts were shown to contain appreciable concentrations of the redox active metals Fe and Cu quantified using the chromogenic chelators bathophenantroline-disulphonate (BPS) and bathocuproine-disulphonate acid (BCS) in each of the three dung cake particle suspensions. These methods gave Fe concentrations of 7.1, 3.6, and 8.5 μg/mg in DC 1, 2 and 3 respectively and 26.2, 26.2 and 23.8 μg/mg for Cu. In comparison ROFA contained 23.7 ± 0.30 μg/mg Fe and 5.9 ± 0.73 μg/mg of Cu, whilst both metals were undetectable in the diesel and gasoline samples, using either chromogenic chelator.
Figure 5 Rates of ascorbate consumption seen with aqueous and organic extracts of dung cake (DC) particle extracts. All data represent the mean (SD) of three separate experiments. '*' Indicates that the rate of AA depletion seen with the complete, aqueous and organic dung cake extracts was significantly greater (P < 0.05) than that observed in the particle free control; 'a' that the activity of the aqueous and organic extracts were significantly less than the complete extract; and, 'b' that the rate of AA depletion by the organic extract, with or without DTPA, was less than that seen with the aqueous fraction.
Discussion
Indoor air pollution poses a significant health risk worldwide. WHO estimates suggest that up to 6.5% of the annual disease burden in developing nations is attributable to the burning of solid fuels in the indoor environment [24,25]. Smoke from cooking stoves burning biomass fuels contains carbon monoxide, fine particulates, nitrogen dioxide and hydrocarbons; all at concentrations far in excess of what is considered unsafe in outdoor air [7]. In this study we investigated whether fine particles derived from the burning of the biofuel dung cake also displayed high levels of intrinsic oxidative activity relative to traffic and industrial derived PM. We wanted to examine the hypothesis that the health effects associated with exposure to biofuel derived PM were not solely a function of the high exposure concentrations but also because of their high content of redox active components.
It has been proposed that the capacity of inhaled particles to elicit inflammation and injury in the lung, as well as systemically, may be related to their capacity to cause oxidative stress [17]. In this working paradigm, inhaled particles generate oxidative stress through three inter-related pathways: first, by directly introducing oxidising species into the lung, such as redox active transition metals [16] or quinones [19] absorbed onto their surface. Second, by introducing surface absorbed PAHs that can undergo bio-transformation in vivo into quinones species through the action of the cytochrome P450, epoxide hydrolase and dihydrodial dehydrogenase detoxification pathway [26] and finally by stimulating inflammatory cells to undergo the oxidative burst. In this final case, activation of inflammatory cells may be triggered by endotoxin on the surface of inhaled particles [27], futile phagocytic processing of PM [28], or by the up-regulation of redox sensitive transcription factors directing the synthesis of pro-inflammatory cytokines [29]. The integrated sum of all these processes can be considered the 'total' oxidative activity of the particle.
In this study we measured PM oxidative activity using an in vitro screening procedure that assessed the capacity of PM associated pro-oxidant components (metals and quinones) to deplete physiologically relevant antioxidants, ascorbate, urate and reduced glutathione from a synthetic model of the RTLF [21,22]. This 'intrinsic' activity, measured in a cell free system, only reflects the oxidative activity attributable to redox active metals and quinone compounds and not 'latent' activities that may be associated with PAHs or endotoxin. With this caveat, we found PM samples derived from the combustion of dung cake to be significantly more active, on an equal mass basis, than either metal-rich ROFA or PAH-rich vehicle exhaust PM, despite the greater surface area of these samples. This activity was manifest by the capacity of the PM suspensions to deplete both AA and GSH from synthetic RTLF. Notably, the endotoxin and PAH content of dung cake and other biofuels have been shown to be high [24,30] suggesting that their 'total' oxidative activity is likely to be far in excess of that associated with traffic derived PM. This very high oxidative activity in animal dung combustion particles supports studies demonstrating increased pulmonary toxicity in mice following instillation of particles derived from dried municipal sewage combustion, relative to coal alone [31].
The depletion of both AA and GSH from the model was prevented by co-incubation with the metal chelator diethylenetriaminepentaacetate (DTPA) indicating that the losses observed were driven by redox active metals such as Fe, Cu, Ni, and Cr. DTPA has five acetate groups linked to a molecular backbone that permits it to form tight complexes with a broad range of metals, preventing them from catalysing damaging oxidation reactions. Desferoxamine (DFO) was not used in these studies as it has been reported to reduce chelated Cu that would have resulted in interpretive difficulties when examining its protective role in mixtures of soluble metals [32]. The contribution of superoxide and hydrogen peroxide to the observed antioxidant losses was examined in co-incubation experiments using the antioxidant enzymes Cu, Zn superoxide dismutase (SOD) and catalase (CAT). Limited protection was observed with these enzymatic antioxidants suggesting that the contribution of these reactive oxygen species to the ascorbate and glutathione losses was minor compared with those attributable to their direct oxidation during the reduction of Fe3+ and Cu2+.
Thus we conclude that AA and GSH oxidation occurred predominately by their direct reduction of Fe3+ to Fe2+ and Cu2+ to Cu+. The superoxide formed by the subsequent oxidation of ferric and cupric ions could undergo dismutation to hydrogen peroxide, reduce Fe3+ and Cu2+, or oxidise ascorbate, urate or glutathione within the synthetic RTLF. As the reaction rate between Fe3+ and superoxide (1.5 × 108 M-1s-1) greatly exceeds that its dismutation at physiological pH (5.4 × 105 M-1s-1, pH7.4) it seems likely that the former reaction predominated, especially as little loss of ascorbate or glutathione could be attributed to superoxide or hydrogen peroxide production. Interestingly, we saw no evidence of urate depletion, despite the importance of this antioxidant in protecting the airway against oxidant gases [33], peroxynitrite [34] and hydroxyl radicals [35]. The rate of the reaction of urate with hydroxyl radicals (7.2 × 109 M-1s-1, pH 6–7) is broadly similar to that of both ascorbate (1.6 × 109 – 1.1 × 1010 M-1s-1, pH 7–7.4) and glutathione (9.0 × 109 – 1.3 × 1010 M-1s-1, pH 8 and 7.8, respectively). Thus the absence of UA depletion in this model supported the contention that superoxide dismutation to hydrogen peroxide was not occurring to any great extent, with little evidence of hydroxyl radical generation. Whilst the redox potentials of UA and AA at pH7 (E°' = 590 and 282 mV respectively) [36] may suggest that the urate radical could be reduced back to UA at the expense of AA we do not believe that this occurred, as removal of urate from the RTLF had no impact on the observed rate of ascorbate depletion (data not shown).
We also observed that the capacity of DTPA to inhibit ascorbate oxidation was significantly reduced in the ascorbate only incubation experiment: only 100 μM being required for full inhibition in the complete synthetic RTLF as opposed to 200 μM in the ascorbate only RTLF model. This finding may imply that either GSH or UA is limiting the bioavailability of Fe, either through chelation, as has been proposed for UA [37], or by interfering with the capacity of AA to solubilise ferric iron from the particle surface [16]. We are currently investigating these potential actions of UA and GSH. Irrespective of this, when the concentration of DTPA was increased in the AA only model, all AA oxidation was blocked indicating the absence of a quinone-dependent activity in the DC particles. This contention was supported by the observation that only a fraction of the measured oxidative activity was present in organic DC extracts. Whilst other groups have emphasised the importance of PM associated quinones/hydroquinones in the oxidative activity of ambient PM [19] our data would tend to emphasise metal content as the major determinant of DC oxidative activity. Clearly the contribution of metal and organic components to PM oxidative activity may vary depending on its source. In addition it is also likely that the age and storage conditions of the filters used in this study resulted in losses of potentially reactive organic species. These cautionary caveats only further emphasise that we are probably underestimating the 'true' oxidative capacity of freshly generated DC particulates.
As these findings implicated redox active metals in the oxidation process we measured the bioavailable Fe and Cu content of the DC particles. This measurement included reduced and oxidised forms of these metals, both water-soluble and surface mobilisable through ligation to the chromogenic chelators bathophenantroline-disulphonate (BPS) [38] and bathocuproine-disulphonate acid (BCS) [39]. Using these approaches we detected considerable Fe and Cu content. The especially high content of Cu is likely to explain extensive glutathione oxidation observed with the dung cake samples, due to copper's high reactivity toward this antioxidant [18], as well as the lack of reactivity of the ROFA sample toward GSH. It should be noted, however, that the variation in the content of these metals in the three DC samples did not match their observed variation in oxidative activity implying that Fe and Cu were not the sole determinates of the observed activity. The compositional data pertaining to the ROFA sample (PM2.5), used in the current study have been described previously [23,40] using ICP-MS. These analyses have confirmed the relatively high concentrations of total Fe, and low concentrations of Cu in the ROFA sample derived from the burning of heavy fuel oil (N° 5). These metals are however less abundant in the ROFA sample than either vanadium (58.6 μg/g) or nickel (10.6 μg/g). In contrast, ICP-MS analysis of both the gasoline and diesel samples revealed these metals to be below detectable limits [41], which concurs with their low reactivity in this assay system and these measurements made using the chromogenic chelators. Parallel ICP-MS analysis of dung cake PM obtained under identical burn conditions has also subsequently revealed appreciable concentrations of the redox active metals Ni and Cr (45 and 40 μg/g PM, respectively) in these samples, but no detectable V. This elemental analysis will be described in detail in a subsequent manuscript.
The high metal content of the dung cake may reflect both the presence of biological metals; especially Fe and Cu in dung [42], as well as metals associated with the local soil with which the animal dung is mixed to make the bricklets. These data therefore support the initial hypothesis that fine particles derived from the controlled burning of dung cake are highly oxidative in nature due to their content of redox active metals.
Conclusion
These data demonstrate that fine particles derived from the burning of the biomass fuel dung cake are highly oxidising. This activity appears to be largely related to their content of redox active metals, with the caveats mention above. As oxidative activity is a biologically meaningful index, the high activity associated with dung cake fine particles allied to their high concentrations in homes using dung cake fuel emphasises their potential negative health impact.
Methods
Sample Collection and Filter extraction
Dung-cake was obtained from Eksaal, near Mumbai. Dung cake has a high ash content, 25–30% and is made by mixing dung with mud/clay and straw to improve its mechanical strength. It is then patted down, moulded and sun-dried for 24 h, which achieves moisture content of 7–8%. As dung-cake combustion results in copious particle emissions a sampling rate of 3–4 L per minute was used in the particle sampler, and a burn time of 12 min, to ensure no clogging of the Teflon filter substrates used for particle collection. The particle sampler employed had a cyclone inlet to exclude particles larger than 2.6 μm diameter [4]. All Teflon filters were conditioned for 12-h at 50% RH and 25°C prior to weighing. Filters were shipped to the UK where the PM was extracted from the Teflon matrix using a standardized vortexing and sonication protocol into Chelex-treated water contain 5% methanol, pH 7.0 [21]. All subsequent dilutions of this particle solution were also performed in Chelex-treated water contain 5% methanol, pH 7.0. Diesel and gasoline particles in the fine mode (0.1–2.5 μm) were obtained from Professor Thomas Sandstrom (University Hospital, Umeå, Sweden) derived from idling diesel (Volvo TD45, 4.5 L, 4cylinders, 1991) and gasoline (2.0 L "Opel Omega" with a catalyst from 1989) engines. Details of the diesel fuel used have been published previously [43] whilst the gasoline engine utilized 95-octane gasoline. Residual fly ash was kindly donated by Professor Ken Donaldson (University of Edinburgh, Scotland, UK). Diesel, gasoline and ROFA PM samples were resuspended in Chelex-treated water contain 5% methanol, pH 7.0 at the required concentration and the pH of the resultant suspension adjusted to neutral pH.
Particle Incubations and Measurement of Antioxidant Loss
For the initial experiments dung cake extracts were diluted to either 100 or 50 μg/ml and incubated at 37°C in synthetic RTLF (200 μM, AA, UA and GSH, pH7.0) for a period of 4 h. After incubation, samples were either acidified with metaphosphoric acid to a final concentration of 5% (w/v) for UA and AA measurement or diluted into 100 mM phosphate buffer for determination of GSH. Antioxidant determinations have been described previously [21]. Subsequent inhibitor studies were performed at PM concentrations of 50 μg/ml by following the loss of AA (starting concentration 200 μM, pH7.0, 37°C) at 265 nm over a 4 h time course with readings every 5 minutes. Incubations were performed with dung cake extracts only, as well as in the presence of a range of free radical scavengers and transition metal chelators: SOD/CAT = superoxide dismutase (50 U/mL) and catalase (150 U/mL); hiSOD/CAT = heat inactivated antioxidant enzymes; CAT = catalase only (150 U/mL), and DTPA = Diethylenetriamine pentaacetate (200 μM).
Hexane extractions
1 ml of each dung cake particle suspension at 55.6 μg/ml was vortexed hard for 2 minutes with an equal volume of HPLC grade hexane. At the end of this mixing period the organic and aqueous phases were separated, and the former dried under nitrogen at room temperature. The hexane extract was then resuspended according to the standard protocol in an equal volume of Chelex-treated water containing 5% methanol.
Total Fe and Cu determinations
Total iron concentrations in the particle suspensions were determined using the Fe2+-specfic chromogenic chelator, bathophenantroline disulphonate (BPS) [38]. DC particle suspensions (55.6 μg/ml) were prepared in Chelex-100 resin treated ultra-pure (18Ω) water containing 5% HPLC grade methanol (pH 7). These samples were then incubated with BPS (1 mM) in the presence of 10 mM ascorbate (final pH 4) for 30 minutes in the dark. These samples were then briefly vortexed and then centrifuged at 13,000 rpm for 15 minutes. The absorbance of the BPS-Fe complex was then determined in the particle free supernatant at 535 nm. Iron concentrations were determined against a standard curve of ferrous ammonium sulphate (0–50 μM) in ultrapure water, 5% methanol, in the presence of 10 mM ascorbic acid. Particle free and filter blanks were also ran in parallel to the extracted dung cake samples and these background concentrations subtracted from the determined dung cake Fe concentrations. A further blank containing the dung cake suspension with ascorbate but in the absence of BPS was also included and the background absorbance subtracted from the observed values in the dung cake samples. Total copper was similarly determined using the Cu+-specific chromogenic chelator bathocuproine-disulphonate at 483 nm [39], with the following amendments: the final ascorbate concentration in the samples and standards was 1 mM and the pH 7.0. Coppers concentrations were determined a standard curve of copper sulphate (0–20 μM) prepared in Chelex-100 resin treated water.
Statistical analysis
All data are expressed throughout as means with SD. Specific experimental details are summarised the each Figure legend. Experimental group comparisons were performed using two way repeated measures ANOVA. Post-hoc comparisons of group means were performed using the Student-Newman-Kuels test. All statistical analyses were performed using SPSS for windows, version 11.5 or the Unistat Exel plug-in, version 4.53.
List of Abbreviations used
AOX Antioxidant
AA Ascorbate
UA Urate
GSH Reduced glutathione
GSSG Glutathione disulphide
RTLF Respiratory tract lining fluid
DC Dung cake
DFO Desferoxamine
BPS Bathophenantroline-disulphonate
BCS Bathocuproine-disulphonate
DPTA Diethylene triamine pentaacetate
SOD Superoxide dismutase
hiSOD Heat inactivated superoxide dismutase
CAT Catalase
PAH Polyaromatic hydrocarbons
PM Particulate matter
WHO World Health Organisation
Competing interests
Ian S Mudway: None
Sean T Duggan: None
Chandra Venkataraman: None
Frank J Kelly: None
Jonathon Grigg: None
Authors' contributions
Ian S Mudway: Study design and experimental work, data analysis, preparation of manuscript
Sean T Duggan: Study design and experimental analysis
Chandra Venkataraman: Collection of particle samples from dung cake combustion
Gazala Habib: Collection of particle samples from dung cake combustion
Frank J Kelly: Study conception and design, manuscript review
Jonathon Grigg: Study conception and design
Acknowledgements
The authors would like to express their gratitude to Professors Thomas Sandstrom and Ken Donaldson for their gift of the ROFA, diesel and gasoline particles used as control particles in this study.
==== Refs
Barnes DF Openshaw K Smith KR van der Plas R What makes people cook with improved biomass stoves? A Comparative International Review of Stove Programs World Bank Technical Paper 1994 242 World Bank
Balakrishnan K Sankar S Parikh J Padmavathi R Srividya K Venugopal V Prasad S Pandey VL Daily average exposures to respirable particulate matter from combustion of biomass fuels in rural households of southern India Environ Health Perspect 2002 110 1069 75 12417476
Raiyani CV Jani JP Desai NM Shah SH Shah PG Kashyap SK Assessment of indoor exposure to polycyclic aromatic hydrocarbons for urban poor using various types of cooking fuels Bull Environ Contam Toxicol 1993 50 757 63 8490282 10.1007/BF00194673
Venkataraman C Rao GUM Emission factors of carbon monoxide and size-resolved aerosols from biofuel combustion Environ Sci Technol 2001 35 2100 7 11393993 10.1021/es001603d
Smith KR Samet JM Romieu I Bruce N Indoor air pollution in developing countries and acute lower respiratory infections in children Thorax 2000 55 518 32 10817802 10.1136/thorax.55.6.518
Ezzati M Kammen DM The health impacts of exposure to indoor air pollution from solid fuels in developing countries: knowledge, gaps, and data needs Environ Health Perspect 2002 110 1057 68 12417475
US Environmental Protection Agency National Ambient Air Quality Standards 11645578
Chen BH Hong CJ Pandey MR Smith KR Indoor air pollution in developing countries World Health Stat Q 1990 43 127 38 2238693
Mishra V Retherford RD Cooking smoke increases the risk of acute respiratory infection in children Natl Fam Health Surv Bull 1997 1 4
Mishra V Indoor air pollution from biomass combustion and acute respiratory illness in preschool age children in Zimbabwe Int J Epidemiol 2003 32 847 53 14559763 10.1093/ije/dyg240
Viegi G Simoni M Scognamiglio A Baldacci S Pistelli F Carrozzi L Annesi-Maesano I Indoor air pollution and airway disease Int J Tuberc Lung Dis 2004 8 1401 15 15636485
Bruce N Perez-Padilla R Albalak R Indoor air pollution in developing countries: a major environmental and public health challenge Bull World Health Organ 2000 78 1078 92 11019457
Kulkarni NS Prudon B Panditi SL Abebe Y Grigg J Carbon loading of alveolar macrophages in adults and children exposed to biomass smoke particles Science of the Total Environment
Donaldson K Stone V Current hypotheses on the mechanisms of toxicity of ultrafine particles Ann Ist Super Sanita 2003 39 405 10 15098562
Xia T Korge P Weiss JN Li N Venkatesen MI Sioutas C Nel A Quinones and aromatic chemical compounds in particulate matter induce mitochondrial dysfunction: implications for ultrafine particle toxicity Environ Health Perspect 2004 112 1347 58 15471724
Aust AE Ball JC Hu AA Lighty JS Smith KR Straccia AM Veranth JM Young WC Particle characteristics responsible for effects on human lung epithelial cells Res Rep Health Eff Inst 2002 1 65 12578113
Roberts ES Richards JH Jaskot R Dreher KL Oxidative stress mediates air pollution particle-induced acute lung injury and molecular pathology Inhal Toxicol 2003 15 1327 46 14569496
Stohs SJ Bagchi D Oxidative mechanisms in the toxicity of metal ions Free Radic Biol Med 1995 18 321 36 7744317 10.1016/0891-5849(94)00159-H
Squadrito GL Cueto R Dellinger B Pryor WA Quinoid redox cycling as a mechanism for sustained free radical generation by inhaled airborne particulate matter Free Radic Biol Med 2001 31 1132 8 11677046 10.1016/S0891-5849(01)00703-1
Takizawa H Abe S Okazaki H Kohyama T Sugawara I Saito Y Ohtoshi T Kawasaki S Desaki M Nakahara K Yamamoto K Matsushima K Tanaka M Sagai M Kudoh S Diesel exhaust particles upregulate eotaxin gene expression in human bronchial epithelial cells via nuclear factor-kappa B-dependent pathway Am J Physiol Lung Cell Mol Physiol 2003 284 L1055 62 12576300
Mudway IS Stenfors N Duggan ST Roxborough H Zielinski H Marklund SL Blomberg A Frew AJ Sandstrom T Kelly FJ An in vitro and in vivo investigation of the effects of diesel exhaust on human airway lining fluid antioxidants Arch Biochem Biophys 2004 423 200 12 14871482 10.1016/j.abb.2003.12.018
Mudway IS Kelly FJ Modeling the interactions of ozone with pulmonary epithelial lining fluid antioxidants Toxicol Appl Pharmacol 1998 148 91 100 9465268 10.1006/taap.1997.8318
Miller CA Linak WP King C Wendt JOL Fine particle emissions from heavy fuel oil combustion in a firetube package boiler Combust Sci Technol 1998 134 477 502
World Health Organisation World Health Report 2001 Geneva, WHO
Smith KR Mehta S The burden of disease from indoor air pollution in developing countries: comparison of estimates Int J Hyg Environ Health 2003 206 279 89 12971683 10.1078/1438-4639-00224
Bonvallot V Baeza-Squiban A Baulig A Brulant S Boland S Muzeau F Barouki R Marano F Organic compounds from diesel exhaust particles elicit a proinflammatory response in human airway epithelial cells and induce cytochrome p450 1A1 expression Am J Respir Cell Mol Biol 2001 25 515 21 11694458
Monn C Naef R Koller T Reactions of macrophages exposed to particles <10 microm Environ Res 2003 91 35 44 12550086 10.1016/S0013-9351(02)00021-X
Soukup JM Becker S Human alveolar macrophage responses to air pollution particulates are associated with insoluble components of coarse material, including particulate endotoxin Toxicol Appl Pharmacol 2001 171 20 6 11181108 10.1006/taap.2000.9096
Bonvallot V Baulig A Boland S Marano F Baeza A Diesel exhaust particles induce an inflammatory response in airway epithelial cells: involvement of reactive oxygen species Biofactors 2002 16 15 7 12515912
Madsen AM Martensson L Schneider T Larsson L Microbial dustiness and particle release of different biofuels Ann Occup Hyg 2004 48 327 38 15191942 10.1093/annhyg/meh016
Fernandez A Wendt JO Cenni R Young RS Witten ML Resuspension of coal and coal/municipal sewage sludge combustion generated fine particles for inhalation health effects studies Sci Total Environ 2002 287 265 74 11993968 10.1016/S0048-9697(01)00984-6
Van Reyk DM Dean RT The iron-selective chelator desferal can reduce chelated copper Free Radic Res 1996 24 55 60 8747892
Mudway IS Blomberg A Frew AJ Holgate ST Sandstrom T Kelly FJ Antioxidant consumption and repletion kinetics in nasal lavage fluid following exposure of healthy human volunteers to ozone Eur Respir J 1999 13 1429 38 10445623
Squadrito GL Cueto R Splenser AE Valavanidis A Zhang H Uppu RM Pryor WA Reaction of uric acid with peroxynitrite and implications for the mechanism of neuroprotection by uric acid Arch Biochem Biophys 2000 376 333 7 10775420 10.1006/abbi.2000.1721
Becker BF Towards the physiological function of uric acid Free Radic Biol Med 1993 14 615 31 8325534 10.1016/0891-5849(93)90143-I
Buettner GR The pecking order of free radicals and antioxidants: lipid peroxidation, alpha-tocopherol, and ascorbate Arch Biochem Biophys 1993 300 535 43 8434935 10.1006/abbi.1993.1074
Davies KJ Sevanian A Muakkassah-Kelly SF Hochstein P Uric acid-iron ion complexes. A new aspect of the antioxidant functions of uric acid Biochem J 1986 235 747 54 3753442
Nilsson UA Bassen M Savman K Kjellmer I A simple and rapid method for the determination of "free" iron in biological fluids Free Radic Res 2002 36 677 84 12180193 10.1080/10715760290029128
Cecconi I Moroni M Vilardo PG Dal Monte M Borella P Rastelli G Costantino L Garland D Carper D Petrash JM Del Corso A Mura U Oxidative modification of aldose reductase induced by copper ion. Factors and conditions affecting the process Biochemistry 1998 37 14167 74 9760253 10.1021/bi981159f
Hoffman GP Huggins FE Shah N Huggins R Pugmire RJ Meuzelaar HLC Seehra MS Manivannan A Characterisation of fine particulate matter produced by combustion of residual fuel oil J Air & Waste Manage Assoc 2000 50 1106 1114
Bloeman HJT Gerlofs-Nijland ME Janssen NAH Sandstrom T van Bree L Cassee FR Chemical characterisation and source apportionment estimates of particulate matter collected within the framework of the EU project HEPMEAP RIVM report 863001002/2005 2005
Anthony WB Animal waste value – nutrient recovery and utilization J Anim Sci 1971 32 799 802 5106328
Salvi S Blomberg A Rudell B Kelly F Sandstrom T Holgate ST Frew A Acute inflammatory responses in the airways and peripheral blood after short-term exposure to diesel exhaust in healthy human volunteers Am J Respir Crit Care Med 1999 159 702 9 10051240
|
16202154
|
PMC1262769
|
CC BY
|
2021-01-04 16:38:30
|
no
|
Part Fibre Toxicol. 2005 Oct 4; 2:6
|
utf-8
|
Part Fibre Toxicol
| 2,005 |
10.1186/1743-8977-2-6
|
oa_comm
|
==== Front
Part Fibre ToxicolParticle and Fibre Toxicology1743-8977BioMed Central London 1743-8977-2-71620216210.1186/1743-8977-2-7ResearchUltrafine particles cause cytoskeletal dysfunctions in macrophages: role of intracellular calcium Möller Winfried [email protected] David M [email protected] Wolfgang G [email protected] Vicki [email protected] GSF National Research Center for Environment and Health, Clinical research group 'Inflammatory Lung Diseases', Robert Koch Allee 29, D-82131 Munich-Gauting, Germany2 GSF National Research Center for Environment and Health, Institute for Inhalation Biology, and Focus Network Aerosols and Health, Ingolstädter Landstr. 1, D-85746 Neuherberg/München, Germany3 Napier University, School of Life Sciences, Edinburgh EH10 5DT, UK2005 4 10 2005 2 7 7 17 5 2005 4 10 2005 Copyright © 2005 Möller et al; licensee BioMed Central Ltd.2005Möller et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Particulate air pollution is reported to cause adverse health effects in susceptible individuals. Since most of these particles are derived form combustion processes, the primary composition product is carbon with a very small diameter (ultrafine, less than 100 nm in diameter). Besides the induction of reactive oxygen species and inflammation, ultrafine particles (UFP) can cause intracellular calcium transients and suppression of defense mechanisms of alveolar macrophages, such as impaired migration or phagocytosis.
Methods
In this study the role of intracellular calcium transients caused by UFP was studied on cytoskeleton related functions in J774A.1 macrophages. Different types of fine and ultrafine carbon black particles (CB and ufCB, respectively), such as elemental carbon (EC90), commercial carbon (Printex 90), diesel particulate matter (DEP) and urban dust (UD), were investigated. Phagosome transport mechanisms and mechanical cytoskeletal integrity were studied by cytomagnetometry and cell viability was studied by fluorescence microscopy. Macrophages were exposed in vitro with 100 and 320 μg UFP/ml/million cells for 4 hours in serum free medium. Calcium antagonists Verapamil, BAPTA-AM and W-7 were used to block calcium channels in the membrane, to chelate intracellular calcium or to inhibit the calmodulin signaling pathways, respectively.
Results
Impaired phagosome transport and increased cytoskeletal stiffness occurred at EC90 and P90 concentrations of 100 μg/ml/million cells and above, but not with DEP or UD. Verapamil and W-7, but not BAPTA-AM inhibited the cytoskeletal dysfunctions caused by EC90 or P90. Additionally the presence of 5% serum or 1% bovine serum albumin (BSA) suppressed the cytoskeletal dysfunctions. Cell viability showed similar results, where co-culture of ufCB together with Verapamil, W-7, FCS or BSA produced less cell dead compared to the particles only.
Ultrafine particlescytoskeletonstiffnessviscoelasticityphagosome transportrelaxationintracellular calcium
==== Body
Background
Epidemiological studies suggest increased health risks (decreased lung function, increased morbidity and mortality) after exposure to environmental particles [1,2]. Recent epidemiological studies show, that not only the mass of inhaled urban particles (i.e. PM10 or PM2.5 ≡ mass of all particles being smaller than 10 μm or 2.5 μm, respectively) is associated with increased morbidity and mortality, but also the number of particles with even higher significance [3,4]. Because of the smaller mass of ultrafine (diameter < 100 nm) compared to fine particles, they contribute little to the total mass, but are important in the overall number of particles. Most urban particles result from combustion processes; therefore the dominant fraction contains ultrafine carbonaceous particles. These particles are less soluble and therefore can reside in the lung for longer times. Larger particles and bacteria are phagocytized by alveolar and airway macrophages and dendritic cells within hours and further digested, keeping the lung surface in a sterile condition (under physiological conditions). UFP are not only phagocytosed by alveolar macrophages, but can enter epithelial cells, can penetrate into the circulation, being further redistributed to other organs of the body [5]. Therefore, UFP can reach sites in the body far away from the site of entrance, where they can cause inflammatory reactions, in contrast to larger particles [6,7]. With decreasing particle size the particle surface increases in relation to the particle volume or particle mass. Particles forming aggregates being composed of subunits of much smaller size can further enhance their surface area in relation to the mass (specific surface area). Because of their high specific surface area, UFP can catalyze chemical reactions, which may induce chronic inflammation. Animal studies and in vitro cell studies have shown that high concentrations of ultrafine particles can induce inflammatory processes and increased calcium transients [8-13], can induce the production of reactive oxygen species (ROS, like H2O2, O2-) [14-16], can induce oxidation of cytoskeletal proteins [17,18] and cytoskeletal dysfunctions [19,20], or can act as vehicles to transport toxic gasses and substances to the lung periphery [21].
Phagocytic cells in the lung, such as macrophages, polymorphonuclear leukocytes and dendritic cells play a key-role in the defense-reaction. They ingest foreign materials, digest bacteria and viruses, and present antigens in order to trigger specific (immunological) defense mechanisms [22-24]. Macrophages reside in the alveoli and chemotactic gradients direct the macrophages within minutes to the site of particle deposition. During phagocytosis the particles are incorporated into a membranous vesicle and ingested into intracellular phagosomes, which fuse with lysosomes [25]. Phagolysosomes are acidic (pH ≈ 5) and contain reactive oxygen species (H2O2) [26,27]. They are the site where most bacteria and fungi are digested. Non-digestible particles are retained in the lung for longer periods of time, for example iron oxide tracer particles have a clearance half-time of 120 days in healthy non-smoking subjects [28]. UFP are not only engulfed by phagocytosis and therefore must not necessarily reside in phagolysosomes, which may imply much longer residence times in the lung. The cytoskeleton of the AM is crucially involved in the phagocytic defense reactions, including locomotion and cell migration, phagocytosis, intracellular transport, phagosome-lysosome fusion and signal transduction [29-31]. The cytoskeleton consists of three different filamentous structures. Microfilaments (actin) are involved in dynamic processes of the cell, like crawling and phagocytosis. Microtubuli are involved in cell shape and intracellular vesicle transport. Intermediate filaments contribute to the static part of the cytoskeleton. Every filamentous structure has its own family of motor proteins being responsible for the transport of molecules and vesicles [32]. Those transport processes require energy in the form of ATP [33,34]. The cytoskeleton is sensitive to ROS and oxidative stress, due to the presence of thiol groups located on the actin microfilaments which are sensitive to oxidation, leading to cross linking and reduced motility [18,35], and the cytoskeletal dynamics is regulated by cytoplasmic calcium.
We have developed protocols to study cytoskeleton associated functions in macrophages in vivo and in vitro, such as phagosome transport, mechanical integrity (viscoelastic properties, stiffness) and phagocytosis, using ferromagnetic microparticles [36-38]. In in vitro studies the magnetic microparticles (1.8 μm diameter) are incubated for 24 hours together with cultivated macrophages, after which more than 95% are ingested. The particles are then magnetized and aligned in a short magnetic field pulse, and can be detected by a magnetic field sensor. Intracellular phagosome transport causes stochastic disorientations of the particles, which results in a decay of the remanent magnetic field of the cell probe (relaxation). In addition mechanical viscoelastic properties of the cytoskeleton can be investigated by twisting the micromagnets in a weak magnetic field. This method is called 'Magnetic Twisting Cytometry' (MTC) [39,40] and has been used to investigate the role of the different cytoskeletal structures in macrophage function after co-incubation with cytoskeletal drugs (Cytochalasin D, Nocodazole [37]). Additionally it has been shown that MTC can monitor cytotoxicity of GaAs-particles in vivo in lung macrophages of animals [41,42] and in vitro caused by ultrafine particles [20].
In this study, we have investigated the role of intracellular calcium transients produced by environmentally relevant fine and ultrafine model particles in the induction of adverse reactions on the cytoskeleton of macrophages, using the MTC technique. Cultivated macrophages from the cell line J774A.1 were incubated with increasing amounts of particles. Intracellular phagosome transport (relaxation), stiffness of the cytoskeleton and cell viability was recorded. Particles on the basis of carbon black (CB) having different chemical compositions and impurities, and a wide range of diameters (12 nm - 1.5 μm) and specific surface areas (≈ 1 m2/g - 600 m2/g) were tested.
Materials and methods
Target cells
J774A.1 macrophages originate from a BALB/c/NIH mouse [43] and were obtained from the German Collection of Animal Cell Cultures (Tumorbank, DKFZ Heidelberg, Germany). Cells were grown in RPMI 1640 Medium (Sigma, Taufkirchen, Germany), supplemented with 5% fetal calf serum (FCS), 100 U/ml penicillin, 100 μg/ml streptomycin, 2.5 μg/ml amphotericin and 0.3 g/L-glutamine in NaHCO3. The cell culture grows with a doubling time of 2 days and was sub-cultured every 4 days. In cytotoxicity studies cells were incubated with particles and with Calcium agonists in serum-free medium. In addition cells were incubated with 5% FCS to test the possible cytotoxic effect of lack of serum in the medium.
Magnetic particle binding assay
A number of 0.2 million macrophages were incubated with 10 μg of 1.8 μm spherical ferromagnetic microparticles in glass vials (12 mm outer diameter) for 24 hours prior to adding calcium antagonists or ultrafine particles. This ensured the adherence of the macrophages and more than 95% phagocytosis of the magnetic beads. Before the addition of UFP or calcium antagonists, non-adherent cells and free particles were removed by washing with medium. Ferromagnetic 1.8 μm magnetite microparticles (beads) were prepared with narrow size variation (geometric standard deviation < 1.1) and spherical shape [44], which is important for data analysis using mathematical models to estimate the viscous and elastic properties of the cytoskeleton. The particles were not further coated in order to avoid the activation of specific cell surface receptors. It is reported that the non-specific Scavenger-receptor mediates phagocytosis of the magnetite particles [45].
Ultrafine test-particles and calcium modulating drugs
Table 1 gives an overview of the test particles used in the study, together with their physical properties. Because most urban particles result from combustion processes, the dominant fraction contains fine and ultrafine carbonaceous particles. Ultrafine carbon (EC90) particles were produced by an electrical spark generator [46] under standardized conditions with low impurities (for example transition metals, polycyclic hydrocarbons; about 5% organic carbon). The mobility diameter in air of these particles was 90 nm and the high specific surface area of 600 m2/g indicated that aggregates had much smaller subunits. Alternatively, commercially available ultrafine carbon particles were used (Printex, P90, Degussa, Frankfurt, Germany), which have low organic impurities of (≈ 1%), but some impurities of transition metals (iron etc.). Diesel particulate matter (Standard Reference Material 1650, DEP) and urban dust (Standard Reference Material 1649a, UD) were used to simulate environmental particle exposure [47,48]. The certificate of analysis of UD does not give an estimation of the specific surface area; therefore a value from literature was obtained for a comparable urban dust (PM2.5) probe [49]. The particles were suspended within glass tubes in deionised water by ultrasonication and then further diluted in medium without serum. Microscopic investigation showed that a large fraction of the carbonaceous particles appeared as aggregates in the cell probes.
Table 1 Fine and ultrafine test particles used in the study together with their physical properties, such as count median diameter and specific surface area. Elemental carbon particles (EC90) were prepared in own laboratory (GSF-IHB, [46]).
Material type Diameter Specific surface area Source
Printex90 (carbon, P90) 12 nm 300 m2/g Degussa
Elemental carbon (EC90) 90 nm+ 600 m2/g GSF-IHB
Diesel exhaust particles (DEP) 120 nm 108 m2/g NIST*
Urban dust (UD) 1.5 μm ≈ 1 m2/g++ NIST*
+ size of the airborne agglomerates; primary particles are expected to be < 10 nm [46].
* National Institute of Standards and Technology (NIST), Washington, USA [47, 48].
++estimated from other PM2.5 ([49])
Particles were incubated at concentrations of 100 and 320 μg/ml/million cells for 4 h in serum-free medium, or in combination with calcium antagonists, such as Verapamil (100 μM, Ca2+ channel blocker), BAPTA-AM (20 μM, Ca2+ chelator), or W-7 (N-(6-aminohexyl)-5-chloro-1-naphthalene-sulfonamide hydrochloride, 25 μM, calmodulin inhibitor). The concentration of the Ca-antagonists was obtained from previous studies and tested for non-toxic responses in own preliminary studies. The antioxidant Nacystelin (NAL, 200 μM) was also included. Nacystelin (NAL), a thiol antioxidant compound that is a lysinated derivative of N-acetyl cysteine (NAC), possesses potent mucolytics capacities and has been shown to inhibit reactive oxygen species effects [50]. It has the advantage of having a neutral pH compared with NAC, which is acidic, and thus can be administered to the airways without the local airway irritation that occurs with NAC. Additionally the role of serum (FCS, 5%) or bovine serum albumin (BSA, 1%) in the medium was investigated. In some experiments long-term incubation (24 hours) was carried out to compare with previous data.
Cell viability
Cell viability was tested by the propidium iodide (PI) exclusion test. Necrotic cells allow the penetration of PI into the cell and the nucleus, where the fluorescent dye can be visualized. After 10 min PI incubation with adherent cells, the probes were analyzed by fluorescence microscopy, using an inverted microscope (Axiovert 25, Zeiss GmbH, Jena, Germany) together with a mercury-arc lamp and a filter system for the detection of propidium iodide (PI) dye and a digital camera system (CoolSnap-pro, Mediacybernetics Inc., Silver Spring, MD, USA). Digital images were recorded and processed using Image Pro software package (Mediacybernetics Inc., Silver Spring, MD, USA). For every probe a transmission image was recorded as well as a fluorescence image. The transmission image was used to count the total number of cells in a region of interest (ROI), defined by a grid overlay and covering about 80–100% of the image. Each ROI covered at least 100 cells. In the second fluorescence image the number of PI positive cells was analyzed in the same ROI, allowing the fraction of dead cells (PI+) to be assessed.
Magnetic twisting cytometry
For in vitro measurement of relaxation and viscoelastic cell properties a magnetic twisting cytometry device (MTC) was used [40]. A glass vial (12 mm diameter) containing adherent macrophages with ingested magnetic particles is positioned in a second gradiometer array of flux-gate sensors (Förster GmbH, Reutlingen, Germany; Figure 1). The particles in the cells were aligned parallel to the direction of the sensors by a 200 mT, 10 μs magnetic field pulse. The probe was rotated at 6 Hz and the signal of the fluxgate array was amplified and phase sensitive detected which significantly reduced system noise and improved the sensitivity. 10 μg of ferromagnetic particles induce a remanent magnetic field (RMF) of ≈ 1 nT in the sensor array. Particle twisting was performed in a magnetic twisting field (1 – 2, 5 mT) perpendicular to the direction of detection (parallel to the axis of rotation).
Figure 1 Magnetic twisting device to measure relaxation and twisting (10 sec) of aligned ferromagnetic microparticles ingested by macrophages and detection by an array of magnetic fluxgate sensors (Förster devices, Förster GmbH, Reutlingen, Germany).
Stochastic phagosome motion (relaxation)
The motion of vesicles and phagosomes happens continuously within living cells and is part of the intracellular transport system. Relaxation describes the decay of the magnetic field after particle alignment in a magnetic pulse field and originates from the randomization of the magnetic particles. We assume that the phagosomes are coupled to the cytoskeletal filaments by motor proteins and that the hydrolization of ATP provides the energy to move the phagosomes and to induce rotational random kicks to the phagosome [33,39,51,52]. A hydrodynamic relaxation model was developed under the assumption that the intracellular randomization energy Er behaves like thermal energy kT, implying a rotational Brownian motion process together with an exponential decay in a Newtonian viscosity. The hydrodynamic relaxation model was fitted to the experimental data by a non-linear regression algorithm. Additionally two robust relaxation parameters were analyzed, being independent of any model (Figure 2A). This is the normalized RMF after 1 minute, b1 = B(1 min)/B0, which characterizes the initial fast phase of decay, and that after 5 minutes, b5 = B(5 min)/B0, which is characteristic for the decay in the following slow phase. Figure 2A illustrates the decay of aligned particles and the impairment by UFP and by Cytochalasin D, a microfilament disruption agent.
Figure 2 Measurement of stochastic intracellular phagosome transport (relaxation, A) and cytoskeletal stiffness (B) by Magnetic Twisting Cytometry (MTC). Control and cytochalasin D (CyD) probes do not contain UFP. Other probes show co-incubation with 100 μg/ml P90 without serum (w/o FCS) and with 100 μM Verapamil or 20 μM BAPTA-AM for 4 hours.
Magnetic phagosome twisting
Application of a weak magnetic field BTW induces twisting of a magnetic dipole particle and allows investigating cytoplasmic rheology and mechanical integrity [36,53]. Particles being suspended in a viscosity η rotate in an external twisting force according to Newton's law:
where dθ/dt is the shear rate and σ is the applied shear stress (, M = remanent magnetization of particles, κ = rotational shape factor). In case of elasticity the applied stress is proportional to the elastic deformation (strain θ) and we get with the elasticity modulus ν Hook's law:
σ = νθ (2)
Strain was estimated from the measurement of the cell field B(t) according to θ(t) = arccos B(t)/B0. First the particles were magnetized and aligned parallel to the field sensors (Figure 2B). After 20 sec relaxation the twisting field was applied for 10 sec duration and viscoelastic recoil was recorded for another 3 minutes. Viscoelastic recoil does not force the dipoles back to the undisturbed relaxation curve, indication of viscous shear, which can reflect a permanent deformation (break of filamentous interactions) of the cytoskeletal structure. Figure 2B shows the change of particle alignment during magnetic field twisting together with the elastic recoil in comparison to a relaxation curve recorded without any external forces to the magnetic beads. Permanent deformation is characterized by the difference in complete elastic recoil and undisturbed relaxation, as illustrated in Figure 2B. Cell stiffness was estimated as the ratio between mean stress and strain after a constant twisting duration of 10 sec. This analysis of particle twisting does not view specific viscous or elastic properties. Therefore this parameter provides an integral description of the cytoskeletal mechanical properties.
Data analysis
The data presented in Figures 3 and 4 under the influence of particles or drugs are normalized to results of control probes without particle co-incubation. An inhibited relaxation (slower decay, higher b5) results then in a normalized value larger than one. An accelerated relaxation displays as a normalized value smaller than one. Every set of UFP/calcium antagonist measurements was performed on at least 5 separate probes together with a separate set of control measurements. All parameters estimated under the set of antagonists were normalized to the set of control measurements. A significant deviation of the normalized parameters from unity denotes an influence of the UFP/antagonist. Using a 2-sided Student's t-Test, deviations from unity were analyzed for their level of statistical significance. An influence by the appropriate antagonist was accepted when the level of significance was p < 0.05. Pearson's correlation analysis was performed using WINSTAT, Version 2001.1, Fitch Software, Cambridge, USA.
Figure 3 Relaxation of ingested magnetic particles (A, relative decay after 5 min, b5, normalized to control probes without particles) and mechanical integrity (B, stiffness, mean between low and high stress) of J774A.1 macrophages after 4 h incubation with different types of 100 μg ultrafine particles/ml/million cells under different incubation conditions, such as without serum (w/o FCS), 100 μM Verapamil, 20 μM BAPTA-AM or 25 μM W-7 (normalized values +/- SD; N = 5; **: p < 0.01, *: p < 0.05).
Figure 4 Relaxation of ingested magnetic particles (A, relative decay after 5 min, b5, normalized to control probes w/o particles) and mechanical integrity (B, stiffness, mean between low and high stress) of J774A.1 macrophages after 4 h incubation with different types of 100 μg ultrafine particles/ml/million cells under different incubation conditions, such as without (w/o) and with 5% serum (wFCS), 1% bovine serum albumin (BSA) or 200 μM Nacystelin (NAL); normalized values +/- SD; N = 5; **: p < 0.01, *: p < 0.05.
Results
Influence of UFP on phagosome motion and on magnetic particle twisting
The results of stochastic intracellular phagosome transport and of cytoskeletal mechanical integrity after 4 hours of incubation of J774A.1 macrophages with 100 μg/ml fine or ultrafine particles are shown in Figure 3 (summarized in Table 2 and Table 3). Compared to cells without particles (control), P90 and EC90 cause a retardation of relaxation both, for concentrations of 100 μg/ml and of 320 μg/ml (data not shown). This retardation is not seen for DEP or UD particles. The retardation of relaxation can be inhibited in part by the co-incubation with Verapamil, but not with BAPTA-AM. Co-incubation of W-7 with 100 μg/ml P90 in part inhibits the retardation of relaxation, while it does not with 100 μg/ml EC90 particles. UD significantly retarded relaxation only at the high concentration of 320 μg/ml; the retardation could be inhibited by the calcium antagonists. The measurements of cytoskeletal stiffness (Figure 3B) reflect the same results as were found on relaxation measurements. P90 and EC90 cause an increase of cytoskeletal stiffness, when no serum is present. Verapamil lower the stiffening caused by P90 or by EC90. W-7 inhibited the stiffening only for P90 co-incubation, but not with EC90. BAPTA-AM did not influence the particle induced stiffening for both particle types. The higher particle concentration of 320 μg UFP/ml reflects a comparable response. The urban particles (DEP or UD) do not modulate cytoskeletal stiffness, neither alone or in combination with calcium antagonists (data not shown). In summary the data show cytoskeletal dysfunctions caused by P90 and EC90, but not by DEP and by UD. Cytoskeletal dysfunctions can be inhibited in part by the Ca2+ channel blocker Verapamil and by the calmodulin inhibitor W-7, but not by the Ca2+ chelator BAPTA-AM.
Table 2 Summary of effects of 100 μg particles/ml/million cells on phagosome transport (relaxation) together with effect of co-incubation with different Ca-modulating drugs (100 μM Verapamil, 20 μM BAPTA-AM or 25 μM W-7) as well as medium with serum (wFCS), with bovine serum albumin (BSA), or with 200 μM Nacystelin (NAL); inhib. – inhibition, n.e. – no effect, accel. – acceleration; N = 5; **: p < 0.01, *: p < 0.05.
Drug P90 EC90 DEP UD
w/oFCS inhib.** inhib.** accel.* n.e.
Verap inhib.* n.e. n.e. n.e.
BAPTA inhib.** inhib.** n.e. n.e.
W-7 inhib.** inhib.** n.e. n.e.
wFCS n.e. n.e. n.e. n.e.
BSA n.e. n.e. n.e. n.e.
NAL inhib.** inhib.** n.e. n.e.
Table 3 Summary of effects of 100 μg particles/ml/million cells on cytoskeletal stiffness together with effect of co-incubation with different Ca-modulating drugs (100 μM Verapamil, 20 μM BAPTA-AM or 25 μM W-7) as well as medium with serum (wFCS), with bovine serum albumin (BSA), or with 200 μM Nacystelin (NAL); stiff. – stiffening, n.e. – no effect; N = 5; **: p < 0.01, *: p < 0.05.
Drug P90 EC90 DEP UD
w/oFCS stiff.* stiff.** n.e. n.e.
Verap stiff.* n.e. n.e. n.e.
BAPTA stiff.** stiff.** n.e. n.e.
W-7 n.e. stiff.** n.e. n.e.
wFCS n.e. n.e. n.e. n.e.
BSA n.e. n.e. n.e. n.e.
NAL stiff.** stiff.** n.e. n.e.
Influence of serum, BSA and Nacystelin on phagosome motion and on magnetic particle twisting
Figure 4 shows the results of serum and of BSA on phagosome transport and on cytoskeletal stiffness after co-culture with P90 and EC90 (summary in Table 2 and Table 3). Both, 5% FCS and 1% BSA are able to inhibit the particle induced retardation of relaxation (Figure 4A) and the increase in cell stiffness (Figure 4B). The antioxidant Nacystelin does not inhibit the particle induced cytoskeletal dysfunctions. Serum, BSA and Nacystelin do not influence any of the effects of DEP or UD, therefore, the data are not shown.
Influence on cell viability
After 4 hours incubation time none of the particles or drugs decreased cell viability to below 90% at a particle concentration of 100 μg/ml/million cells. Figure 5 shows the results of the cytotoxicity test (PI exclusion) of the J774A.1 cells for a particle concentration of 320 μg/ml and 4 hours incubation time. P90 and EC90 significantly enhanced the fraction of dead cells in medium w/o serum, which in part could be suppressed by Verapamil and W-7, but not by BAPTA-AM. In addition FCS and BSA, but not NAL reduced the fraction of dead cells after co-culture with P90 or EC90. Compared to control, there was no significant increase in the fraction of dead cells after co-culture with DEP or DU. An incubation time of 24 hours further enhanced the cytotoxic effect of the UFP (data not shown). The viability of the cells under control conditions (w/o particles) was about 90% with FCS, BSA or NAL and decreases to about 80% w/o FCS (p < 0.01 compared to 4 hours incubation time). 24 h incubation with the Ca-modulating drugs showed a reduced viability of 80% for Verapamil and BAPTA-AM, and 70% for W-7. This reflects non-physiological conditions of the Ca-modulating antagonists after long-term exposure. 320 μg/ml P90 and EC90 raised the fraction of dead cells to up to 80% and the effect of the Ca-modulating drugs was not uniform. Long-term co-culture with DEP did not raise the fraction of dead cells, compared to control incubation conditions. UD caused a doubling of the fraction of dead cells, which could be inhibited in part by co-culture with Verapamil, W-7, and with NAL, but not with BAPTA-AM.
Figure 5 Cell viability (PI exclusion) of J774A.1 macrophages after 4 hours of incubation without (control) and with 320 μg UFP/ml/million under different incubation conditions. A): without serum (w/oFCS), 100 μM Verapamil, 20 μM BAPTA-AM and 25 μM W-7. B): w/oFCS, with 5% serum (wFCS), 1% bovine serum albumin (BSA) or 200 μM Nacystelin (NAL); mean +/- SD, **: p < 0.01, *: p < 0.05.
Discussion and conclusion
Necrosis caused by UFP
Figure 5 shows both the cytotoxic potential of the particles in inducing cell necrosis used in the study, together with the Ca-modulating antagonists, particularly after 4 hours incubation time. The ultrafine carbon black particles (P90 and EC90) had the most significant effect on cell viability, with a doubling of the fraction of dead cells after 4 hours and a triplication after 24 hours. EC90 particles caused more dead cells due to the higher specific surface area compared to P90. FCS and BSA both reduce the cytotoxic potential of P90 and EC90. In the presence of FCS or BSA, the particle surface may become coated with proteins, which may shield reactive sites on the particle surface and reduced the potential of causative oxidative reactions. Although serum free medium allows clearer experimental conditions in case of receptor dependent responses, the serum co-culture may be closer to physiological conditions in the lung, where surfactant coats and opsonizes the particles. On the other hand the inhibitory effects of the Ca-antagonists would not have been seen with the presence of serum in the medium. Long-term incubation of J774A.1 macrophages without serum induced a decrease in cell viability, showing that this by itself is a non-physiological condition. Interestingly, cell viability could be enhanced by the presence of serum or BSA in the medium. But for short time periods (4 hours), the absence of serum did not injure the cells.
DEP and UD produce less cytotoxicity compared to the pure carbon black particles, despite other reactive material may be present, like transition metals, and adsorbed organic substances. One reason for the lower cytotoxicity of DEP and UD may be the lower specific surface area. In a previous study we could show cytoskeletal dysfunctions caused by DEP and UD after 24 hours incubation time [20]. In this study different incubation conditions were used, where particles were dispersed in medium containing serum. It was shown that the particles form larger aggregates without serum proteins in the medium, causing particle clusters in the μm-size range [54]. Such clusters are subject to different mechanisms of uptake and processing by the cells (phagocytosis versus endocytosis for smaller particles), although both are actin based mechanism. Phagocytosed materials are processed via the lysosomal digestion route, which is not the case after endocytosis. In addition the specific surface area is smaller for aggregated particles. It was also shown that organics adsorbed to the particle surface can lower the specific surface area, can shield reactive sites; therefore make the particles less toxic than plain particles [55]. We suggest here that the organic substances are stable when bound to the surface, in conjunction with low concentrations in the medium. The bound organics on the particles may cause a surface coating, which can shield reactive sites of the carbonaceous core.
Influence of UFP on phagosome motion and on magnetic particle twisting
MTC studies investigate the transport of micron-sized phagosomes from two different perspectives. Relaxation directly monitors the coupling dynamics between phagosomes and the cytoskeleton [39] while phagosome twisting examines the mechanical integrity (viscoelastic properties, stiffness) of the cytoskeletal filaments, which are linked to the phagosomes [56]. Phagosome transport requires an intact cytoskeleton, intact phagosomes including motor proteins and energy (ATP) [33,57]. Additionally, intracellular calcium plays an important role for cytoskeletal functions and intracellular signaling in relation to immunological and non-immunological defense reactions [58,59]. Cytotoxic reactions can involve the energy metabolism of the cell, intracellular calcium transients [13], the dynamics of cytoskeletal filaments and the transport mechanisms of phagosomes [41,42]. A previous study has shown that destroying microfilaments by cytochalasin D resulted in a retarded relaxation and an increased stiffness [37,53]. Colchicine, which disrupts microtubuli, results in an accelerated relaxation and in a moderately increased stiffness. Figure 2A illustrates the retardation of relaxation caused by 100 μg/ml P90, and the modulation by Verapamil and BAPTA-AM together with the retardation produced by 4 μM Cytochalasin D. In comparison to the effects induced by the cytoskeletal drugs the dysfunctions produced by the carbonaceous UFP suggests destructions of microfilamentous structures and hence dysfunctions of the cytoskeleton. This is also in agreement with other studies which showed impairment of the phagocytic capacity (an actin based mechanism) after incubation with different types of UFP [19,20].
Intracellular calcium plays a central role in the modulation of the defense mechanisms causing inflammation. Further studies have shown that ufCB causes a transient increase of intracellular calcium, which can be inhibited by the calcium channel blocker Verapamil [13,60]. This was reflected in this study by the inhibition of ufCB induced cytoskeletal dysfunctions in the presence of Verapamil. Interestingly, the intracellular calcium chelator BAPTA-AM did not induce the expected suppression of cytoskeletal dysfunction. The reason for this lack of effect is not clear but suggests that not only the level of intracellular calcium is responsible for the ufCB induced dysfunctions.
Calmodulin (CaM) is a ubiquitous calcium binding protein that can bind to and regulate a multitude of different protein targets by affecting many different cellular functions. CaM mediates processes such as inflammation, metabolism, apoptosis muscle contraction and cellular movement. It has been shown that calmodulin has a direct impact on actin polymerization [61-63], acto-myosin binding and cross linking [64,65]. CaM is thought to activate the myosin light chain kinase (MLCK) and CaM kinase II by displacement of their auto inhibitory domains. Many of the proteins that bind CaM are unable to bind calcium themselves and as such use CaM as a calcium sensor and signal transducer. The fact that the CaM-inhibitor W-7 can suppress most of the cytotoxic reactions caused by ufCB co-culture shows that the calcium dependent signaling pathway is crucial for the cytotoxic effect and cytoskeletal dysfunctions. The lack of suppression after co-culture with EC90 particles may reflect their higher specific surface area.
The antioxidant Nacystelin did not inhibit the UFP induced dysfunctions of the cytoskeleton. Oxygen radicals cause an oxidation and depletion of cytoskeletal proteins (thiols), disruption of actin filaments and the inhibition of F-actin formation, and an actin cross linking [18]. The actin system is the most sensitive constituent of the cytoskeleton to the oxidant attack. Nacystelin can inhibit the thiol oxidation [66], but did not result in a suppression of UFP induced cytoskeletal dysfunctions. This shows that the thiol oxidation may not be the only response to UFP exposure (besides the modulation of the calcium metabolism), but the cross linking of actin filaments primarily may describe the retarded relaxation and the increased stiffness of the cytoskeleton.
The specific surface area of the particles is discussed to be a significant parameter relating to cytotoxic responses of ultrafine particles [16,67]. Our studies in part support this hypothesis. Cell viability seems to correlate with the specific surface area of the particles, where EC90 induces more dead cells compared to P90 or to UD and DEP. Impairment of relaxation and stiffening of the cytoskeleton seem not to depend on the specific surface area, but W-7 could inhibit the particle induced impairments in P90, but not in EC90, which has the higher specific surface area. Besides the different surface characteristics of DEP and UD compared to EC90 and P90, the lack of cytoskeletal dysfunctions of these materials seems to be in part induced by the smaller specific surface area.
List of abbreviations
UFP ultrafine particles
ufCB ultrafine carbon black
PM2.5 particle mass with aerodynamic diameter <= 2.5 μm
EC90 elemental carbon particles, 90 nm mobility diameter
P90 commercial carbon particles, Printex 90 (Degussa)
DEP diesel exhaust particles
UD urban dust
W-7 (N-(6-aminohexyl)-5-chloro-1-naphthalene-sulfonamide hydrochloride, 25 μM, calmodulin inhibitor
CaM calmodulin
BAPTA-AM Ca2+ chelator
Verap Verapamil
BSA bovine serum albumin
FCS fetal calf serum
ATP adenosine triphosphate
PI probidium iodide
ROI region of interest
MTC magnetic twisting cytometry
RMF remanent magnetic field of cell/magnetic particle probe
b1 = B(1 min)B0 RMF after 1 min relaxation
b5 = B(5 min)B0 RMF after 1 min relaxation
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
WM has long experience in MTC measurements and investigation of cytoskeletal functions using MTC. He has conducted the studies together with Dr. David Brown during a visit at the lab in Gauting/Germany. DB and VS have long experience in particle toxicology and the induction of reactive oxygen species. They have a particular knowledge and experience in the role of calcium transients in intracellular signalling pathways. They designed the studies and interpreted the results, and DB conducted the studies during a visit at the lab in Gauting/Germany. WK is a specialist in studies of ultrafine particle clearance and translocation, and in ultrafine particle toxicology. He produced the EC90 particles, characterized them and contributed to the study design and interpretation.
Acknowledgements
This work was supported by the CEC under contract FIGD-CT-2000-00053.
==== Refs
Dockery DW Pope CA 3rdXu X Spengler JD Ware JH Fay ME Ferris BG Speizer FE An association between air pollution and mortality in six U.S. cities N Engl J Med 1993 329 1753 1759 8179653 10.1056/NEJM199312093292401
Pope CA Thun MJ Namboodiri MM Dockery DW Evans JS Speizer FE Heath CW Particulate air pollution as a predictor of mortality in a prospective study of U.S. adults Am J Respir Crit Care Med 1995 151 669 674 7881654
Peters A Wichmann HE Tuch T Heinrich J Heyder J Respiratory effects are associated with the number of ultrafine particles Am J Respir Crit Care Med 1997 155 1376 1383 9105082
Wichmann HE Spix C Tuch T Wölke G Peters A Heinrich J Kreyling WG Heyder J Daily mortality and fine and ultrafine particles in Erfurt, Germany part I: role of particle number and particle mass Research report/HEI 2000 98 5 86
Oberdörster G Sharp Z Atudorei V Elder A Gelein R Lunts A Kreyling WG Cox C Extrapulmonary translocation of ultrafine carbon particles following whole-body inhalation exposure of rats J Toxicol Env Heal A 2002 65 1531 1543 10.1080/00984100290071658
Kreyling WG Semmler M Möller W Dosimetry and toxicology of ultrafine particles J Aerosol Med 2004 17 140 152 15294064 10.1089/0894268041457147
Oberdörster G Oberdörster E Oberdörster J Nanotoxicology: An emerging discipline evolving from studies of ultrafine particles Environ Health Perspect 2005 113 823 839 16002369
Oberdoerster G Ferin J Gelein R Soderholm SC Finkelstein J Role of the alveolar macrophage in lung injury: studies with ultrafine particles Environ Health Perspect 1992 97 193 199 1396458
Johnston CJ Finkelstein JN Gelein R Baggs R Oberdoerster G Characterization of the early pulmonary inflammatory response associated with PTFE fume exposure Toxicol Appl Pharmacol 1996 140 154 163 8806881 10.1006/taap.1996.0208
Li XY Gilmour PS Donaldson K MacNee W In vivo and in vitro proinflammatory effects of particulate air pollution (PM10) Environ Health Perspect 1997 105 1279 1283 9400738
Li XY Brown D Smith S MacNee W Donaldson K Short-term inflammatory responses following intratracheal instillation of fine and ultrafine carbon black in rats Inhal Toxicol 1999 11 709 731 10477444 10.1080/089583799196826
Brown DM Stone V Findlay P MacNee W Donaldson K Increased inflammation and intracellular calcium caused by ultrafine carbon black is independent of transition metals or other soluble components Occup Environ Med 2000 57 685 691 10984341 10.1136/oem.57.10.685
Stone V Tuinman M Vamvakopoulos JE Shaw J Brown D Petterson S Faux SP Borm P MacNee W Michaelangeli F Donaldson K Increased calcium influx in a monocytic cell line on exposure to ultrafine carbon black Eur Respir J 2000 15 297 303 10706495 10.1034/j.1399-3003.2000.15b13.x
Zhang Q Kusaka Y Sato K Nakakuki K Kohyama N Donaldson K Differences in the extent of inflammation caused by intratracheal exposure to three ultrafine metals: role of free radicals J Toxicol Environ Health 1998 53 423 438 10.1080/009841098159169
MacNee W Donaldson K How can ultrafine particles be responsible for increased mortality? Monaldi Arch Chest Dis 2000 55 135 139 10949875
Beck-Speier I Dayal N Karg E Maier KL Schumann G Schulz H Semmler M Takenaka S Stettmaier K Bors W Oxidative stress and lipid mediators induced in alveolar macrophages by ultrafine particles Free Radic Biol Med 2005 38 1080 1092 15780766 10.1016/j.freeradbiomed.2005.01.004
Tang JX Janmey PA Stossel TP Ito T Thiol oxidation of actin produces dimers that enhance the elasticity of the F-actin network Biophys J 1999 76 2208 2215 10096915
Dalle-Donne I Rossi R Milzani A Di Simplicio P Colombo R The actin cytoskeleton response to oxidants: from small heat shock protein phosphorylation to changes in the redox state of actin itself Free Radic Biol Med 2001 31 1624 1632 11744337 10.1016/S0891-5849(01)00749-3
Lundborg M Johard U Lastbom L Gerde P Camner P Human alveolar macrophage phagocytic function is impaired by aggregates of ultrafine carbon particles Environ Res 2001 86 244 253 11453675 10.1006/enrs.2001.4269
Möller W Hofer T Ziesenis A Karg E Heyder J Ultrafine particles cause cytoskeletal dysfunctions in macrophages Toxicol Appl Pharmacol 2002 182 197 207 12183099 10.1006/taap.2002.9430
Niessner R Wilbring P Ultrafine Particles as Trace Catchers for Polycyclic Aromatic Hydrocarbons: The Photoelectric Aerosol Sensor as a Tool for In Situ Sorption and Desorption Studies Anal Chem 1989 61 708 714 10.1021/ac00182a014
Brain JD Mechanisms, measurement, and significance of lung macrophage function Environ Health Perspect 1992 97 5 10 1396468
Zhang P Summer WR Bagby GJ Nelson S Innate immunity and pulmonary host defense Immunol Rev 2000 173 39 51 10719666 10.1034/j.1600-065X.2000.917306.x
Stumbles PA Upham JW Holt PG Airway dendritic cells: Co-ordinators of immunological homeostasis and immunity in the respiratory tract APMIS 2003 111 741 755 12974776 10.1034/j.1600-0463.2003.11107806.x
Tjelle TE Lovdal T Berg T Phagosome dynamics and function Bioessays 2000 22 255 263 10684585 10.1002/(SICI)1521-1878(200003)22:3<255::AID-BIES7>3.0.CO;2-R
Nyberg K Johansson A Camner P Intraphagosomal pH in alveolar macrophages studied with fluorescein- labeled amorphous silica particles Exp Lung Res 1989 15 49 62 2537190
Mullins C Bonifacino JS The molecular machinery for lysosome biogenesis Bioessays 2001 23 333 343 11268039 10.1002/bies.1048
Möller W Barth W Kohlhäufl M Häussinger K Stahlhofen W Heyder J Human alveolar long-term clearance of ferromagnetic iron-oxide microparticles in healthy and diseased subjects Exp Lung Res 2001 27 547 568 11597117 10.1080/019021401753181827
Valerius NH Stendahl OI Hartwig JH Stossel TP Distribution of actin-binding protein and myosin in neutrophils during chemotaxis and phagocytosis Adv Exp Med Biol 1982 141 19 28 7115550
Stossel TP On the crawling of animal cells Science 1993 260 1086 1094 8493552
Janmey PA The cytoskeleton and cell signaling: component localization and mechanical coupling Physiol Rev 1998 78 763 781 9674694
Hirokawa N Kinesin and dynein superfamily proteins and the mechanism of organelle transport Science 1998 279 519 526 9438838 10.1126/science.279.5350.519
Nemoto I Ogura K Toyotama H Estimation of the energy of cytoplasmic movements by magnetometry: effects of temperature and intracellular concentration of ATP IEEE Trans Biomed Eng 1989 36 598 607 2731946 10.1109/10.29454
Qian H A simple theory of motor protein kinetics and energetics. II Biophys Chem 2000 83 35 43 10631478 10.1016/S0301-4622(99)00121-0
Bellomo G Mirabelli F Oxidative stress and cytoskeletal alterations Ann N Y Acad Sci 1992 663 97 109 1482106
Möller W Takenaka S Rust M Stahlhofen W Heyder J Probing mechanical properties of living cells by magnetopneumography J Aerosol Med 1997 10 173 186 10174196
Möller W Nemoto I Matsuzaki T Hofer T Heyder J Magnetic phagosome motion in J774A.1 macrophages: influence of cytoskeletal drugs Biophys J 2000 79 720 730 10920006
Möller W Kreyling WG Kohlhäufl M Häussinger K Heyder J Macrophage functions measured by magnetic microparticles in vivo and in vitro J Magn Magn Mater 2001 225 218 225 10.1016/S0304-8853(00)01261-0
Gehr P Brain JD Bloom SB Valberg PA Magnetic particles in the liver: a probe for intracellular movement Nature 1983 302 336 338 6835369 10.1038/302336a0
Valberg PA Butler JP Magnetic particle motions within living cells. Physical theory and techniques Biophys J 1987 52 537 550 3676435
Aizawa Y Takata T Karube H Tatsumi H Inokuchi N Kotani M Chiyotani K Magnetometric evaluation of the effects of gallium arsenide on the clearance and relaxation of iron particles Ind Health 1993 31 143 153 7512540
Okada M Karube H Niitsuya M Aizawa Y Okayasu I Kotani M In vitro toxicity of gallium arsenide in alveolar macrophages evaluated by magnetometry, cytochemistry and morphology Tohoku J Exp Med 1999 189 267 281 10739163 10.1620/tjem.189.267
Ralph P Nakoinz I Phagocytosis and cytolysis by a macrophage tumour and its cloned cell line Nature 1975 257 393 394 1101071 10.1038/257393a0
Möller W Scheuch G Sommerer K Heyder J Preparation of spherical monodisperse ferrimagnetic iron-oxide microparticles between 1 and 5 μm diameter J Magn Magn Mater 2001 225 8 16 10.1016/S0304-8853(00)01221-X
Kobzik L Lung macrophage uptake of unopsonized environmental particulates. Role of scavenger-type receptors J Immunol 1995 155 367 376 7541421
Roth C Ferron GA Karg E Lentner B Schumann G Takenaka S Heyder J Generation of ultrafine particles by spark discharging Aerosol Sci Technol 2004 38 228 235 10.1080/02786820490247632
NIST Standard Reference Material 1649a, Urban Dust, Certificate of Analysis 1998 Gaithersburg, MD 20899, USA: National Institute of Standards and Technology (NIST), USA 1 4
NIST Standard Reference Material 1650a, Diesel Particulate Matter, Certificate of Analysis 2000 Gaithersburg, MD 20899 USA: National Institute of Standards and Technology (NIST), USA 1 7
Dagher Z Garcon G Gosset P Ledoux F Surpateanu G Courcot D Aboukais A Puskaric E Shirali P Pro-inflammatory effects of Dunkerque city air pollution particulate matter 2.5 in human epithelial lung cells (L132) in culture J Appl Toxicol 2005 25 166 175 15744755 10.1002/jat.1050
Gillissen A Jaworska M Orth M Coffiner M Maes P App EM Cantin AM Schultze-Werninghaus G Nacystelyn, a novel lysine salt of N-acetylcysteine, to augment cellular antioxidant defence in vitro Respir Med 1997 91 159 168 9135855 10.1016/S0954-6111(97)90052-4
Nemoto I A model of magnetization and relaxation of ferrimagnetic particles in the lung IEEE Trans Biomed Eng 1982 29 745 752 7173941
Nemoto I Möller W A viscoelastic model of phagosome motion within cells based on cytomagnetometric measurements IEEE Trans Biomed Eng 2000 47 170 182 10721624 10.1109/10.821751
Valberg PA Feldman HA Magnetic particle motions within living cells. Measurement of cytoplasmic viscosity and motile activity Biophys J 1987 52 551 561 3676436
Möller W Takenaka S Buske N Felten K Heyder J Relaxation of ferromagnetic nanoparticles in macrophages: in vitro and in vivo studies J Magn Magn Mater 2005 293 245 251 10.1016/j.jmmm.2005.02.017
Dasenbrock C Peters L Creutzenberg O Heinrich U The carcinogenic potency of carbon particles with and without PAH after repeated intratracheal administration in the rat Toxicol Lett 1996 88 1 3 10.1016/0378-4274(96)03712-5
Pollard TD Borisy GG Cellular motility driven by assembly and disassembly of actin filaments Cell 2003 112 453 465 12600310 10.1016/S0092-8674(03)00120-X
Wang N Ingber DE Control of cytoskeletal mechanics by extracellular matrix, cell shape, and mechanical tension Biophys J 1994 66 2181 2189 8075352
Hartwig JH Janmey PA Stimulation of a calcium-dependent actin nucleation activity by phorbol 12-myristate 13-acetate in rabbit macrophage cytoskeletons Biochim Biophys Acta 1989 1 64 71 2642389
Glogauer M Arora P Chou D Janmey PA Downey GP McCulloch CA The role of actin-binding protein 280 in integrin-dependent mechanoprotection J Biol Chem 1998 273 1689 1698 9430714 10.1074/jbc.273.3.1689
Brown DM Donaldson K Stone V Role of calcium in the induction of TNFa expression by macrophages on exposure to ultrafine particles Ann Occup Hyg 2002 46 S219 222 10.1093/annhyg/mef028
Yin HL Hartwig JH Maruyama K Stossel TP Ca2+ control of actin filament length. Effects of macrophage gelsolin on actin polymerization J Biol Chem 1981 256 9693 9697 6270098
Stossel TP Hartwig JH Yin HL Southwick FS Zaner KS The motor of leukocytes Fed Proc 1984 43 2760 2763 6540718
Hu TH Bei L Qian ZM Shen X W-7 primes or inhibits the fMLP-stimulated respiratory burst in human neutrophil by concentration-dependent dual expression of the formyl peptide receptors on cell surface Biochim Biophys Acta 2000 1496 243 251 10771092 10.1016/S0167-4889(00)00017-3
Gorga JA Fishbaugher DE VanBuren P Activation of the calcium-regulated thin filament by myosin strong binding Biophys J 2003 85 2484 2491 14507711
Nguyen H Higuchi H Motility of myosin V regulated by the dissociation of single calmodulin Nat Struct Mol Biol 2005 12 127 132 15665867 10.1038/nsmb894
Antonicelli F Brown D Parmentier M Drost EM Hirani N Rahman I Donaldson K MacNee W Regulation of LPS-mediated inflammation in vivo and in vitro by the thiol antioxidant Nacystelyn Am J Physiol Lung Cell Mol Physiol 2004 286 L1319 1327 15136298 10.1152/ajplung.00329.2003
Duffin R Clouter A Brown DM Tran CL MacNee W Stone V Donaldson K The importance of surface area and specific reactivity in the acute pulmonary inflammatory response to particles Ann Occup Hyg 2002 46 242 245
|
16202162
|
PMC1262770
|
CC BY
|
2021-01-04 16:38:30
|
no
|
Part Fibre Toxicol. 2005 Oct 4; 2:7
|
utf-8
|
Part Fibre Toxicol
| 2,005 |
10.1186/1743-8977-2-7
|
oa_comm
|
==== Front
Reprod Biol EndocrinolReproductive biology and endocrinology : RB&E1477-7827BioMed Central London 1477-7827-3-451615015110.1186/1477-7827-3-45ResearchAugmented cell survival in eutopic endometrium from women with endometriosis: Expression of c-myc, TGF-beta1 and bax genes Johnson M Cecilia [email protected] Marisa [email protected] Alessandra [email protected] Ketty [email protected] Ariel [email protected] Margarita [email protected] M Angélica [email protected] Institute of Maternal and Child Research, School of Medicine, University of Chile; and San Borja Arriarán Clinical Hospital, Santiago, Chile2005 8 9 2005 3 45 45 29 6 2005 8 9 2005 Copyright © 2005 Johnson et al; licensee BioMed Central Ltd.2005Johnson et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
Endometriosis is a common gynaecological disorder characterized by the presence of endometrial tissue outside of the uterus. The fragments in normal menstruation are composed of necrotic and living cells, which do not survive in ectopic locations because of programmed cell death. The aim of this study was to evaluate if the balance between cell proliferation and apoptosis is changed in eutopic endometrium from women with endometriosis throughout the menstrual cycle by studying bax (pro-apoptotic), c-myc (regulator of cell cycle) and TGF-beta1 (involved in cell differentiation) genes.
Methods
Eutopic endometrium was obtained from: 30 women with endometriosis (32.8 +/- 5 years) and 34 fertile eumenorrheic women (36 +/- 5.3 years). We analyzed apoptosis (TUNEL: DNA fragmentation); cell proliferation (immunohistochemistry (IHC) for Ki67); c-myc, bax and TGF-beta1 mRNA abundance (RT-PCR) and TGF-beta1 protein (IHC) in endometrial explants.
Results
Cell proliferation strongly decreased from proliferative to late secretory phases in glands, but not in stroma, in both endometria. Positive staining in glands and stroma from proliferative endometrium with endometriosis was 1.9- and 2.2-fold higher than control endometrium, respectively (p < 0.05). Abundance of c-myc mRNA was 65% higher in proliferative endometrium from endometriosis than normal tissue (p < 0.05). TGF-beta1 (mRNA and protein) augmented during mid secretory phase in normal endometrium, effect not observed in endometrium with endometriosis. In normal endometrium, the percentage of apoptotic epithelial and stromal cells increased more than 30-fold during late secretory phase. In contrast, in endometrium from endometriosis, not only this increase was not observed, besides bax mRNA decreased 63% versus normal endometrium (p < 0.05). At once, in early secretory phase, apoptotic stromal cells increased 10-fold with a concomitant augment of bax mRNA abundance (42%) in endometria from endometriosis (p < 0.05).
Conclusion
An altered expression of c-myc, TGF-beta1 and bax was observed in eutopic endometrium from endometriosis, suggesting its participation in the regulation of cell survival in this disease. The augmented cell viability in eutopic endometrium from these patients as a consequence of a reduction in cell death by apoptosis, and also an increase in cell proliferation indicates that this condition may facilitate the invasive feature of the endometrium.
==== Body
Introduction
Endometriosis is a common gynaecological disorder, frequently associated with infertility and pelvic pain that occurs almost exclusively in menstruating women of reproductive age. Although little is known about its aetiology and pathogenesis, the theory of retrograde menstruation and metaplasia of the mesothelium or peritoneum and implantation of viable endometrial cells has been widely accepted [1,2]. The morphology of eutopic endometrium from women with endometriosis is similar to normal endometrium, but its physiology and biochemistry are different. Recent reports show an abnormal survival capability at the epithelial and stromal levels of the eutopic endometrium from patients with endometriosis that may result in their continuous growth [3,4]
In proliferative phase, the endometrium is characterized by proliferation of cells from the basal layer that respond to 17β-estradiol, reaching its maximum at ovulation time and then, falling due to the effects of progesterone with marked changes in glandular epithelium and stroma. Besides steroids, other factors are involved in these processes. In fact, after the priming of estradiol, the expression of transforming growth factor beta1 (TGF-β1) is up regulated in human endometrium coincidently with the increase of plasma progesterone concentration, being down regulated during progesterone withdrawal. Mashburn et al. (1994) [5] reported that the highest concentration of this cytokine in the stroma was during the secretory phase and suggested that it may regulate epithelial cell proliferation and differentiation (for revision see Godkin and Dore, 1998) [6]. In epithelial cell lines, TGF-β plays a central role in maintaining the homeostasis through limitation of growth by repression of growth-promoting transcription factors, such as c-Myc [7]. As known, c-Myc is a transcription regulator that can both activate or inhibit gene expression in favour of cell proliferation.
It is well established that apoptosis is a temporary physiologic process by which tissues eliminate dysfunctional cells by a regulated mechanism that involves a sequence of intracellular molecular events including members of the Bcl-2 family, such as Bax, an inducer of apoptosis. In human endometrium, Bax is expressed differentially throughout the menstrual cycle, with an expression mainly in late phase according to the number of apoptotic cells detected at the end of the menstrual cycle [8,9].
The aim of the present study was to evaluate whether the balance between cell proliferation and apoptosis is changed in eutopic endometrium from women with endometriosis throughout the menstrual cycle. For this purpose, we evaluated: cell proliferation by the detection of Ki67 and c-myc mRNA abundance, apoptosis by DNA fragmentation detection and bax mRNA abundance, and the expression of TGF-β1 in eutopic endometrium obtained from women with and without endometriosis.
Materials and methods
Subjects
Eutopic endometrial tissue was obtained from 30 women undergoing laparatomy for endometriosis associated with chronic pelvic pain, severe dysmenorrhea and infertility, and from 34 eumenorrheic women undergoing laparoscopy for tubal sterilization. The age of these women was 32.8 ± 5 years; range 23 to 40 years and 36 ± 5.3 years; range 25 to 43 years, respectively (p < 0.05). Both groups of patients had normal body mass index (BMI ≤ 25) and they were without hormonal treatment and at least 3 months without contraceptive pills before surgery. The endometrial samples were obtained with a Pipelle suction curette from the corpus of the uterus and washed several times with ice-cold phosphate buffered saline (PBS) to remove blood. Each patient signed a written informed consent and the Institutional Ethic Committee approved this study.
The endometria were dated according to established criteria [10] by an experienced histopathologist and classified as proliferative (days 1–14, women without (n = 10) or with (n = 8) endometriosis) or secretory endometria (days 15–28, women without (n = 24) or with (n = 22) endometriosis). For both groups of endometria, the secretory phase was divided in early (days 15–18, n = 15), mid (days 19–23, n = 18) and late (days 24–28, n = 13). Once obtained, the tissue was cut into slices and some pieces were immediately frozen in liquid nitrogen for mRNA preparation or placed in 4% formalin/PBS (pH 7.2–7.4) for histological evaluation, cell death determination and immunohistochemistry.
Immunohistochemistry of Ki67 and TGF-β1
Sections (4 to 6 μm thick) of human endometrial tissue at different stages of the menstrual cycle were deparaffinized in xylol and hydrated gradually through graded alcohols. The sections for Ki67 detection were incubated in 10 mmol/L citrate buffer (pH 6.0) at 98°C for 20 min. All slides were quenched in 3% H2O2 for 5 min at room temperature, then blocked with 2% BSA in PBS (w/v) for 1 h at room temperature. Primary antibodies for Ki67 (1:75, Dako, Carpinteria, LA) and TGF-β1 (1:1000, CBL 778 clon TB21; Cymbus Biotechnology LTD, Hamts, NF) were applied to the samples and incubated at 37°C for 2 or 1 h, respectively. Immunodetection was performed by using the streptavidin-biotin peroxidase system (LSAB 2, Dako), diaminobenzidine (DAB) as chromogen and counterstained with hematoxylin. The immunohistochemical evaluation of Ki67 and TGF-β1 was determined as the percentage of positive stained cells. In all cases, three blinded observers evaluated at least 1,000 cells in one section from each sample.
RNA preparation, cDNA synthesis, Reverse Transcription-Polymerase Chain Reaction
Total RNA was isolated from frozen proliferative and secretory endometria using RNA-solv Reagent (Omega Bio-teck, Lilburn, GA) plus glycogen (Chemicon International, Inc., Temecula, CA); then the purified pellet was resuspended in diethylpyrocarbonate-treated water. Complementary DNA (cDNA) was synthesized from 2 μg of total RNA digested previously by DNase I (Fermentas AB, Vilnius, Lithuania) using random primers (Invitrogen, Carlsbad, CA) and 200 U revertAid H Minus M-MuLV reverse transcriptase (Fermentas) following the manufacturer's instructions.
The amplification of bax mRNA was performed as indicated in Johnson et al. (2004a) [11]; c-myc, and TGF-β1 mRNAs were studied using specific pairs of primers: c-myc (330 bp), NID: g34815, upstream 5'-GAT TCT CTG CTC TCC TCG A-3' and downstream 5'-CTC TGA CAC TGT CCA ACT TG-3'; TGF-β1 (307-bp), NID:gi 10863872, upstream 5'-CAC CAA CTA TTG CTT CAG C-3' and downstream 5'-GAT CAT GTT GGA CAG CTG-3'. Two μl of cDNA per reaction were adjusted to a total volume of 25 μL by adding PCR buffer containing 3 mmol/L MgCl2, 0.125 U of Taq DNA polymerase (Invitrogen), 0.25 mmol/L nucleotide mix and 0.4 μmol/L of each specific human primer (Invitrogen). As internal control, 18S rRNA cDNA [12] was amplified in each sample in the same conditions described above. The reaction was performed in ThermalCycler PT-100 (MJ Research Inc. Watertown, MA) at denaturation: 94°C for 45 sec; annealing: 55°C for 60 sec for c-myc, bax and 18S rRNA, or 53°C for TGF-β1; extension: 72°C for 60 sec, and repeat for 18 cycles for 18S rRNA, 30 cycles for c-myc and 33 cycles for bax and TGF-β1. To determine that the amplification of all the genes were within a linear range, we previously evaluated the linearity of amplification of the corresponding transcripts in human endometria explants and the number of cycles were then decided.
Amplified fragment products were visualized on a 1.0% agarose gel using ethidium bromide staining. Semiquantification of PCR-products was performed by image analysis (Kodak EDAS 290 Electrophoresis Documentation and Analysis System, Kodak 1D Image Analysis Software, Rochester, NY), and its identity was confirmed by sequencing (ABI PRISM model 310 version 3.4 automatic sequencer; Perkin Elemer cetus, Norwak, CT).
Apoptosis detection system
DNA fragmentation was assessed by TUNEL according to the specifications of the manufacturer and adapted by us [13]. Briefly, paraffin tissue sections mounted on silane-coated slides (4–6 μm thick) were deparaffinized with xylol, rehydrated and fixed with 4% methanol-free formaldehyde/PBS; then deproteinated for 10 min (20 μg/mL proteinase K in PBS), washed and post-fixed again. To avoid artificial DNA fragmentation, only nuclease-free solutions were used during the whole procedure.
The labeling of 3'-OH ends of fragmented DNA was performed by incubation with fluorescein-12-deoxy-UTP using 12.5 U terminal deoxynucleotidyl transferase (TdT) for 1 h at 37°C in a humidified chamber. As positive control, normal endometrium pre-treated with DNase displayed nuclear positive reactivity in the majority of cells, demonstrating the incorporation of the fluorescent nucleotide at 3'-OH ends to the fragmented DNA. To estimate non-specific binding and autofluorescence, negative controls were included in all assays where tissue sections were incubated without the TdT enzyme. Slides were analyzed by fluorescence microscopy with a wide-band excitation barrier filter suitable for analyzing both green (fluorescein labelled fragmented DNA) and red (propidium iodide counterstain) fluorescence. Three independent observers selected different optical fields in a random manner to determine the percentage of positive staining in at least 1,000 cells for each sample. The assessment of DNA fragmentation by the TUNEL analysis has been shown to detect mainly apoptosis, although positive staining in necrotic cells cannot be ruled out. Therefore, in this study we confirmed the presence of morphological apoptosis by light microscopy and the absence of necrosis was established by morphological criteria.
Statistical analysis
Results were expressed as mean ± SEM of the number of experiments indicated in the figure legends. The age of the women was expressed as mean ± SD. Data were statistically analyzed by Student's t-test (age women) or one-way ANOVA followed by Tukey's Multiple Comparison test (Tables and Figures). Difference was considered statistical significant at p less than 0.05.
Results
Cell proliferation
Nuclear immunostaining of Ki67 was detected in the cell nucleus of both epithelial and stromal compartments during the menstrual cycle in eutopic endometrium of women with or without endometriosis (Figure 1A–D). Cell proliferation decreased from proliferative to mid and late secretory phases in the normal epithelial gland of both endometria (p < 0.05), with no important changes in the stromal compartment (Table 1). The positive staining observed in epithelial and stromal proliferative endometrium obtained from women with endometriosis was 1.9- and 2.2-fold higher than normal proliferative endometrium, respectively (p < 0.05). A tendency of higher cell proliferation was observed in epithelium from endometriosis than in normal epithelium obtained during early and mid secretory phases (Table 1).
Figure 1 Immunohistochemical staining for Ki67, TGF-β1 and apoptosis by TUNEL in explants of human endometrium throughout the menstrual cycle obtained from normal (A, C, E, G, I, J and K) and endometriosis (B, D, F, H and L) women. Representative human endometrium explants from (A, B, E, F and I) proliferative phase and (C, D, G, H, J, K and L) secretory phase of the menstrual cycle with positive immuno-staining for (A-D) Ki67 and (E-H) TGF-β1 or positive DNA fragmentation by TUNEL (K and L). Immunohistochemitry (I) and TUNEL (J) negative controls. Cell nuclei are stained with haematoxylin (immunohistochemitry) or propidium iodine (TUNEL). g: glandular, s: stroma, rc: red blood cell. Magnification, 400×.
Table 1 Percentage of Ki67 in human eutopic endometrium obtained from women without and with endometriosis during the menstrual cycle.
Normal Endometriosis
Gland Stroma Gland Stroma
Proliferative 12.9 ± 4.2 7.7 ± 2.4 27.3 ± 5.3* 16.9 ± 3.8*
Secretory Early 12.4 ± 5.7 5.3 ± 0.6 26.1 ± 5.2 9.3 ± 2.0
Secretory Mid 0.8 ± 0.1# ° 4.3 ± 0.8 10.8 ± 5.7° 5.7 ± 1.6#
Secretory Late 1.0 ± 0.7# ° 10.9 ± 2.4 1.3 ± 0.9# ° 10.5 ± 3.0
Percentage of positive cells was performed in eutopic endometrium explants obtained from women without (normal) and with endometriosis at different stages of the menstrual cycle as indicated in Material and Methods. Results are given as mean ± SEM from: 8 and 7 endometria without and with endometriosis in each stage, respectively. *p < 0.05 vs. normal endometrium. #p < 0.05 vs. proliferative phase. °p < 0.05 vs. early secretory phase.
Detection of TGF-β1 and c-myc in eutopic endometrium obtained from women with and without endometriosis during the menstrual cycle
During the menstrual cycle, the mRNA of TGF-β1 increased more than 50% in mid and late secretory phases in normal endometrium (p < 0.05), effect that was not detected in eutopic endometrium obtained from women with endometriosis, remaining unchanged throughout the cycle (Table 2).
Table 2 TGF-β1 amplification from endometrium cDNA of women without and with endometriosis during the menstrual cycle.
Normal Endometriosis
Proliferative 1.12 ± 0.1 0.92 ± 0.2
Secretory Early 0.82 ± 0.2 0.86 ± 0.1
Secretory Mid 1.56 ± 0.2# 1.10 ± 0.1*
Secretory Late 1.49 ± 0.1# 1.05 ± 0.1*
Data correspond to amplification rate relative to 18S rRNA. Results are given as mean ± SEM from: 8 and 5 proliferative; 7 and 8 early; 4 and 8 mid and 4 late secretory endometria obtained from women without (normal) and with endometriosis, respectively. *p < 0.05 vs. normal endometrium; #p < 0.05 vs. early secretory phase.
Positive TGF-β1 immunostaining was homogenously distributed in the cytoplasm of stromal cells, and a weak intensity was observed in glands of both types of endometrium (Figure 1E–H). The cytokine was also detected as secretion in the glands lumen and extracellular matrix, besides an intense brown staining in blood vessels and red blood cells (1E and 1H). In normal endometrium, the highest percentage of TGF-β1 positive stromal cell was 53 ± 9% at mid secretory phase with an increase by 2-fold compared to proliferative phase (p < 0.05), whereas no differences were observed in epithelial cells. In eutopic endometrium from endometriosis, the percentage of TGF-β1 positive proliferative stromal cells was similar to normal proliferative stromal cells (28.18 ± 10.6 and 27.3 ± 6.5, respectively) and did not change during the menstrual cycle. It was almost undetected in eutopic endometrium gland from endometriosis.
The abundance of c-myc mRNA was not significantly changed throughout the menstrual cycle in both types of endometrium; however, the proto-oncogen was 65% higher in proliferative endometrium with endometriosis than normal endometrium, respectively (p < 0.05) (Figure 2).
Figure 2 PCR amplification from endometrium cDNA of women without (Normal) and with endometriosis using primers for c-myc (330-bp) and 18S rRNA (192-bp). Representative gel is shown. Graph illustrates the corresponding amplification relative to 18S rRNA and the results are given as mean ± SEM from: 6 and 6 proliferative (P); 6 and 5 early secretory (ES); 5 and 4 mid secretory (MS) and 5 and 4 late secretory (LS) eutopic endometria obtained from women without and with endometriosis, respectively. (-) PCR amplification without template. *p < 0.05 vs. normal endometria.
Cell death by apoptosis
DNA fragmentation was strongly detected by TUNEL mostly in epithelial and stromal cells in normal endometrium from late secretory phase (Figure 1K), effect that was not observed in late endometrium from women with endometriosis (Figure 1L). The percentage of apoptotic cells in late normal endometrium increase more than 30-fold in both cell compartments compared to proliferative endometrium (p < 0.05) (Figure 3A and 3B). In proliferative, early and mid secretory endometria, a reduced number of positive apoptotic cells were detectable in both groups of women. Interestingly, in early secretory endometrium from women with endometriosis, the number of apoptotic cells was increased respect to normal endometrium by 10-fold in stromal cells (p < 0.05) (Figure 3A and 3B).
Figure 3 Percentage of apoptotic cells in epithelial (A) and stromal cell (B) compartments in human endometrium explants obtained from normal women and women with endometriosis. Histological sections were evaluated by TUNEL technique (see text). The results are given as mean ± SEM from: 5 and 6 proliferative; 5 and 6 early secretory; 4 and 5 mid secretory; 7 and 6 late secretory eutopic endometria obtained from women without (normal) and with endometriosis, respectively. *p < 0.05 vs. normal endometria. °p < 0.05 vs. early secretory phase. #p < 0.05 vs. late secretory phase.
The pro-apoptotic gene bax was amplified in both groups of endometrium (Figure 4). The abundance of bax mRNA increased at the end of the menstrual cycle in normal endometrium. Although no significant difference was observed in proliferative and mid secretory phases, bax mRNA increased 42% in early and reduced 63% in late secretory endometria from women with endometriosis respect to normal endometrium, respectively (p < 0.05) coincident to the results obtained for apoptotic index.
Figure 4 PCR amplification from endometrium cDNA of women without (Normal) and with endometriosis using primers for bax (334-bp) and 18S rRNA (192-bp). Representative gel is shown. Graph illustrates the corresponding amplification relative to 18S rRNA (192-bp) and the results are given as mean ± SEM from 5 and 5 proliferative (P); 6 and 5 early secretory (ES); 5 and 6 mid secretory (MS) and 4 late secretory (LS) eutopic endometria obtained from women without (normal) and with endometriosis, respectively. (-) PCR amplification without template. *p < 0.05 vs. normal endometria. #p < 0.05 vs. proliferative phase.
Discussion
It is known that eutopic endometrium from women with endometriosis presents physiological differences with normal endometrium. In this study we show that eutopic endometrium from these patients has a significant increase in cell survival, and alterations on DNA fragmentation and on the expression of factors involved in cell proliferation, cell differentiation and programmed cell death induction, such as Ki67, c-myc, TGF-β1 and bax.
In agreement with previous reports and in accordance with results of light microscopy [10,14,15], our data show that epithelial cells from normal endometrium exhibit a higher proliferation rate than stromal cells during proliferative and early secretory phases, stages of the menstrual cycle regulated mainly by estradiol. Although eutopic endometrium from patients with endometriosis present a similar Ki67 distribution throughout the menstrual cycle, a significant augment of cells into cell cycle compared to normal endometrium was observed, in accordance with other investigators [16-18]. Even though Ki67 is expressed in the cell during M, G1, S and G2 phases of cell cycle and is absent in resting cells (Go), a good correlation between Ki67 and mitotic indices has been reported in human endometrium [19]. In contrast, other groups [4,20] have reported no differences between normal and eutopic endometria from women with endometriosis, probably due to the use of different techniques and the detection of proliferating cell nuclear antigen (PCNA), a nuclear protein restricted to S phase.
Several genes involved in cell proliferation, such as the proto-oncogene c-myc are up-regulated by estradiol in different cell types that include human breast cancer cell line MCF-7 [21,22], rat hepatocyte [23], and pituitary and somatolactotrophic cell line GH3 [24]. Little information is available of the proto-oncogen c-myc in the endometrium. The presence of the protein c-Myc in nuclei and cytoplasm of endometrial glands and nuclei stroma from eutopic and ectopic endometria has been reported [25], even though its role in the endometrium is not clear. In the present investigation, the c-myc mRNA abundance was not modified throughout the menstrual cycle in both types of endometria studied, although it was augmented in eutopic endometrium from women with endometriosis compared to normal women. Several findings suggest that in endometriosis, ectopic endometrium grows and regresses in an estrogen-dependent manner. Even more, eutopic endometrium of women with endometriosis contains not only estradiol receptors, but also P450Arom, the enzyme that catalyzes the conversion of androgen to estrogen, provoking an estrogenic micro-environment in this tissue that may act through intracrine pathways [26-28]. Therefore, the estrogenic microenvironment could induce the expression of the transcription factor c-myc and facilitates cell proliferation in the eutopic endometria in this disease. This effect may be improved by the reduction of TGF-β1 detected in the endometrium of these patients, as probable consequence of the loss of the negative regulation on P450AROM described in choriocarcinoma cell line (JEG-3) [29].
On the other hand, TGF-βs are likely candidates to partially mediate the effects of progesterone on the expression of other proteins, which have been suggested to be involved in the regulation of endometrial cell proliferation and differentiation [5,6]. In addition, the cyclic expression of TGF-β1, mainly observed in stromal cells during the secretory phase as reported in this study, is in accord to progesterone receptor location [30]. In several cell types, TGF-β1 plays a role in maintaining tissue homeostasis by modulating genes involved in the arresting of cell cycle including down-regulation of c-Myc [7,31]. Even more, in rabbit uterine epithelial cells, a coordinated but inverse regulation of cell proliferation and apoptosis by this cytokine was described [32]. The fact that the cyclicity of TGF-β1 observed in normal endometrium was absent in eutopic endometrium from women with endometriosis indicates an impaired local production in these patients that may favors the cell proliferation. More studies are required to understand the complex and strict interactions between these molecules involved in the deregulation on cell proliferation in the eutopic endometrium of these women.
Proliferation and apoptosis are important biologic processes in the endometrium remodeling with cell cycle changes that take place during the menstrual cycle [14,33]. In the late secretory phase, and related with a decrease of endometrial progesterone receptor and plasma progesterone concentration, cell death by apoptosis increases in the functional endometrial layer, and at the time of menstruation it becomes necrotic and hypoxic and is shed. In the present investigation, the expected increase of apoptotic cells during late secretory phase was not observed in eutopic endometrium from women with endometriosis. Concomitantly to the absence of apoptosis in those endometria, the pro-apoptotic gene bax mRNA abundance was unchanged during the menstrual cycle, at difference at the increase exhibited in normal endometrium, suggesting that the reduction of bax expression may be a mechanism that could explain the decreased incidence of apoptosis in endometriosis. In agreement with these results, recent publications [3,34,35] also reported the absence of apoptotic cells and the increase of Bcl-2 and the reduction of Bax expression in eutopic endometrium from women with endometriosis, although Béliard et al. (2004) [4] were unable to find these differences between patients with endometriosis and normal women. Curiously, we detected that both bax mRNA abundance and apoptotic cells were increased in early secretory phase in eutopic endometrium from women with endometriosis but not from normal women. The implication of this augmented cell death in eutopic endometrium from the patients needs more studies.
It is well known that menstrual fragments are composed of both necrotic and living cells, which do not survive in ectopic locations because of its programmed cell death [36,37]. The loss of cell death by apoptosis in late secretory phase in these patients seems to be consistent with cell ectopic survival and implantation of endometrial cells in the peritoneal cavity, effect that is independent of the disease stages as it was reported by Dmowski et al. (2001) [34]. The altered expression of c-myc, TGF-β1 and bax observed in eutopic endometrium from women with endometriosis, suggests their participation in the deregulation of the cell survival in this tissue. It is known that the deregulation of apoptotic signaling can play a primary or secondary role in various diseases; in fact, an insufficient apoptosis degree contributes to pathogenic processes including cancer. The present investigation shows that cell viability is augmented in eutopic endometrium from women with endometriosis compared to normal endometrium, primarily due to a reduction in cell death by apoptosis, and also, an increase in cell proliferation. These data strongly suggest that this condition may facilitate the invasive character of the endometrium in these patients. Until now, it is not clear why misplaced endometrial cells from healthy women do not implant and do not develop into endometriotic lesions as occurs in women that develop endometriosis, indicating that some factor(s) which facilitate(s) their survival and implantation may be involved. However, we cannot rule out the possible presence of endometrium cells that belong to basal layer, aspect not studied in the present work.
Conclusion
The altered expression of c-myc, TGF-β1 and bax observed in eutopic endometrium from women with endometriosis, suggests the participation of these molecules in the regulation of the cell survival in this disease. The augmented cell viability in eutopic endometrium from these patients as consequence of a reduction in cell death by apoptosis, and also an increase in cell proliferation indicates that this condition may facilitate the invasive character of the endometrium. The estrogenic microenvironment reported in this tissue in women with endometriosis may explain this improvement of proliferation/apoptosis balance.
Contribution of the authors
MCJ conceived and designed the study, participated in the analysis and interpretation of data, and drafted the manuscript. MT carried out the Ki67 IHQ and TUNEL, participated in the analysis of data. AA carried out TGF-β1 IHQ and participated in analysis of data. KB carried out the mRNA studies and participated in the analysis of data, AF carried out the statistical analysis and helped to draft the article, MV participated revising the manuscript critically and helped to draft the article, MAB participated in the analysis and interpretation of data and helped to draft the manuscript. MT, AA and KB performed the microscopic evaluation.
Acknowledgements
The authors are grateful to Juan Carlos Barros M.D. and Alberto Palomino M.D. (National Health Service, Chile) for their role in recruiting and performing surgical procedures for patients providing endometrial samples; Fernando Gabler M.D. (School of Medicine, University of Chile, Chile) for patient endometrial dating; and the women who donated tissue.
This work was supported by Departamento de Investigación y Desarrollo (DID) #ENL03/05 of University of Chile, and Fondo Nacional de Ciencias y Tecnología (FONDECYT) # 1040412 of Chile.
==== Refs
Sampson J Peritoneal endometriosis due to the menstrual dissemination of endometrial tissue into the peritoneal cavity Am J Obstet Gynecol 1927 14 422 469
Fujii S Secondary mullerian system and endometriosis Am J Obstet Gynecol 1991 165 219 225 1853900
Meresman GF Vighi S Buquet RA Contreras-Ortiz O Tesone M Rumi LS Apoptosis and expression of Bcl-2 and Bax in eutopic endometrium from women with endometriosis Fertil Steril 2000 74 760 766 11020520 10.1016/S0015-0282(00)01522-3
Béliard A Noel A Foidar JM Reduction of apoptosis and proliferation in endometriosis Fertil Steril 2004 82 80 85 15236993 10.1016/j.fertnstert.2003.11.048
Mashburn PB Arici M Casey L Expression of transforming growth factor-β messenger ribonucleic acid and the modulation of deoxyribonucleic acid synthesis by transforming growth factor-β1 in human endometrial cell Am J Obstet Gynecol 1994 170 1152 1158 8166201
Godkin JD Dore JJ Transforming growth factor β and the endometrium Rev Reprod 1998 3 1 6 9509983 10.1530/ror.0.0030001
Siegel PM Shu W Massagué J Mad upregulation and id2 repression accompany transforming growth factor (TGF)-β-mediated epithelial cell growth suppression J Biol Chem 2003 278 35444 354450 12824180 10.1074/jbc.M301413200
Vaskivuo TE Stenback F Karhumaa P Risteli Dunkel L Tapanainen JS Apoptosis and apoptosis-related proteins in human endometrium Mol Cell Endocrinol 2000 165 75 83 10940486 10.1016/S0303-7207(00)00261-6
Castro A Johnson MC Anido M Cortinez A Gabler F Vega M Role of nitric oxide and bcl-2 family genes in the regulation of human endometrial apoptosis Fertil Steril 2002 78 587 595 12215338 10.1016/S0015-0282(02)03304-6
Noyes RW Hertig AT Rock J Dating the endometria biopsy Fertil Steril 1950 1 3 25
Johnson MC Maliqueo M Boric MA Villavicencio A Vantman D Vega M Differential in vitro actions of nitric oxide on human endometrial cell survival Fetil Steril 2004 81 176 184 10.1016/j.fertnstert.2003.05.018
Schmittgen TD Zakrajsek BA Effect of experimental treatment on housekeeping gene expression: validation by real-time, quantitative RT-PCR J Biochem Biophys Methods 2000 46 69 81 11086195 10.1016/S0165-022X(00)00129-9
Vega M Urrutia L Iñiguez G Gabler F Devoto L Johnson MC Nitric oxide induces apoptosis in the human corpus luteum in vitro Mol Human Reprod 2000 6 681 687 10.1093/molehr/6.8.681
Von Rango U Classen-Linke I Krusche CA Beier HM The receptive endometrium is characterized by apoptosis in the glands Human Reprod 1998 13 3177 3189 10.1093/humrep/13.11.3177
Mertens HJMM Heineman MJ Evers JLH The expression of apoptosis-related proteins Bcl-2 and Ki67 in endometrium of ovulatory menstrual cycles Gynecol Obstet Invest 2002 53 224 230 12186988 10.1159/000064569
Jones RK Searle RF Bulmer JN Apoptosis and Bcl-2 expression in normal human endometrium, endometriosis and adenomyosis Human Reprod 1998 13 3496 3502 10.1093/humrep/13.12.3496
Nisolle M Casanas-Roux F Donnez J Immunohistochemical analysis of proliferative activity and steroid receptor expression in peritoneal and ovarian endometriosis Fertil Steril 1997 68 912 919 9389825 10.1016/S0015-0282(97)00341-5
Meresman GF Augé L Barañao RI Lombardi E Tesone M Sueldo C Oral contraceptives suppress cell proliferation and enhance apoptosis of eutopic endometrial tissue from patients with endometriosis Fertil Steril 2000 77 1141 1147 12057719 10.1016/S0015-0282(02)03099-6
Brenner RM Slayden OD Rodgers WH Critchley HOD Carroll R Nie XJ Mah K Immunocytochemical assessment of mitotic activity with an antibody to phosphorylated histone H3 in the macaque and human endometrium Human Reprod 2003 18 1185 1193 10.1093/humrep/deg255
Jurgensen A Mettler L Volkov NI Parwaresch R Proliferative activity of the endometrium throughout the menstrual cycle in infertile women with and without endometriosis Fertil Steril 1996 66 369 375 8751731
Carrol JS Swabrick A Musgrove EA Sutherland RL Mechanism of growth arrest by c-myc antisense oligonucleotides in MCF-7 breast cancer cells: implications for the antiproliferative effects of antiestrogens Cancer Res 2002 62 3126 3131 12036924
Paruthiyil S Parmar H Kerekatte V Cunha GR Firestone GL Leitman DC Estrogen receptor beta inhibits human breast cancer cell proliferative and tumor formation by causing a G2 cell cycle arrest Cancer Res 2004 64 423 428 14729654 10.1158/0008-5472.CAN-03-2446
Lee CH Edwards AM Differential expression of c-fos and c-myc protooncogenes by estrogens, xenobiotics and other growth-stimulatory agents in primary rat hepatocytes Arch Toxicol 2003 77 150 159 12632255
Fujimoto N Igarashi K Kanno J Honda H Inoue T Identification of estrogen-responsive genes in the GH3 cell line by cDNA microarray analysis J Steroid Biochem Mol Biol 2004 91 121 129 15276619 10.1016/j.jsbmb.2004.02.006
Schenken RS Johnson JV Riehl RM c-myc proto-oncogene polypeptide expression in endometriosis Am J Obstet Gynecol 1991 164 1031 1037 1707594
Bulun SE Economos K Miller D Simpson ER CYP19 (aromatase cytochrome P450) gene expression in human malignat endometrial tumors J Clin Endocrinol Metab 1994 79 1831 1834 7989490 10.1210/jc.79.6.1831
Noble LS Simpson FR Johns A Bulun SF Aromatase expression in endometriosis J Clin Endocrinol Metab 1996 81 174 179 8550748 10.1210/jc.81.1.174
Johnson MC Pinto C Alves A Palomino A Fuentes A Boric MA Vega M P450 Arom y microambiente estrogénico en endometrios eutópicos de mujeres con endometriosis Rev Med Chile 2004 132 1475 1482 15743158
Luo S Wu D Peng Ch Transforming growth factor-β1 inhibits steroidogenesis in human trophoblast cells Mol Human Reprod 2002 8 318 325 10.1093/molehr/8.4.318
Mote PA Balleine RL McGowan EM Colocalization of progesterone receptors A and B by dual immunofluorescent histochemistry in human endometrium during the menstrual cycle J Clin Endocrinol Metab 1999 84 2963 2971 10443705 10.1210/jc.84.8.2963
Khanna A Concerted effect of transformation growth factor-β, cyclin inhibitor p21, and c-myc on smoth muscle cell proliferation Am J Physiol Heart Circ Physiol 2004 286 H1133 H1140 14766678 10.1152/ajpheart.00462.2003
Rotello RJ Lieberman RC Purchio AF Gerschenson LE Coordinated regulation of apoptosis and cell proliferation by transforming growth factor β1 in cultured uterine epithelial cells Proc Natl Acad Sci USA 1991 88 3412 3415 2014262
Kokawa K Shikone T Nakano R Apoptosis in the menstrual uterine endometrium during the menstrual cycle J Clin Endocrinol Metab 1996 81 4144 4147 8923873 10.1210/jc.81.11.4144
Dmowski WP Ding J Shen J Rana N Fernandez BB Braun DP Apoptosis in endometrial glandular and stromal cells in women with and without endometriosis Human Reprod 2001 16 1802 1808 10.1093/humrep/16.9.1802
Braun DP Ding J Shen J Rana N Fernandez BB Dmowski WP Relationship between apoptosis and the number of macrophages in eutopic endometrium from women with and without endometriosis Fertil Steril 2002 78 830 835 12372464 10.1016/S0015-0282(02)03334-4
Keetel WC Stein RJ The viability of the cast-off menstrual endometrium Am J Obstet Gynecol 1951 61 440 442 14799567
Bartosik D Jacobs SL Kelly LJ Endometrial tissue in peritoneal fluid Fertil Steril 1986 46 796 800 3780999
|
16150151
|
PMC1262771
|
CC BY
|
2021-01-04 16:37:12
|
no
|
Reprod Biol Endocrinol. 2005 Sep 8; 3:45
|
utf-8
|
Reprod Biol Endocrinol
| 2,005 |
10.1186/1477-7827-3-45
|
oa_comm
|
==== Front
Reprod Biol EndocrinolReproductive biology and endocrinology : RB&E1477-7827BioMed Central London 1477-7827-3-471616228210.1186/1477-7827-3-47ResearchCalcium alginate microencapsulation of ovarian follicles impacts FSH delivery and follicle morphology Heise Matthew [email protected] Richard [email protected] Alan J [email protected] Elizabeth A [email protected] McGowan Institute for Regenerative Medicine, University of Pittsburgh, 100 Technology Dr. Suite 200, Pittsburgh, PA 15219, USA2 Magee-Womens Research Institute, 204 Craft Ave, Pittsburgh PA, 15213, USA3 Department of Obstetrics, Gynecology and Reproductive Medicine, University of Pittsburgh, Pittsburgh PA 15213, USA2005 14 9 2005 3 47 47 19 4 2005 14 9 2005 Copyright © 2005 Heise et al; licensee BioMed Central Ltd.2005Heise et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
We have previously shown that suspension culture prevents follicle flattening and maintains three-dimensional follicle architecture better than culture on flat plates. However, many of the follicles cultured in suspension do eventually rupture, as basement membrane integrity is lost and the three-dimensional structure of the follicle is altered. Therefore, the objective of this study is to support three-dimensional follicle architecture during in vitro growth of ovarian follicles through encapsulation in calcium alginate, while maintaining responsiveness to FSH stimulation.
Methods
Preantral follicles (150 – 160 micrometers in diameter) were isolated from the ovaries of juvenile rats and grown in culture tubes or encapsulated in calcium alginate and grown in culture tubes. Previous studies revealed that follicles maintained structural integrity but did not grow as well when encapsulated in calcium alginate. In these studies, we evaluated the effect of calcium alginate on FSH-stimulated follicle growth, survival, and morphology in suspension culture. Follicles were grown under 5 culture conditions: 1) not encapsulated; with FSH in the medium, 2) encapsulated in the absence of FSH, grown in medium without FSH, 3) encapsulated with calcium alginate containing FSH but grown in medium without FSH, 4) encapsulated without FSH but grown in medium containing FSH and 5) encapsulated with calcium alginate containing FSH and in medium containing FSH. To assess growth rates, follicles were cultured for 72 hours and analyzed for follicle size increase and DNA content. Survival analysis for encapsulated and unencapsulated follicles was performed by constructing a Kaplan Meier survival curve of daily observations of intact follicle survival. Three-dimensional architecture was assessed histologically and by analysis of the pattern of connexin 43 expression in the cultured follicles.
Results
In the absence of FSH, follicle diameter increased by only 6.4%. When FSH was included in the alginate bead alone or the media alone, the follicle diameter increased by 13.5% and 19.9% respectively. This was greater than follicles cultured in the absence of FSH (p < 0.05), but less than that of the FSH-treated unencapsulated follicles (p < 0.05). However, when follicles were cultured with FSH included in both the media and the bead, a 32.6% increase in follicle diameter was observed, statistically no different than the growth rate of the unencapsulated follicles grown with FSH.
Conclusion
Microencapsulation supports three-dimensional follicle growth, but may limit access to hormones in the medium resulting in altered development compared to unencapsulated follicles. Inclusion of FSH in the alginate bead restores the follicle growth response to FSH, while also providing a scaffold of support for three-dimensional growth. The application of tissue engineering principles to the problems of follicle culture in vitro may provide advances applicable to fertility preservation in women and endangered species.
==== Body
Background
The culture of intact ovarian follicles is useful for the study of the regulation of folliculogenesis, but may also provide an alternative to ovarian transplant for the preservation of fertility [1]. Classically, ovarian follicle culture has been performed in culture dishes, on a flat surface. Although ovarian follicles from the mouse can readily be grown on culture plates [2,3], the maturation of follicles from larger mammals has proven to be much more difficult [1]. One cause of difficulty may be that larger follicles have different structural needs. Mouse follicles can ovulate at 400 μm in diameter [4], whereas rat preovulatory follicles are often greater than 800 μm in diameter [5]. Though this represents only a doubling of diameter, the volume of the rat preovulatory follicle is at least 8 times that of the mouse. The application of standard tissue engineering principles to the growth of ovarian follicles in vitro may provide additional tools necessary to overcome the structural challenges presented by follicles from larger mammals and make significant advances in this field.
It has been shown that suspension culture both enhances cell proliferation [6] and prevents follicle flattening and maintains three-dimensional follicle architecture better than culture on flat plates [7]. Though mouse follicles grown in a hanging drop system grew very well compared to follicles grown on flat plates, the follicle diameter was still well under 500 after 6 days of culture [6]. Rat follicles often rupture at 200 to 250 μm on flat plates. However in suspension culture follicles routinely maintain intact survival to about 400 μm. Past this size, basement membrane integrity is lost and the three-dimensional structure of the follicle is altered. One tissue engineering approach to this problem is to encapsulate the follicles. Microencapsulation has been used to provide structural support for a variety of tissues such as pancreatic islets [8] and thyroid follicles [9].
Alginate is one of the most commonly applied biomaterials for microencapsulation due to its biocompatibility, high affinity to water, and ability to form gels under mild conditions when in the presence of calcium ions [10-14]. Alginate is comprised of chains of alternating blocks of mannuronic acid, which contributes the elastic property of the gel; and guluronic acid, which contributes mechanical strength, stability, porosity, and gel forming properties [15,16]. Alginates are extracted from all species of brown algae and contain differing compositions of mannuronic acid/mannuronic acid, mannuronic acid/glucoronic acid, and glucoronic acid/glucoronic acid blocks offering a variation in strength and stability. Alginate gel beads are reported to have a high porosity range and only limit the diffusion of large proteins [12]. It has been reported that substrates of molecular weight less than 2 × 104, such as glucose, L-tryptophan (MW = 204), and α-lactoalbumin (MW = 1.54 × 104), are able to diffuse freely into and from calcium alginate beads at approximately the same diffusion rate as in water, while larger proteins, such as albumin (MW = 6.9 × 104), could not diffuse freely into the calcium alginate beads [17]. Although there is resistance for larger proteins (MW > 2 × 104) diffusing into these beads, diffusion from the bead into a surrounding solution devoid of the substrate is not hindered until the molecular weight of the substrate approaches 3 × 105 [11]. FSH has a molecular weight of 3 × 104 [18]. Thus, FSH is in the size range where hindrance may play a role in FSH availability to follicles encapsulated in calcium alginate beads.
In this study, the role of a calcium alginate scaffold on follicle growth, survival, and morphology in suspension culture was explored. Our previous studies suggested that alginate encapsulation slowed follicle growth [7]. Therefore, in these studies we have determined the growth rates of follicles in response to different delivery methods of FSH to encapsulated follicles. As a result, we have established the ability of calcium alginate to support three-dimensional follicular growth of intact follicles, but discovered that encapsulation may limit follicular access to FSH unless it is included in the alginate bead itself.
Materials and methods
Animals and Ovarian Dissection
All animal experiments were performed in accordance with National Institutes of Health guidelines and with institutional approval. Sprague-Dawley rats were obtained from Hilltop Lab Animals (Pittsburgh, PA) and housed under standard conditions. The animals were sacrificed by CO2 exposure and cervical dislocation. Ovaries were carefully dissected and placed immediately in warmed culture medium, consisting of Leibovitz L-15 Medium (Gibco BRL) with 1% bovine serum albumin (Sigma). The follicles were then mechanically dissected from the ovary as previously described [19]. All follicles used in the experiments were measured in two dimensions, using an inverted microscope fitted with an ocular micrometer. Only intact follicles that were between 150 and 160 microns in diameter were used in culture. There was no statistical difference in the starting follicle diameter of any of the groups.
Calcium Alginate Microencapsulation
After dissection, 20 to 30 follicles per group were transferred with glass pipettes to a solution of sodium alginate (1% w/v; Sigma) in distilled water. (For the groups that required FSH in the bead, follicles were transferred to a solution of sodium alginate containing recombinant Follicle Stimulating Hormone, rFSH, (Serono Laboratories, Geneva) at a concentration of 1 iu/mL.) The mixture of follicles in sodium alginate was slowly released through a 25-gauge needle as droplets falling into a beaker containing a stirred solution of CaCl2 (0.1 M). The droplets immediately gelled to form beads. A stream of 0.2 μm-filtered air was positioned at the tip of the needle to cut the mixture stream into small droplets to obtain beads with diameters between 250 μm to 400 μm. Beads containing individual follicles (Figure 1) were then removed from the beaker using glass pipettes and immediately transferred to media in 12 × 75 mm polypropylene culture tubes. Follicles were cultured one per tube.
Figure 1 Ovarian Follicle Encapsulated in a Calcium Alginate Bead. Scale bar represents 50 μm.
Follicle Culture
Culture media consisted of α-Minimal Essential Medium (Gibco BRL, Invitrogen Corporation, Grand Island, NY) with additives of 8-bromo-cGMP (5 mM), ITS+ (1 % solution of insulin, 10 mg/L; transferrin, 5.5 mg/L; linoleic acid, 4.7 mg/L; selenium, 5 mg/L), Pen/Strep (1 %, penicillin 100 U/ml, streptomycin 100 μg/ml), all from Sigma Chemical Co. (St. Louis, MO), and recombinant Follicle Stimulating Hormone, rFSH, (1 iu/ml; Serono Laboratories, Geneva). In the basal or negative control group, FSH was not included. Culture media was placed into 12 × 75 mm polypropylene culture test tubes (500 μl/tube) and cultured in 5 % CO2 and 37°C humidified incubator.
Suspension was attained by placing the 12 × 75 mm polypropylene culture tubes in a circular rotator plate (Glas-Col, Terre Haute, IN), having a diameter of 30.5 cm, which was rotated around its horizontal axis at rate between 8–15 rpm. Therefore, as the plate rotates, the tubes slowly orbit the axis of the plate and the follicle is maintained is fluid suspension. After 72 hours in culture, follicle diameter was measured once again and follicles were collected for DNA quantification and histological analysis. For survival analysis, follicles were observed daily for basement membrane integrity and signs of atresia. Atresia is readily apparent as darkening of the follicle appearance under the dissecting microscope [7,20]. Survival analysis cultures were continued for 7 days.
DNA Quantification
To verify that increased follicle size represented increased follicle cell number, DNA quantification was performed on the cultured follicles as previously described [7]. At 72 hours of culture, the follicles were released from the beads by allowing the capsules to dissolve in sterile PBS (pH 7.4). Follicles (n = 5 for each group) were then trypsinized and DNA was extracted from each individual follicle and quantified by using the fluorescent dye, Hoechst 33258 (bisbenzimidazole; Sigma), and a microplate fluorescence reader (Perkin Elmer Life Sciences, Boston, MA) at 365 nm excitation and 450 nm emission wavelengths [21]. A range of known dilutions of salmon testes DNA (Sigma) was used to plot a standard curve from which follicle DNA content was extrapolated as previously described [7]. This experiment was repeated twice with the same result.
Histology and Immunohistochemistry
To confirm that cultured follicles encapsulated in calcium alginate retain normal anatomic relationships between the cell types within a follicle, immunohistochemistry for connexin 43 was performed on follicles that were cultured in calcium alginate beads with and without FSH. Some additional sections were also stained with hematoxylin and eosin. At 72 hours of culture, the follicles were released from the beads by allowing the capsules to dissolve in sterile PBS (pH 7.4). Follicles were embedded in OCT for fluorescent immunostaining.
Frozen sections were cut at 5 μm and air-dried overnight at room temperature. Sections from at least 10 follicles per group were evaluated for connexin staining pattern. After a 10 minute acetone fixation, slides were again allowed to air dry, and then were rehydrated in PBS for 10 minutes. Immunohistochemistry was performed using the M.O.M. immunodetection system (Vector Laboratories, Burlingame, Calif.). All reagents were prepared according to kit directions and all steps were carried out at room temperature. Slides were incubated with blocking reagent for 1 hour. Mouse anticonnexin (Chemicon International, Temecula, CA) was prepared in diluent to a final concentration of 0.1 μg/mL and applied to the sections for 1 hour in a humidified chamber. For negative control sections, mouse IgG was substituted for the anti-connexin antibody. After a PBS wash, sections were incubated with biotinylated anti-mouse IgG for 10 minutes. Slides were again washed for 10 minutes in PBS, and then incubated with fluorescein avidin DCS (1:62.5 in PBS) for 5 minutes. After a final 10 minute PBS wash, slides were mounted with Vectashield mounting medium (Vector Laboratories) and digitally imaged using a Leica CMR fluorescence microscope.
Statistical analysis
Statistical analysis for DNA quantification and follicle diameters was performed by analysis of variance (ANOVA) with post hoc testing. Survival data was analyzed with the Kaplan Meier program contained in the Sigma Stat 3.0 software package (Chicago, Il). Probability values less than 0.05 were used to determine significance.
Results
Effect of calcium alginate encapsulation and FSH treatments on follicle diameter
Average follicle diameters at the beginning and end of the 72-hour culture period are depicted in Figure 2A. To determine the effects of alginate alone, follicles were encapsulated in calcium alginate, but grown in the complete absence of FSH as a negative control. A positive control comparison group consisted of follicles grown in the standard fashion and treated with FSH, but without encapsulation [7]. As expected, the negative control group of follicles encapsulated and grown in the absence of FSH, experienced minimal growth (diameter increase; 6.3%). The positive control group of unencapsulated follicles with FSH in the medium experienced a 35.4% increase in diameter, consistent with our previous studies [7]. When encapsulated follicles were grown in medium containing FSH, a 19.9% increase in follicle diameter over the 72-hour culture period was observed (p < 0.05 relative to the negative control). When FSH was included in the calcium alginate capsule but not the medium, a 13.5% increase in follicle diameter was observed. This was different from the negative control (p < 0.05), but not the previously described FSH in medium group (p > 0.05). However, when follicles were cultured with FSH included in both the medium and the bead, a 32.6% increase in follicle size was observed.
Figure 2 (A) Diameter of follicles cultured under different conditions over 72 h. Each line represents measured follicle diameter before and after 72 h of culture (n = 20 – 30 follicles per group). Data points are average diameter with standard error bars included for each treatment. Treatment conditions are listed to the right of the data point representing average diameter at 72 hours for each group. a, b, c represent statistically different groups (p < 0.05). (B) Kaplan Meier Survival Plot. Mean survival time for unencapsulated follicles is 4.5 days while mean survival time for encapsulated follicles is 6 days (p < 0.05).
The size of positive control follicles and encapsulated follicles with FSH placed in the medium and the bead were significantly different than the other three treatment groups (Figure 2A, p < 0.05). Furthermore, there was no significant difference between the growth of unencapsulated FSH-treated cultures and the encapsulated follicles with FSH present in both the bead and the medium (positive control, 54.3 ± 10.1 μm; FSH in bead and media, 49.2 ± 6.8 μm; p > 0.05).
Survival Analysis
The effect of encapsulation on follicle survival was evaluated by Kaplan Meier analysis of daily observation of cultured follicles. Survival plots for unencapsulated and encapsulated FSH-treated follicles are shown in Figure 2B. Mean intact survival time for encapsulated follicles, 6 days, was greater than the mean intact survival time for unencapsulated follicles, 4.5 days (p < 0.05).
DNA Quantification confirms follicle growth
DNA content of individual follicles was consistent with measured follicle size. The unencapsulated follicles in FSH supplemented media (Positive Control) contained 50.7 ± 19.9 ng of DNA after 72 hours of culture. The encapsulated follicles with FSH present in the media and the bead (Experimental Group) contained 53.9 ± 15.9 ng of DNA, not significantly different from the positive control. Both FSH-treated groups contained significantly greater amounts of DNA than the encapsulated follicles cultured in the absence of FSH (Negative Control) which had only 27.8 ± 11.7 ng of DNA (p < 0.05).
Connexin 43 expression in cultured follicles
As seen in Figure 3A, connexin expression (green stain) is very low in follicles in the absence of FSH. In the presence of FSH, connexin expression is readily apparent in the unencapsulated follicle (Figure 3C). However when the follicle is encapsulated and treated with optimal FSH (both in the alginate and in the medium), excellent expression of connexin 43 is seen throughout the cross section of the follicle (Figure 3E). This pattern is very similar to the follicle grown in vivo (Figure 3G). The negative control pictured (Figure 3H) is a serial section from the same follicle from 3E and demonstrates that there is minimal background staining in this system. Follicle sections stained with H&E are adjacent to the corresponding connexin photograph. In the absence of FSH (Figure 3B), the follicle remains intact but the granulosa cell layer is poorly organized and contains numerous pyknotic appearing cells. Figure 3D is of an FSH treated follicle that was not encapsulated. There is a thickened appearance of the outer granulosa cells though the cells around the oocyte appear fairly normal. This is consistent with the area that stains more heavily for connexin. Figure 3F depicts a typical follicle grown in the bead with optimal FSH. There is good morphology with a fairly uniform granulosa layer and a healthy appearing oocyte.
Figure 3 Connexin 43 expression and histology of representative cultured follicles. (A and B) Encapsulated follicle in the absence of FSH, (C and D) Unencapsulated follicle grown in FSH supplemented media, (E and F) Encapsulated follicle with FSH present in both the media and the bead, (G) Connexin staining in preantral follicle of an intact ovary section. (H) Negative control for the immunohistochemical staining, serial section from the same follicle as in C. A, C, E, G, H, represent connexin immunohistochemistry. B, D, and F are H&E stained follicle sections. In panel C, white line is location of basement membrane. Scale bars represents 50 μm in length.
Discussion
In this study, we have demonstrated that calcium alginate encapsulation can act as a scaffold to support three-dimensional relationships between the cells of a follicle in suspension culture. However we have also demonstrated that calcium alginate encapsulation can impede access of the follicle to FSH in the medium.
We have previously shown that FSH is a growth factor for preantral follicles in culture [19]. Other investigators have shown that the dosing and timing of follicle exposure to FSH has threshold limits for continued growth and survival of mouse follicles [21]. Our previous work has demonstrated that encapsulation slowed the rate of follicle growth [7]. There were two hypotheses considered to explain the decreased growth response to FSH. Either FSH diffusion into the alginate bead was significantly limited, or the bead itself physically hindered the growth of the follicle.
With FSH present only in the bead, a concentration gradient is created between the bead and the media. Naturally this gradient would drive the flux of FSH out of the calcium alginate capsule. For this case, because the molecular weight is less than 3 × 105, the FSH will freely diffuse out of the bead until equilibrium is reached [17]. For a substrate with the molecular weight of FSH, equilibrium would be reached within 3 to 5 hours [11], resulting in a much lower concentration of FSH within the bead for the bulk of the 72 hour culture. As a result, follicular growth response was greatly diminished compared to the positive control (Figure 2).
Similarly, by adding FSH to only the media, a concentration gradient is established in the opposite direction, driving the flux of FSH from the media into the bead. However, due to the molecular weight of the hormone being greater than 2 × 104, the diffusion into the capsule will be impeded [11]. The question is then, will the decreased diffusion rate into the calcium alginate bead be significant enough to cause the observed reduction in follicular growth (Figure 2).
To answer this, equal concentrations of FSH were placed in both the bead and the culture media. By eliminating the concentration gradients, FSH was not driven from the bead, leaving the concentration of FSH within the bead comparable to that present for the positive control. Therefore, because the addition of FSH to the alginate bead restored the follicular growth rate to that of the positive control, the likely cause of the diminution of follicle growth in our earlier studies [7] is a limitation of follicle access to FSH by the encapsulation process.
In order to confirm appropriate cell-cell interactions within the follicle, we next evaluated the expression of the gap junction protein connexin 43 in cultured follicles. Connexin 43 plays a key role in follicle development by promoting communication between granulosa cells via the exchange of small ions and molecules [22,23]. Connexin 43 deficient mice have severely impaired follicle development beyond the primary stage [24]. The expression and localization of connexin 43 is also highly FSH dependant [22]. The evaluation of connexin in our cultured follicles provides both a structural and a hormonal assessment of the effect of alginate on follicle development and FSH-stimulated differentiation.
In this study, it was found that the patterned expression of connexin 43 is not affected by calcium alginate encapsulation when the FSH is included in both the bead and the medium (Figure 3). Encapsulated preantral follicles cultured with FSH, maintain a well developed theca, a basement membrane, a normally configured granulosa compartment and a healthy appearing egg with a zona pellucida (Figure 3E and 3F). Interestingly, it was observed that the positive control lacked the punctuate staining at its periphery (Figure 3C), while the encapsulated follicle demonstrated consistent punctuate staining throughout the entire granulosa layer (Figure 3E) similar to in vivo preantral follicles (Figure 3G). This may suggest that encapsulating the follicle might protect it from possible deleterious shearing forces in the suspension culture system, since shear stress has been shown to affect connexin expression [25]. Further studies will be needed to more fully evaluate this phenomenon.
Although a variety of gels and scaffolds exist, calcium alginate has some desired advantages. Unlike collagen based gels [26,27], the follicle can be easily viewed through the calcium alginate capsule, allowing for daily observations to be made. Alginate is easily washed away with PBS, unlike collagen. The alginate also does not seem to interact with the follicles unless ECM or its components are included in the gel [28]. Matrigel has also been used as an imbedding medium for follicles. However matrigel is like serum in that it is derived from undefined biological sources and can have considerable content variation from lot to lot. In contrast, calcium alginate is a well-defined gel and fits with the goal of establishing a completely defined culture system that can be stably reproduced and still meet all the nutritional and structural needs of the developing follicle.
Conclusion
These combined findings have demonstrated that microencapsulation can improve the support of three dimensional growth of preantral follicles; but requires the inclusion of FSH in the scaffold. Further work is necessary to determine if intact follicles can be maintained for longer periods in culture with this scaffolding approach. However, care must be taken to determine if the scaffolding material itself alters the microfollicular environment. Though follicle development is a prolonged and complex process, careful application of tissue engineering principles may facilitate the eventual development of a consistent, standardized in vitro process for follicle growth of large mammals that will be reliable enough for gamete production.
List of Abbreviations
FSH: Follicle-Stimulating Hormone
DNA: Deoxyribonucleic Acid
MW: Molecular Weight
w/v: Weight per volume
IU: International Unit
ml: Milliliter
μm: Micrometer
mg: Milligram
L: Liter
Rpm: revolutions per minute
OCT: Optimal Cutting Temperature Compound
PBS: Phosphate Buffered Solution
Avidin DCS: Cell Sorting Grade Avidin D
ECM: extracellular matrix
Authors' contributions
All authors participated in group discussions of the project conception and design and the analysis of data. MH and EAM drafted the manuscript, RK and AR edited and approved the final manuscript. All authors read and approved the final manuscript.
Acknowledgements
We wish to acknowledge Jill Brekosky for excellent technical assistance with immunohistochemistry and histology and Astrid de Ridder for editorial assistance.
Support: This work has been supported by a Pittsburgh Tissue Engineering Initiative Award from PTEI, Pittsburgh, PA and Women's Reproductive Health Research Career Development Award from NICHD K12 HD38513-01 to EAM.
==== Refs
Hovatta O Cryopreservation and culture of human ovarian cortical tissue containing early follicles Eur J Obstet Gynecol Reprod Biol 2004 S50 54 15041132 10.1016/j.ejogrb.2003.11.012
Spears N Boland NI Murray AA Gosden RG Mouse Oocytes Derived From In Vitro Grown Primary Ovarian Follicles are Fertile Hum Reprod 1994 9 527 532 8006146
Murray AA Molinek MD Baker SJ Kojima FN Smith MF Hillier SG Spears N Role of Ascorbic Acid in Promoting Follicle Integrity and Survival in Intact Mouse Ovarian Follicles In Vitro Reproduction 2001 121 89 96 11226031 10.1530/rep.0.1210089
Rose UM Hanssen RG Kloosterboer HJ Development and characterization of an in vitro ovulation model using mouse ovarian follicles Biol Reprod 1999 61 503 511 10411533
Eisenhauer KM Chun SY Billig H Hsueh AJ Growth hormone suppression of apoptosis in preovulatory rat follicles and partial neutralization by insulin-like growth factor binding protein Biol Reprod 1995 53 13 20 7545438
Wycherley G Downey D Kane MT Hynes AC A novel follicle culture system markedly increases follicle volume, cell number and oestradiol secretion Reproduction 2004 127 669 677 15175503 10.1530/rep.1.00040
Rowghani NM Heise MK McKeel D McGee EA Koepsel RR Russell AJ Maintenance of morphology and growth of ovarian follicles in suspension culture Tissue Eng 2004 10 545 552 15165471 10.1089/107632704323061906
de Groot M Schuurs TA van Schilfgaarde R Causes of limited survival of microencapsulated pancreatic islet grafts J Surg Res 2004 121 141 150 15313388 10.1016/j.jss.2004.02.018
Glaser C Marti U Burgi-Saville ME Ruchti C Gebauer M Buchler MW Gerber H Burgi U Peter HJ Inhibition of iodine organification and regulation of follicular size in rat thyroid tissue in vitro Endocrine 1999 11 165 170 10709764 10.1385/ENDO:11:2:165
Dulieu C Poncelet D Neufeld R Encapsulation and Immobilization Techniques Cell Encapsulation Technology and Therapeutics 1999 3 17
Smidsrod O Skjak-Braek G Alginate as Immobilization Matrix for Cells Trends Biotechnol 1990 8 71 78 1366500 10.1016/0167-7799(90)90139-O
Amsden B Turner N Diffusion Characteristics of Calcium Alginate Gels Biotechnol Bioeng 1999 65 605 610 10516587 10.1002/(SICI)1097-0290(19991205)65:5<605::AID-BIT14>3.0.CO;2-C
Martinsen A Skjak-Braek G Smidsrod O Alginate as Immobilization Material: I. Correlation Between Chemical and Physical Properties of Alginate Gel Beads Biotechnol Bioeng 1989 33 79 83 10.1002/bit.260330111
Martinsen A Alginate as Immobilization Material: III. Diffusional Properties Biotechnol Bioeng 1992 39 186 195 10.1002/bit.260390210
Zimmermann U Biocompatible Encapsulation Materials: Fundamentals and Application 1999 Birkhäuser, Boston, Basel, Berli: Cell Encapsulation Technology and Therapeutics 40 52
Yang H Wright JR Jr Calcium Alginate 1999 Birkhäuser, Boston, Basel, Berli: Cell Encapsulation Technology and Therapeutics 79 89
Tanaka H Matsumura M Veliky IA Diffusion Characteristics of Substrates in Ca-Alginate Gel Beads Biotechnol Bioeng 1984 26 53 58 10.1002/bit.260260111
Palagiano A Nesti E Pace L FSH: urinary and recombinant Eur J Obstet Gynecol Reprod Biol 2004 S30 33 15196713 10.1016/j.ejogrb.2004.01.023
McGee EA Spears N Minami S Hsu S Chun S Billig H Hsueh AJW Preantral Ovarian Follicles in Serum-Free Culture: Suppression of Apoptosis after Activation of the Cyclic Guanosine 3',5'-Monophosphate Pathway and Stimulation of Growth and Differentiation by Follicle Stimulating Hormone Endocrinology 1997 138 2417 2424 9165031 10.1210/en.138.6.2417
Spears N Murray AA Allison V Boland NI Gosden RG Role of Gonadotrophins and Ovarian Steroids in the Development of Mouse Follicles In Vitro J Reprod Fert 1998 113 19 26
Cortvrindt R Smitz J Van Steirteghem AC In Vitro Maturation, Fertilization and Embryo Development of Immature Oocytes From Early Preantral Follicles From Prepubertal Mice in a Simplified Culture System Hum Reprod 1996 11 2656 2666 9021369
Johnson ML Redmer DA Reynolds LP Bilski JJ Grazul-Bilska AT Gap junctional intercellular communication of bovine granulosa and thecal cells from antral follicles: effects of luteinizing hormone and follicle-stimulating hormone Endocrine 2002 18 261 270 12450318 10.1385/ENDO:18:3:261
Sommersberg B Bulling A Salzer U Frohlich U Garfield RE Amsterdam A Mayerhofer A Gap junction communication and connexin 43 gene expression in a rat granulosa cell line: regulation by follicle-stimulating hormone Biol Reprod 2000 63 1661 1668 11090433
Juneja SC Barr KJ Enders GC Kidder GM Defects in the germ line and gonads of mice lacking connexin 43 Biol Reprod 1999 60 1263 1270 10208994
DePaola N Davies PF Pritchard WF JrFlorez L Harbeck N Polacek DC Spatial and temporal regulation of gap junction connexin 43 in vascular endothelial cells exposed to controlled disturbed flows in vitro Proc Natl Acad Sci U S A 1999 16 3154 159 10077653 10.1073/pnas.96.6.3154
Torrance C Telfer E Gosden RG Quantitative study of the development of isolated mouse pre-antral follicles in collagen gel culture J Reprod Fertil 1989 87 367 374 2621708
Loret de Mola JR Barnhart K Kopf GS Heyner S Garside W Coutifaris CB Comparison of two culture systems for the in-vitro growth and maturation of mouse preantral follicles Clin Exp Obstet Gynecol 2004 31 15 19 14998179
Kreeger PK Woodruff TK Shea LD Murine granulosa cell morphology and function are regulated by a synthetic Arg-Gly-Asp matrix Mol Cell Endocrinol 205 1 10 2003 Jul 31 12890562
|
16162282
|
PMC1262772
|
CC BY
|
2021-01-04 16:37:12
|
no
|
Reprod Biol Endocrinol. 2005 Sep 14; 3:47
|
utf-8
|
Reprod Biol Endocrinol
| 2,005 |
10.1186/1477-7827-3-47
|
oa_comm
|
==== Front
Reprod Biol EndocrinolReproductive biology and endocrinology : RB&E1477-7827BioMed Central London 1477-7827-3-511618804210.1186/1477-7827-3-51ReviewEstrogen regulation of testicular function Akingbemi Benson T [email protected] Department of Anatomy, Physiology and Pharmacology, College of Veterinary Medicine, Auburn University, Auburn, AL 36849, USA2005 27 9 2005 3 51 51 7 6 2005 27 9 2005 Copyright © 2005 Akingbemi; licensee BioMed Central Ltd.2005Akingbemi; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Evidence supporting a role for estrogen in male reproductive tract development and function has been collected from rodents and humans. These studies fall into three categories: i) localization of aromatase and the target protein for estrogen (ER-alpha and ER-beta) in tissues of the reproductive tract; ii) analysis of testicular phenotypes in transgenic mice deficient in aromatase, ER-alpha and/or ER-beta gene; and, iii) investigation of the effects of environmental chemicals on male reproduction. Estrogen is thought to have a regulatory role in the testis because estrogen biosynthesis occurs in testicular cells and the absence of ERs caused adverse effects on spermatogenesis and steroidogenesis. Moreover, several chemicals that are present in the environment, designated xenoestrogens because they have the ability to bind and activate ERs, are known to affect testicular gene expression. However, studies of estrogen action are confounded by a number of factors, including the inability to dissociate estrogen-induced activity in the hypothalamus and pituitary from action occurring directly in the testis and expression of more than one ER subtype in estrogen-sensitive tissues. Use of tissue-specific knockout animals and administration of antiestrogens and/or aromatase inhibitors in vivo may generate additional data to advance our understanding of estrogen and estrogen receptor biology in the developing and mature testis.
==== Body
Introduction
The testis consists of two compartments: seminiferous tubules and intertubular tissue, which forms the interstitium. Seminiferous tubules are lined by layers of germ cells in various stages of development (spermatogonia, spermatocytes, spermatids, spermatozoa) and supporting Sertoli cells. The interstitium consists of loose connective tissue, blood and lymphatic vessels, and various cell types, including Leydig cells, fibroblasts, macrophages and leukocytes. Leydig cells are the predominant source of the male sex steroid hormone testosterone. However, recent observations challenge the dogma that the male phenotype is maintained solely by testosterone binding to its protein target, i.e., the androgen receptor. Growing public concerns that exposures to environmental chemicals with estrogenic activity may impact human reproductive health have focused attention on the role of estrogen in male reproductive health [1]. The aromatization of C19 androgens, i.e., testosterone and androstenedione, is a key step in estrogen (E2) biosynthesis and is catalyzed by the aromatase enzyme, which is a product of the CYP19 gene [2]. The serum levels of E2 measure about 40 pg/mL in male rats [3], and ranges between 20 and 40 pg/mL in men [4]. Evidence from several studies indicates that aromatase, ERα and ERβ are encoded by separate genes but are co-expressed with androgen receptors in the male reproductive tract [2,3]. In consonance with localization studies, mice which have targeted deletion of the aromatase gene, ERα and/or ERβ showed altered testicular morphology and derangements of spermatogenesis [5-7], and exposures of laboratory species and wildlife to estrogenic chemicals were found to cause abnormalities of the reproductive tract [8].
Although the present review is focused on direct estrogen action in the testis, estrogen regulation may occur indirectly by changes caused in the hypothalamus and pituitary. Gonadal steroids act on the hypothalamus to affect GnRH pulses, and at the pituitary level to regulate gonadotropin (FSH and LH) secretion. FSH and LH are the primary tropic hormones that regulate testicular function. Indeed, FSH receptors are expressed only in Sertoli cells, and Leydig cells are the only binding sites for LH in the testis. In contrast, ERs have a more diversified pattern of expression. There is conclusive evidence showing that ERα and ERβ are present in several hypothalamic nuclei and in pituitary gonadotropes, indicating that estrogen regulates the hypothalamus-pituitary axis [9,10]. For example, E2 treatment of a mouse gonadotroph cell line (LβT2) increased LH secretion and, following co-incubation with GnRH, increased LHβ mRNA levels [11]. Furthermore, the presence of estrogen-response-elements (EREs) on the promoter region of the β-subunit of the LH gene has been reported, implying that estrogen regulation of LH secretion occurs directly at the level of the LHβ gene [12,13]. There is also evidence that ERα can be transcriptionally activated in gonadotrope cells in an estrogen-independent manner, through the GnRH receptor and signaling via protein kinase C (PKC) and mitogen-activated protein kinase (MAPK) pathways [14]. Together, these observations demonstrate that estrogen regulation of testicular function is also mediated indirectly by changes occurring in the hypothalamus and pituitary.
Estrogen Receptors
ERs are members of the steroid/thyroid hormone super family of nuclear receptors, which share a common structural architecture, and consist of three independent but interacting functional domains: the N-terminal or A/B domains, the C or DNA-binding domain, and the D/E/F or ligand-binding domain (Fig 1). Binding of a ligand to the ER causes a series of downstream events, including receptor dimerization, receptor-DNA interactions mediated by EREs present in the promoter region of target genes, recruitment of and interaction with transcription factors, and the formation of a preinitiation complex. Ligand-receptor interactions ultimately cause changes in target gene expression [15]. The N-terminal domain of nuclear receptors encodes an activation function called AF-1, which mediates protein-protein interactions to induce transcriptional activity. It is thought that this domain is highly active in ERα-mediated stimulation of reporter gene expression from a variety of ERE-constructs but its activity in the ERβ is limited [16]. On the other hand, the C-terminal or ligand-binding domain contains the AF-2 interacting surface that mediates ligand binding and receptor dimerization to stimulate transcriptional activity [17]. Thus, AF-1 and AF-2 are both involved in mediating the transcriptional activation functions of ERs.
Figure 1 An illustration of the structure of the estrogen receptor. The NH2 terminal consists of the A/B domains, the C domain forms the DNA-binding domain (DBD) while domains D/E/F constitute the ligand-binding domain (LBD). The AF-1 and AF-2 activation units are part of the DNA-binding and ligand-binding domains, respectively. The two ER subtypes, ERα and ERβ, are almost identical in the DNA-binding domain (~95% homology) but differ in the ligand-binding domain (about 60% homology). Differences in the ligand-binding domain are responsible in part for ligand-specificity, and the ratio of ERα and ERβ is a critical determinant of cellular response to endogenous estrogen and other ER agonists and antagonists.
Although there is a high degree of homology in the DNA-binding domains of ERα and ERβ (about 95%), only a partial homology exists in the ligand-binding domain (~60%) [18]. Differences in ligand binding, in association with other factors, have the effect of altering the pattern of ER-mediated transcriptional activity. For example, some agonists bind both ER subtypes with the same affinity while others preferentially bind to ERα or ERβ [19-21]. There is general agreement that ERs function as dimers, and co-expression of ERα and ERβ in the same cell causes the formation of homodimers (ERα/ERα and ERβ/ERβ) or heterodimers (ERα/ERβ), which affect ligand-specificity. The interactions between ERs and EREs are complicated by other factors, including the ability of ERβ to modulate ERα transcriptional activity and recruitment of several protein co-activators and repressors by both ER subtypes. Therefore, the relative amounts of ERα and ERβ in a given tissue are key determinants of cellular responses to estrogen and other ER agonists and antagonists [22]. Moreover, ER and other steroid receptors have the ability to mediate biological effects through non-transcriptional mechanisms mediated by protein-protein interactions occurring between ERs and growth factors e.g., IGF-1 and EGF [23,24]. Furthermore, there is growing evidence for the presence of a small pool of ERs localized to the plasma membrane. For example, BSA-conjugated E2, which is unable to gain entry into the cytosol and acts at the plasma membrane, decreased testicular androgen production in vitro [25]. Membrane ER is thought to signal mainly by coupling to GTP-activating proteins and through pathways involving second messengers (e.g., calcium) and kinase cascades [26]. The integration of several pathways implies that estrogen action in any particular tissue and organ is the result of activities mediated by genomic and non-genomic pathways although the physiological significance of specific pathways in the testis remains to be elucidated [27].
I. Localization of aromatase and ERs
Data describing aromatase activity and ER expression in reproductive tissues were collected using a combination of techniques: binding assays, immunohistochemistry, in situ hybridization, reverse transcriptase-polymerase chain reaction (RT-PCR), and RNase protection assays. In spite of the large body of information derived from these studies, localization studies have shortcomings that limit data interpretation regarding ER expression in specific tissues. For example, binding assays do not distinguish between ERα and ERβ while in situ hybridization studies measure mRNA levels but do not determine whether mRNA is translated to protein. Similarly, immunocytochemistry lacked specificity for either ER subtype. However, the availability of antibodies directed against ERα, and much later for ERβ, allowing for discrimination between ER subtypes in subsequent studies has generated substantial information on ER biology.
During fetal development in the rodent, aromatase is expressed in Sertoli cells and Leydig cells but not in spermatogonia. On the other hand, aromatase has been localized to virtually all cell types in the adult testis, including Leydig cells, Sertoli cells, spermatocytes, spermatids and spermatozoa [6,28,29]. Cellular expression of aromatase is age-dependent in the postnatal rat, occurring predominantly in Sertoli cells and germ cells of the prepubertal testis (up to 21 days of age) and in Leydig cells after this period [30]. Taken together, the bulk of data collected from the rodent testis point to a general pattern of ERα expression in Leydig cells and peritubular myoid cells and ERβ in germ cells [10,31-33]. However, ERβ was localized to adult Leydig cells in the mouse [34], and both ERα and ERβ were found to be present in rat fetal and adult Leydig cells [35]. Similarly, ERα and ERβ were immunolocalized to Leydig cells in pubertal rats although treatment with the pure antiestrogen ICI 182,780 abolished ERα, but not ERβ, protein [36]. Consistent with ER expression in diverse cell types in the testis, it is not surprising that administration of ER agonists and antagonists or targeted deletion of the aromatase gene and ERs caused derangements in germ cell development and testicular steroidogenesis.
Unlike in rodents, aromatase activity and estrogen biosynthesis occur mostly in adipose tissue in men, and the testis synthesizes only 10–25% of E2 in circulation [37]. Early studies showed that prenatal exposures to the synthetic estrogen diethylstilbestrol (DES) caused male reproductive tract abnormalities in mice [38] and men [39]. In agreement with these observations, ERα and ERβ were localized to the human testis, and the presence of two variants of ERβ, designated ERβ1 and ERβ2, has been clearly demonstrated [40]. Although ERβ mRNA levels were 3-fold greater than ERα, both were expressed in the testis beginning from 16 weeks of gestation [41]. ERβ1 was more widely expressed in Sertoli cells, germ cells and Leydig cells while ERβ2 mRNA and protein were restricted to spermatogonia [42]. In the adult testis, both ERα and ERβ are expressed in spermatocytes, elongating spermatids, Sertoli cells and Leydig cells [43,44]. Other studies have demonstrated the presence of ERα in spermatids and mature spermatozoa [45], ERβ in all germ cells [46,47], and the absence of ERα in Leydig cells [48]. ERβ1 appears to be expressed at high levels in pachytene spermatocytes and round spermatids but much less so in Sertoli cells and spermatogonia whereas expression of ERβ2 is high in Sertoli cells and spermatogonia and is reduced in spermatocytes [49,50]. Although the physiological significance of ERβ isoforms in the human testis remains to be clarified, it has been suggested that ERβ2 forms heterodimers with ERα thereby attenuating its transcriptional activity; however, it lacks the ability to bind endogenous E2 or recruit cofactors via the AF-2 domain [51].
II. Transgenic mouse studies
Reports of testicular anomalies in men with naturally occurring mutations in the aromatase gene and in individuals lacking a functional ERα, including undescended testis, decreased sperm production, and altered endocrine profiles, reinforced the view that estrogen action is a requirement for normal testicular function [52-55]. Thus, development of knockout or transgenic mice with disruption of molecules related to reproduction and hormone action, e.g., mice with targeted deletion of the aromatase gene (ARKO), ERα (αERKO), ERβ (βERKO) and both ER subtypes (αβERKO), has contributed immensely to our understanding of reproductive endocrinology [56]. A major difference between these lines of mutant mice is that ARKO mice adequately express ERα and ERβ protein and do not make endogenous E2 whereas ER knockout mice are able to synthesize E2 but lack either ERα and/or ERβ protein. Therefore, a major caveat in these studies is the inadvertent removal of estrogen priming of extragonadal tissues during development. In this regard, there is a possibility that absence of endogenous E2 and/or ER-mediated activity during tissue differentiation in the hypothalamus and pituitary jeopardizes developmental maturation of regulatory pathways in the HPT axis.
The spectrum of testicular anomalies exhibited by transgenic mice deficient in E2 biosynthesis and ER protein is summarized in Table 1. ARKO mice have enlarged sex accessory organ weights presumably as a result of elevated serum testosterone levels and enhanced androgen action, and show disturbances of spermatogenesis, which is associated with increased apoptosis of developing germ cells [6,7]. However, the results of fertility assessment in ARKO mice have been rather inconsistent, sexual function being impaired in one line of mice and not in the other; these differences are thought to be due to the amounts of residual aromatase gene products in mutant mice [7,57]. In contrast to the lack of E2, overexpression of the aromatase gene and enhanced E2 production in mice induced cryptorchidism or undescended testis, spermatogenic arrest, Leydig cell hyperplasia, and decreased serum FSH and testosterone levels. Disruption of spermatogenesis was associated with decreased FSH levels while increased exposures to E2 induce Leydig cell hyperplasia [58]. Progressive degeneration of testicular tissue, dilation of the seminiferous tubules, and sexual behavioral problems are typical findings in αERKO mice [5]. Disruption of spermatogenesis has been attributed to fluid retention, which causes pressure atrophy of the seminiferous epithelium [59].
Table 1 Testicular function in mice deficient in estrogen biosynthesis or estrogen receptor protein
Spermatogenesis Steroidogenesis Fertility References
ARKOa Affected Affected Affected 6,7
αERKOb Affected Affected Affected 5
βERKOc Normal Normal Normal 66
αβERKOd Affected Affected Affected 64
The obvious differences in the phenotypes of ARKO and αERKO mice, as were determined in early studies, implied that ERα is not the sole mediator of estrogen action and that another ER protein may be present in testicular cells. These speculations were confirmed by cloning of ERβ in the rat prostate and ovary [60]. Subsequently, the bulk of experimental evidence shows that ERβ regulates germ cell development. For example, ERα inactivation had no effect on the number of Sertoli cells and spermatogonia whereas ERβ inactivation increased the number of spermatogonia by more than 50% in neonatal mice [61]. However, it is surprising that in spite of the evidence for ERβ regulation of mitosis in spermatogonia, which serve as stem cells for the process of spermatogenesis, disturbances of sperm production were not evident in βERKO mice. On the other hand, the presence of ERα in Sertoli cells has not been demonstrated. Paradoxically, spermatogenic arrest occurs in αERKO mice, which have ERβ protein. These observations suggest that testicular cells regulate Sertoli cell support of germ cell development through unidentified ERα-mediated mechanisms. This line of thinking is supported by data from experiments in which germ cells were transplanted from donor males homozygous for the mutation ERα-/- to testes of wild-type ERα+/+ recipient mice depleted of germ cells. When mated to wild-type females, the recipients sired offspring heterozygous for the mutation ERα+/- but retained the coat-color marker of the ERα-/- donor mice. This finding confirmed that somatic cells in the testis, but not germ cells, have a requirement for ERα in order to support the process of spermatogenesis [62,63]. In contrast to the αERKO, βERKO males retain full fertility but tend to show increased incidence of prostate hyperplasia with advancing age [64]. Perhaps not unexpectedly, male αβERKO mice are infertile, which is likely due to ERα deficiency because these effects are absent in βERKO mice [65,66].
Alteration of the endocrine profile is a consistent finding in transgenic mice with targeted deletion of aromatase gene or ERs (Table 2). For example, serum LH levels were elevated in adult ARKO mice [6] while serum testosterone concentrations, though were increased at 12–14 wk of age, were comparable in wild-type and mutant mice [7]. Similarly, the concentrations of serum testosterone, LH, and FSH were increased in αERKO males compared to their wild-type littermates [5,67]. The changes in serum gonadotropin levels presumably result from alleviation of estrogen feedback regulation on the hypothalamus-pituitary axis. The regulation of testicular steroidogenesis appears to be mediated primarily by ERα because changes in serum steroid hormone levels seen in the αERKO are absent in βERKO mice. Moreover, administration of a synthetic estrogen, estradiol benzoate, reduced serum LH and testosterone levels in wild-type but not αERKO mice (Fig. 2A), and treatment with the pure antiestrogen ICI 182,780 decreased androgen biosynthesis in wild-type but not αERKO Leydig cells [67]. The differences in androgen biosynthesis between αERKO and wild-type Leydig cells were associated with changes in steroidogenic enzyme activity because ERα deficiency enhanced gene expression for cytochrome P450 hydroxylase/17α lyase and 17β-hydroxysteroid dehydrogenase type III; these enzymes take part in reactions involved in the conversion of the steroid substrate cholesterol to testosterone (Fig. 2B). Consistent with these observations, a recent report showed that DES decreased testosterone production of wild-type fetal and neonatal testes but not ERα-/- [68].
Table 2 Endocrine profiles of mice deficient in estrogen biosynthesis or estrogen receptor protein
FSH LH Testosterone 17β-estradiol References
ARKOa Elevated Elevated Elevated Absent 6,7
αERKOb ND Elevated Elevated ND 5,67
βERKOc Normal Normal Normal ND 66
αβERKOd ND Elevated Elevated ND 64
ND, Not determined.
Figure 2 Estrogen regulates testicular steroidogenesis, acting via ERα because administration of an estrogenic chemical, estradiol benzoate, suppressed pituitary LH secretion and serum testosterone levels in wild-type but not αERKO mice (ref. 67)(A). ERα deficiency enhanced gene expression for cytochrome P450 hydroxylase/17α lyase and 17β-hydroxysteroid dehydrogenase type III in αERKO Leydig cells compared to wildtype (WT) (B), indicating that ERα regulates androgen biosynthesis by mediating changes in steroidogenic enzyme activity (ref. 67). Copyright 2003, The Endocrine Society.
III. Studies of xenoestrogens
Although there had been long standing evidence that male reproductive tract development is subject to estrogen action [39,69], scientific attention to the role of estrogen in reproductive activity was highlighted only recently by public concerns that exposures to environmental chemicals may adversely affect the endocrine and reproductive systems. Exposures of laboratory animals and wildlife to high levels of estrogenic chemicals resulted in a number of abnormalities, including reduced gonad size, feminization of genetic males, and low sperm count and quality. In this regard, estrogenic activity has been attributed to a diverse array of steroidal and non steroidal compounds, including industrial chemicals (e.g., polychlorobiphenyls, alkyphenols, pesticides (e.g., DDT derivatives, methoxychlor, kepone), pharmaceutical agents (e.g., DES, tamoxifen, raloxifene), phthalates (e.g., di-2-ethylhexylphthalate, di-n-butyl phthalate), and phytoestrogens (e.g., genistein, daidzen) [1,70,71]. While there is no clear data demonstrating that environmental chemicals are the cause of reproductive anomalies in humans, the homology in organ systems between animal models and humans indicates a potential for adverse effects on sexual development and function.
Although binding affinity of xenoestrogens for ERs is low, ranging from 0.0001% to 1% of E2 levels, these chemicals have the ability to activate ERα and ERβ as agonists or prevent their binding by endogenous ligands when acting as antagonists [72,73]. Just as diverse as the number of chemicals known to exhibit estrogenic activity, the profile of biological responses to exogenous chemicals is affected by a variety of factors in reproductive tissues: animal strain and species differences, relative amounts of ER subtypes, presence of EREs, recruitment of co-regulatory proteins (co-activators and repressors), binding to plasma proteins, chirality of chemicals, and multiple mechanisms of action (e.g., estrogenicity versus antiandrogenicity) [71,74]. Specifically, xenoestrogens evoke estrogenic responses and cause their effects by mimicking and/or blocking the actions of endogenous E2 (agonist versus antagonist), and these effects may be result in changes in steroid hormone receptor gene expression, altered steroid hormone metabolism, cross-talk between ERs and other signaling systems (e.g., aryl hydrocarbon and EGF), and interference with serum protein binding [75-77]. It has also been suggested that the presence of xenoestrogens in the hormonal milieu of estrogen-sensitive tissues has the effect of potentiating E2 action [78]. The nature of the ERE in the promoter region of target genes may affect cellular response as indicated by the ability of ERβ to activate EREs from the vitellogene while ERα showed greater activation at the more divergent LH EREs in COS-1 cells [79]. Because ERs function as dimers, estrogen responsive genes may respond differently to ERα and ERβ homodimers or ERα/ERβ heterodimers following ER activation [80]. Furthermore, there are ligand-dependent differences in the ability of ERα and ERβ to bind co-regulatory proteins [21,81]. Thus, the cellular response to ER agonists and antagonists is the result of interaction between several factors.
A detailed discussion of the effects of environmental chemicals on male reproduction is outside the scope of the present review and can be found elsewhere [82,83]. However, studies of estrogenic chemicals have been conducted in laboratory species with low (physiological) and high (pharmacological) doses. Data from these investigations indicate that estrogen action is dose-dependent and may be stimulatory or inhibitory. For example, exposure to E2 restored spermatogenesis to the germ cell-depleted testis of hypogonadal mice [84], decreased the rate of apoptosis and stimulated proliferation of mouse and rat spermatogonia in vitro [43,85], and induced renewal of spermatogonial stem cells in the testis of the Japanese eel [86]. On the other hand, incubation with E2 and DES was found to inhibit development of spermatogonia, Leydig cells and Sertoli cells in the fetal rat testis [87]. Administration of low doses of the industrial and estrogenic chemical bisphenol A (BPA) reduced spermatogenesis in mice [88], decreased DNA synthesis by immature rat Leydig cells (author's unpublished observations), and suppressed androgen biosynthesis by mature rat Leydig cells (Fig. 3). The effects of E2 and BPA on spermatogonial divisions and Leydig cell steroidogenesis were blocked by co-incubation with antiestrogens ICI 164384 and ICI 182,780, respectively, indicating that these effects were ER-mediated [1,89]. There is also evidence showing that E2 regulates ER gene expression in a dose-dependent manner because chronic exposures of mice to 0.5 or 50 μg/ml BPA decreased ERβ and increased ERα gene expression in germ cells [90] but a single injection of estradiol benzoate at high doses (500 μg) caused the opposite effect in prepubertal rats, i.e., decreased ERα mRNA levels and increased ERβ expression [91]. Disparities in data from different laboratories are probably due to several factors, which act to moderate estrogen signaling in sensitive tissues, e.g., interaction between transcriptional and non-transcriptional signaling pathways, receptor cross-talks, unpredictable mixture effects, and changes in steroid production and action. In addition, an inverted U-shaped dose-response, in which low doses are stimulatory and high doses are inhibitory, has been proposed for estrogen action in reproductive tissues [92].
Figure 3 ER agonists regulate androgen biosynthesis in Leydig cells. Incubation of mature rat Leydig cells, from 90-day old rats, with estrogenic chemicals, i.e., bisphenol A (BPA)(A), the synthetic estrogen diethylstilbestrol DES (B) and a biologically active metabolite of the pesticide methoxychlor (HPTE)(C), caused an inhibitory effect on androgen biosynthesis albeit at different doses. Using RT-PCR, ERβ was not detected in these cells, implying that inhibitory effects were ERα-mediated (ref. 89). Copyright 2004, The Endocrine Society.
In agreement with studies conducted in rodents, evidence supporting a direct role for estrogen in male reproductive tract development was collected from men. For example, poor semen quality has been a consistent finding in male patients with mutations in ERα [52] as well as those suffering from aromatase deficiency [53,54]. A recent study involving a large cohort of men concluded that prenatal DES exposure is associated with testicular cancer and malformations of the genitalia although fertility was not affected [93]. Increased incidence in testicular cancer was thought to be due to early life-stage exposures to environmental estrogens and/or antiandrogens, which interfere with the ability of gonadal steroids to support tissue differentiation in the fetal period [94,95]. Indeed, elevated blood estrogen levels in dizygotic twin pregnancies are known to increase the risk of testicular cancer in males [96]. Growing epidemiological evidence in support of these observations has led to the hypothesis, which states that a testicular dysgenesis syndrome (TDS) that is characterized by hypospadias, testicular cancer, abnormal spermatogenesis and undescended testis, is the result of interaction between genetic and environmental factors, including inappropriate exposures to endocrine-active chemicals [97,98]. The growing incidence of TDS in the population implies that changes in steroid hormone synthesis and action cause greater effects during sexual differentiation in humans, as in rodents, but it is not clear that sperm function and fertility are affected in adulthood.
A series of data were published lately to highlight aspects of estrogen action in the testis. First, it was observed that neonatal treatment of prepubertal rats with DES alone (0.1 μg) induced only minor effects, which were amplified after suppression of androgen production and action. Thus, it was hypothesized that reduced androgen levels render the reproductive tract more sensitive to estrogen stimulation, and that the ratio between androgen and estrogen, rather than their absolute levels, is the critical determinant of E2 action [99,100]. Curiously, this line of thinking does not explain similarities in the phenotypes of ARKO and αERKO mice, which have comparable serum androgen levels but exhibit different E2 levels (Table 2). However, there are suggestions that phenotypic similarities in ARKO and αERKO mice are possibly due to the confounding effects of growth factors that activate signaling pathways mediating E2 activity, e.g., EGF [59]. Also of interest are recent data showing that the presence of soy in the diet decreases body and testis weights, suppresses gonadotropin secretion, and retards germ cell development in the rat [101]. These findings have implications for analysis of estrogen action in the testis because: 1) The normal rat chow contains significant levels of phytoestrogens (200–300 mg/kg), potentially interfering with the action of E2 and estrogenic chemicals in reproductive tissues [102,103]; and, 2) Putative health benefits associated with soy-based diets may be confounded by phytoestrogen signaling in the testis [104]. Although there is no evidence that consumption of soy-based diets has deleterious effects on testicular function, the possibility that such effects may occur in the prepubertal period, i.e., in infanthood, cannot be discounted as this population is not routinely examined for reproductive health. Because acquisition of adult sexual behavior is dependent on priming of sexually dimorphic hypothalamic nuclei by steroid hormones in the perinatal period [105], such evaluations seem to be warranted.
IV. Conclusion
The major source of E2 biosynthesis is the testis in rodents and adipose tissue in men but the receptor protein (ERα and ERβ) is localized to most cell types in the testis of both species in consonance with a physiological role for estrogen in testicular development and function (Fig. 4). It is therefore not surprising that targeted deletion of the aromatase gene, ERα, and/or ERβ caused a variety of testicular anomalies in mutant mice. For example, evidence for direct ER-mediated action in testicular cells is provided by disruption of spermatogenesis in ARKO and αERKO mice and the requirement for ERα for estrogen action in Leydig cells [6,7]. Moreover, elevated serum LH levels in αERKO mice indicate that ERα deficiency jeopardizes steroid hormone negative feedback mechanisms in the hypothalamus-pituitary axis [5,106]. These observations have clinical relevance because men with disorders of glucose metabolism and those with increased body mass index (i.e., overweight or obese) exhibit elevated serum E2 levels [107] with the potential for enhanced estrogen action in the testis. However, a number of confounding variables need resolution in order to clearly identify the mechanisms associated with the physiological actions of E2 in the testis. In this regard, new experimental approaches are needed, and may include: i) use of tissue-specific knockouts in order to remove effects of concurrent ER-mediated activity in the hypothalamus and/or pituitary [108]; ii) analysis of signaling pathways not mediated by ligand binding of ERα and/or ERβ [109,110]; iii) investigation of the role of membrane ER signaling in the regulation of testicular function [111,112]; iv) assessment of the influence of genetic background on estrogen action [96,113]; v) development of methods for measuring bioavailability of estrogens in the body in order to define dose-effect relationships [114,115]; and vi) use of techniques that maintain the normal hormonal milieu of reproductive tract tissues during investigation, i.e., as related to gonadotropin and androgen action [116]. A combination of these approaches will advance our understanding of the regulatory role of E2 in the mammalian testes.
Figure 4 Endocrine regulation of the testis. Pituitary gonadotropins are the chief regulators of testicular function; FSH acts through its receptors in Sertoli cells (FSHR) to regulate spermatogenesis and LH stimulates androgen production by Leydig cells after binding to LHR. However, gonadal steroids, i.e., androgen and estrogen, and other agents that bind or prevent binding to steroid hormone receptors (androgen receptor AR, ERα, and ERβ), which are present in Sertoli cells, germ cells and Leydig cells also regulate testicular function. The pathway mediated by adenosine-3',5'-cyclic monophosphate (cAMP) appears to be the primary intracellular signaling pathway in all testicular cells. However, several growth factors e.g., insulin like growth factor-1 (IGF-1) and epidermal growth factor (EGF), acting via their receptors, IGF-1R and EGF-R, possibly modulate AR and ER-mediated pathways. Thus, testicular function is regulated by interactions between several signaling pathways, some acting locally, e.g., AR and ER-mediated pathways, and others indirectly by modulating hypothalamus-pituitary function. Hormonal activation of transcriptional gene activity results in changes in cell differentiation and function. PMC, peritubular myoid cell; CRE, cAMP-responsive elements, ARE, androgen-responsive elements; ERE, estrogen-responsive elements.
Abbreviations
ER, estrogen receptor; E2, 17β-estradiol; DES, diethylstilbestrol; HPT, hypothalamus-pituitary-testicular axis; GnRH, gonadotropin releasing hormone; FSH, follicle stimulating hormone; LH, luteinizing hormone; ERE, estrogen response elements; IGF, insulin growth factor; EGF, epidermal growth factor; ARKO, aromatase knockout mice; αERKO, ERα knockout mice; βERKO, ERβ knockout mice, αβERKO mice, mice deficient in both ERα and ERβ; BPA, bisphenol A.
Acknowledgements
Some of the work described in this review was performed during the author's postdoctoral training with Dr. Matthew P. Hardy at the Center for Biomedical Research, Population Council, The Rockefeller University, New York, with funding support by the Fogarty International Center, National Institutes of Health (F05 TW05350) and the National Institute of Environmental Health Sciences (ES 10233). Drs. Y. Tao and T. Braden provided helpful comments on an early draft of this paper.
==== Refs
Colborn T vom Saal FS Soto AM Developmental effects of endocrine-disrupting chemicals in wildlife and humans Environ Health Perspect 1993 101 378 384 8080506
Simpson ER Mahendroo MS Means GD Kilgore MW Corbin CJ Mendelson CR Tissue-specific promoters regulate aromatase cytochrome P450 expression Clin Chem 1993 39 317 324 8432022
Brewster ME Anderson WR Pop E Effect of sustained estradiol release in the intact male rat: correlation of estradiol serum levels with actions on body weight, serum testosterone, and peripheral androgen-dependent tissues Physiol Behav 1997 61 225 229 9035252 10.1016/S0031-9384(96)00364-2
Nagata C Takatsuka N Shimizu H Hayashi H Akamatsu T Murase K Effect of soymilk consumption on serum estrogen and androgen concentrations in Japanese men Cancer Epidemiol Biomarkers Prev 2001 10 179 184 11303585
Eddy EM Washburn TF Bunch DO Goulding EH Gladen BC Lubahn DB Korach KS Targeted disruption of the estrogen receptor gene in male mice causes alteration of spermatogenesis and infertility Endocrinology 1996 137 4796 4805 8895349 10.1210/en.137.11.4796
Fisher CR Graves KH Parlow AF Simpson ER Characterization of mice deficient in aromatase (ArKO) because of targeted disruption of the cyp19 gene Proc Natl Acad Sci U S A 1998 95 6965 6970 9618522 10.1073/pnas.95.12.6965
Robertson KM O'Donnell L Jones ME Meachem SJ Boon WC Fisher CR Graves KH McLachlan RI Simpson ER Impairment of spermatogenesis in mice lacking a functional aromatase (cyp 19) gene Proc Natl Acad Sci U S A 1999 96 7986 7991 10393934 10.1073/pnas.96.14.7986
Sharpe RM Skakkebaek NE Are oestrogens involved in falling sperm counts and disorders of the male reproductive tract? Lancet 1993 341 1392 1395 8098802 10.1016/0140-6736(93)90953-E
Shughrue PJ Lane MV Merchenthaler I Comparative distribution of estrogen receptor-alpha and -beta mRNA in the rat central nervous system J Comp Neurol 1997 388 507 525 9388012 10.1002/(SICI)1096-9861(19971201)388:4<507::AID-CNE1>3.0.CO;2-6
Shughrue PJ Lane MV Scrimo PJ Merchenthaler I Comparative distribution of estrogen receptor-alpha (ER-alpha) and beta (ER-beta) mRNA in the rat pituitary, gonad, and reproductive tract Steroids 1998 63 498 504 9800279 10.1016/S0039-128X(98)00054-3
Nicol L McNeilly JR Stridsberg M Crawford JL McNeilly AS Influence of steroids and GnRH on biosynthesis and secretion of secretogranin II and chromogranin A in relation to LH release in Lbeta2 gonadotroph cells J Endocrinol 2002 174 473 483 12208668 10.1677/joe.0.1740473
Shupnik MA Gharib SD Chin WW Estrogen suppresses rat gonadotropin gene transcription in vivo Endocrinology 1988 122 1842 1846 2452072
Shupnik MA Rosenzweig BA Identification of an estrogen-responsive element in the rat LH beta gene. DNA-estrogen receptor interactions and functional analysis J Biol Chem 1991 266 17084 17091 1894604
Demay F De Monti M Tiffoche C Vaillant C Thieulant ML Steroid-independent activation of ER by GnRH in gonadotrope pituitary cells Endocrinology 2001 142 3340 3347 11459776 10.1210/en.142.8.3340
Kraus WL McInerney EM Katzenellenbogen BS Ligand-dependent, transcriptionally productive association of the amino- and carboxyl-terminal regions of a steroid hormone nuclear receptor Proc Natl Acad Sci U S A 1995 92 12314 12318 8618892
McInerney EM Tsai MJ O'Malley BW Katzenellenbogen BS Analysis of estrogen receptor transcriptional enhancement by a nuclear hormone receptor coactivator Proc Natl Acad Sci U S A 1996 93 10069 10073 8816752 10.1073/pnas.93.19.10069
Giguere V Orphan nuclear receptors: from gene to function Endocr Rev 1999 20 689 725 10529899 10.1210/er.20.5.689
Nilsson S Makela S Treuter E Tujague M Thomsen J Andersson G Enmark E Pettersson K Warner M Gustafsson JA Mechanisms of estrogen action Physiol Rev 2001 81 1535 65 11581496
Chang WY Prins GS Estrogen receptor-beta: implications for the prostate gland Prostate 1999 40 115 124 10386472 10.1002/(SICI)1097-0045(19990701)40:2<115::AID-PROS7>3.0.CO;2-3
Kuiper GG Carlsson B Grandien K Enmark E Haggblad J Nilsson S Gustafsson JA Comparison of the ligand binding specificity and transcript tissue distribution of estrogen receptors alpha and beta Endocrinology 1997 138 863 870 9048584 10.1210/en.138.3.863
Kuiper GG Lemmen JG Carlsson B Corton JC Safe SH van der Saag PT van der Burg B Gustafsson JA Interaction of estrogenic chemicals and phytoestrogens with estrogen receptor beta Endocrinology 1998 139 4252 4263 9751507 10.1210/en.139.10.4252
Hall JM McDonnell DP The estrogen receptor beta-isoform (ERbeta) of the human estrogen receptor modulates ERalpha transcriptional activity and is a key regulator of the cellular response to estrogens and antiestrogens Endocrinology 1999 140 5566 5578 10579320 10.1210/en.140.12.5566
Hewitt SC Korach KS Estrogen receptors: structure, mechanisms and function Rev Endocr Metab Disord 2002 3 193 200 12215714 10.1023/A:1020068224909
Simoncini T Mannella P Fornari L Caruso A Varone G Genazzani AR In vitro effects of progesterone and progestins on vascular cells Steroids 2003 68 831 836 14667975 10.1016/j.steroids.2003.08.006
Loomis AK Thomas P Effects of estrogens and xenoestrogens on androgen production by Atlantic Croaker testes in vitro: evidence for a non-genomic action mediated by an estrogen membrane receptor Biol Reprod 2000 62 995 1004 10727269
Razandi M Pedram A Merchenthaler I Greene GL Levin ER Plasma membrane estrogen receptors exist and functions as dimers Mol Endocrinol 2004 18 2854 2865 15231873 10.1210/me.2004-0115
Bjornstrom L Sjoberg M Mechanisms of estrogen receptor signaling: convergence of genomic and nongenomic actions on target genes Mol Endocrinol 2005 19 833 842 15695368 10.1210/me.2004-0486
Nitta H Bunick D Hess RA Janulis L Newton SC Millette CF Osawa Y Shizuta Y Toda K Bahr JM Germ cells of the mouse testis express P450 aromatase Endocrinology 1993 132 1396 1401 8440194 10.1210/en.132.3.1396
Levallet J Bilinska B Mittre H Genissel C Fresnel J Carreau S Expression and immunolocalization of functional cytochrome P450 aromatase in mature rat testicular cells Biol Reprod 1998 58 919 926 9546721
Bourguiba S Lambard S Carreau S Steroids control the aromatase gene expression in purified germ cells from the adult male rat J Mol Endocrinol 2003 31 83 94 12914527 10.1677/jme.0.0310083
Kuiper GG Gustafsson JA The novel estrogen receptor-beta subtype: potential role in the cell- and promoter-specific actions of estrogens and anti-estrogens FEBS Lett 1997 410 87 90 9247129 10.1016/S0014-5793(97)00413-4
Saunders PT Maguire SM Gaughan J Millar MR Expression of oestrogen receptor beta (ER beta) in multiple rat tissues visualised by immunohistochemistry J Endocrinol 1997 154 R13 16 9379111 10.1677/joe.0.154R013
van Pelt AM de Rooij DG van der Burg B van der Saag PT Gustafsson JA Kuiper GG Ontogeny of estrogen receptor-beta expression in rat testis Endocrinology 1999 140 478 483 9886860 10.1210/en.140.1.478
Rosenfeld CS Ganjam VK Taylor JA Yuan X Stiehr JR Hardy MP Lubahn DB Transcription and translation of estrogen receptor-beta in the male reproductive tract of estrogen receptor-alpha knock-out and wild-type mice Endocrinology 1998 139 2982 2987 9607809 10.1210/en.139.6.2982
Saunders PT Fisher JS Sharpe RM Millar MR Expression of oestrogen receptor beta (ER beta) occurs in multiple cell types, including some germ cells, in the rat testis J Endocrinol 1998 156 R13 17 9582517 10.1677/joe.0.156R013
Oliveira CA Nie R Carnes K Franca LR Prins GS Saunders PT Hess RA The antiestrogen ICI 182,780 decreases the expression of estrogen receptor-alpha but has no effect on estrogen receptor-beta and androgen receptor in rat efferent ductules Reprod Biol Endocrinol 2003 1 75 14613549 10.1186/1477-7827-1-75
Levine AC Kirschenbaum A Gabrilove JL The role of sex steroids in the pathogenesis and maintenance of benign prostatic hyperplasia Mt Sinai J Med 1997 64 20 25 8997072
McLachlan JA Newbold RR Bullock B Reproductive tract lesions in male mice exposed prenatally to diethylstilbestrol Science 1975 190 991 992 242076
Gill WB Schumacher GF Bibbo M Structural and functional abnormalities in the sex organs of male offspring of mothers treated with diethylstilbestrol (DES) J Reprod Med 1976 16 147 153 772199
Brand H Kos M Denger S Flouriot G Gromoll J Gannon F Reid G A novel promoter is involved in the expression of estrogen receptor alpha in human testis and epididymis Endocrinology 2002 143 3397 3404 12193552 10.1210/en.2001-210832
Brandenberger AW Tee MK Lee JY Chao V Jaffe RB Tissue distribution of estrogen receptors alpha (ER-alpha) and beta (ER-beta) mRNA in the midgestational human fetus J Clin Endocrinol Metab 1997 82 3509 3512 9329394 10.1210/jc.82.10.3509
Gaskell TL Robinson LL Groome NP Anderson RA Saunders PT Differential expression of two estrogen receptor beta isoforms in the human fetal testis during the second trimester of pregnancy J Clin Endocrinol Metab 2003 88 424 432 12519886 10.1210/jc.2002-020811
Pentikainen V Erkkila K Suomalainen L Parvinen M Dunkel L Estradiol acts as a germ cell survival factor in the human testis in vitro J Clin Endocrinol Metab 2000 85 2057 2067 10843196 10.1210/jc.85.5.2057
Taylor AH Al-Azzawi F Immunolocalisation of oestrogen receptor beta in human tissues J Mol Endocrinol 2000 24 145 155 10657006 10.1677/jme.0.0240145
Lambard S Galeraud-Denis I Saunders PT Carreau S Human immature germ cells and ejaculated spermatozoa contain aromatase and oestrogen receptors J Mol Endocrinol 2004 32 279 289 14766008 10.1677/jme.0.0320279
Saunders PT Sharpe RM Williams K Macpherson S Urquart H Irvine DS Millar MR Differential expression of oestrogen receptor alpha and beta proteins in the testes and male reproductive system of human and non-human primates Mol Hum Reprod 2001 7 227 236 11228242 10.1093/molehr/7.3.227
Makinen S Makela S Weihua Z Warner M Rosenlund B Salmi S Hovatta O Gustafsson JK Localization of oestrogen receptors alpha and beta in human testis Mol Hum Reprod 2001 7 497 503 11385105 10.1093/molehr/7.6.497
Enmark E Pelto-Huikko M Grandien K Lagercrantz S Lagercrantz J Fried G Nordenskjold M Gustafsson JA Human estrogen receptor beta-gene structure, chromosomal localization, and expression pattern J Clin Endocrinol Metab 1997 82 4258 4265 9398750 10.1210/jc.82.12.4258
Nilsson S Makela S Treuter E Tujague M Thomsen J Andersson G Enmark E Pettersson K Warner M Gustafsson JA Mechanisms of estrogen action Physiol Rev 2001 81 1535 1565 11581496
Saunders PT Millar MR Macpherson S Irvine DS Groome NP Evans LR Sharpe RM Scobie GA ERbeta1 and the ERbeta2 splice variant (ERbetacx/beta2) are expressed in distinct cell populations in the adult human testis J Clin Endocrinol Metab 2002 87 2706 2715 12050238 10.1210/jc.87.6.2706
Ogawa S Inoue S Watanabe T Orimo A Hosoi T Ouchi Y Muramatsu M Molecular cloning and characterization of human estrogen receptor betacx: a potential inhibitor of estrogen action in human Nucleic Acids Res 1998 26 3505 3512 9671811 10.1093/nar/26.15.3505
Smith EP Boyd J Frank GR Takahashi H Cohen RM Specker B Williams TC Lubahn DB Korach KS Estrogen resistance caused by a mutation in the estrogen-receptor gene in a man N Engl J Med 1994 331 1056 1061 8090165 10.1056/NEJM199410203311604
Morishima A Grumbach MM Simpson ER Fisher C Qin K Aromatase deficiency in male and female siblings caused by a novel mutation and the physiological role of estrogens J Clin Endocrinol Metab 1995 80 3689 3698 8530621 10.1210/jc.80.12.3689
Carani C Qin K Simoni M Faustini-Fustini M Serpente S Boyd J Korach KS Simpson ER Effect of testosterone and estradiol in a man with aromatase deficiency N Engl J Med 1997 337 91 95 9211678 10.1056/NEJM199707103370204
Herrmann BL Saller B Janssen OE Gocke P Bockisch A Sperling H Mann K Broecker M Impact of estrogen replacement therapy in a male with congenital aromatase deficiency caused by a novel mutation in the CYP19 gene J Clin Endocrinol Metab 2002 87 5476 5484 12466340 10.1210/jc.2002-020498
Hewitt SC Harrell JC Korach KS Lessons in estrogen biology from knockout and transgenic animals Annu Rev Physiol 2005 67 285 308 15709960 10.1146/annurev.physiol.67.040403.115914
Honda S Harada N Ito S Takagi Y Maeda S Disruption of sexual behavior in male aromatase-deficient mice lacking exons 1 and 2 of the cyp19 gene Biochem Biophys Res Commun 1998 252 445 449 9826549 10.1006/bbrc.1998.9672
Li X Nokkala E Yan W Streng T Saarinen N Warri A Huhtaniemi I Santti R Makela S Poutanen M Altered structure and function of reproductive organs in transgenic male mice overexpressing human aromatase Endocrinology 2001 142 2435 2442 11356692 10.1210/en.142.6.2435
Hess RA Bunick D Lee KH Bahr J Taylor JA Korach KS Lubahn DB A role for oestrogens in the male reproductive system Nature 1997 390 509 512 9393999 10.1038/37352
Kuiper GG Enmark E Pelto-Huikko M Nilsson S Gustafsson JA Cloning of a novel receptor expressed in rat prostate and ovary Proc Natl Acad Sci U S A 1996 93 5925 5930 8650195 10.1073/pnas.93.12.5925
Delbes G Levacher C Pairault C Racine C Duquenne C Krust A Habert R Estrogen receptor beta-mediated inhibition of male germ cell line development in mice by endogenous estrogens during perinatal life Endocrinology 2004 145 3395 3403 15044378 10.1210/en.2003-1479
Mahato D Goulding EH Korach KS Eddy EM Spermatogenic cells do not require estrogen receptor-alpha for development or function Endocrinology 2000 141 1273 1276 10698205 10.1210/en.141.3.1273
Mahato D Goulding EH Korach KS Eddy EM Estrogen receptor-alpha is required by the supporting somatic cells for spermatogenesis Mol Cell Endocrinol 2001 178 57 63 11403895 10.1016/S0303-7207(01)00410-5
Krege JH Hodgin JB Couse JF Enmark E Warner M Mahler JF Sar M Korach KS Gustafsson JA Smithies O Generation and reproductive phenotypes of mice lacking estrogen receptor beta Proc Natl Acad Sci U S A 1998 95 15677 15682 9861029 10.1073/pnas.95.26.15677
Weihua Z Makela S Andersson LC Salmi S Saji S Webster JI Jensen EV Nilsson S Warner M Gustafsson JA A role for estrogen receptor beta in the regulation of growth of the ventral prostate Proc Natl Acad Sci U S A 2001 98 6330 6335 11371645 10.1073/pnas.111150898
Ogawa S Chan J Chester AE Gustafsson JA Korach KS Pfaff DW Survival of reproductive behaviors in estrogen receptor beta gene-deficient (betaERKO) male and female mice Proc Natl Acad Sci U S A 1999 96 12887 12892 10536018 10.1073/pnas.96.22.12887
Akingbemi BT Ge R Rosenfeld CS Newton LG Hardy DO Catterall JF Lubahn DB Korach KS Hardy MP Estrogen receptor-alpha gene deficiency enhances androgen biosynthesis in the mouse Leydig cell Endocrinology 2003 144 84 93 12488333 10.1210/en.2002-220292
Delbes G Levacher C Duquenne C Racine C Pakarinen P Habert R Endogenous estrogens inhibit mouse fetal Leydig cell development via estrogen receptor alpha Endocrinology 2005 146 2454 2461 15661855 10.1210/en.2004-1540
Henry EC Miller RK Comparison of the disposition of diethylstilbestrol and estradiol in the fetal rat. Correlation with teratogenic potency Biochem Pharmacol 1986 35 1993 2001 2424456 10.1016/0006-2952(86)90732-X
Witorsch RJ Low-dose in utero effects of xenoestrogens in mice and their relevance to humans: an analytical review of the literature Food Chem Toxicol 2002 40 905 912 12065211 10.1016/S0278-6915(02)00069-8
Akingbemi BT Hardy MP Oestrogenic and antiandrogenic chemicals in the environment: effects on male reproductive health Ann Med 2001 33 391 403 11585100
Soto AM Sonnenschein C Chung KL Fernandez MF Olea N Serrano FO The E-SCREEN assay as a tool to identify estrogens: an update on estrogenic environmental pollutants Environ Health Perspect 1995 103 113 122 8593856
Routledge EJ Parker J Odum J Ashby J Sumpter JP Some alkyl hydroxy benzoate preservatives (parabens) are estrogenic Toxicol Appl Pharmacol 1998 153 12 19 9875295 10.1006/taap.1998.8544
Boonyaratanakornkit V Edwards DP Receptor mechanisms of rapid extranuclear signalling initiated by steroid hormones Essays Biochem 2004 40 105 120 15242342
Qian YM Sun XJ Tong MH Li XP Richa J Song WC Targeted disruption of the mouse estrogen sulfotransferase gene reveals a role of estrogen metabolism in intracrine and paracrine estrogen regulation Endocrinology 2001 142 5342 5350 11713234 10.1210/en.142.12.5342
Birnbaum LS Fenton SE Cancer and developmental exposure to endocrine disruptors Environ Health Perspect 2003 111 389 394 12676588
Bradlow HL Davis DL Lin G Sepkovic D Tiwari R Effects of pesticides on the ratio of 16 alpha/2-hydroxyestrone: a biologic marker of breast cancer risk Environ Health Perspect 1995 103 147 150 8593862
Rajapakse N Silva E Kortenkamp A Combining xenoestrogens at levels below individual no-observed-effect concentrations dramatically enhances steroid hormone action Environ Health Perspect 2002 110 917 921 12204827
Pennie WD Aldridge TC Brooks AN Differential activation by xenoestrogens of ER alpha and ER beta when linked to different response elements J Endocrinol 1998 158 R11 14 9846177 10.1677/joe.0.158R011
Pettersson K Grandien K Kuiper GG Gustafsson JA Mouse estrogen receptor beta forms estrogen response element-binding heterodimers with estrogen receptor alpha Mol Endocrinol 1997 11 1486 1496 9280064 10.1210/me.11.10.1486
Routledge EJ White R Parker MG Sumpter JP Differential effects of xenoestrogens on coactivator recruitment by estrogen receptor (ER) alpha and ERbeta J Biol Chem 2000 275 35986 35993 10964929 10.1074/jbc.M006777200
Naciff JM Daston GP Toxicogenomic approach to endocrine disrupters: identification of a transcript profile characteristic of chemicals with estrogenic activity Toxicol Pathol 2004 32 59 70 15503665 10.1080/01926230490463812
Safe S Clinical correlates of environmental endocrine disruptors Trends Endocrinol Metab 2005 16 139 144 15860409 10.1016/j.tem.2005.03.004
Ebling FJ Brooks AN Cronin AS Ford H Kerr JB Estrogenic induction of spermatogenesis in the hypogonadal mouse Endocrinology 2000 141 2861 9 10919273 10.1210/en.141.8.2861
Li H Papadopoulos V Vidic B Dym M Culty M Regulation of rat testis gonocyte proliferation by platelet-derived growth factor and estradiol: identification of signaling mechanisms involved Endocrinology 1997 138 1289 1298 9048638 10.1210/en.138.3.1289
Miura T Miura C Ohta T Nader MR Todo T Yamauchi K Estradiol-17beta stimulates the renewal of spermatogonial stem cells in males Biochem Biophys Res Commun 1999 264 230 234 10527870 10.1006/bbrc.1999.1494
Lassurguere J Livera G Habert R Jegou B Time- and dose-related effects of estradiol and diethylstilbestrol on the morphology and function of the fetal rat testis in culture Toxicol Sci 2003 73 160 169 12657744 10.1093/toxsci/kfg065
vom Saal FS Cooke PS Buchanan DL Palanza P Thayer KA Nagel SC Parmigiani S Welshons WV A physiologically based approach to the study of bisphenol A and other estrogenic chemicals on the size of reproductive organs, daily sperm production, and behavior Toxicol Ind Health 1998 14 239 260 9460178
Akingbemi BT Sottas CM Koulova AI Klinefelter GR Hardy MP Inhibition of testicular steroidogenesis by the xenoestrogen bisphenol A is associated with reduced pituitary luteinizing hormone secretion and decreased steroidogenic enzyme gene expression in rat Leydig cells Endocrinology 2004 145 592 603 14605012 10.1210/en.2003-1174
Takao T Nanamiya W Nazarloo HP Matsumoto R Asaba K Hashimoto K Exposure to the environmental estrogen bisphenol A differentially modulated estrogen receptor-alpha and -beta immunoreactivity and mRNA in male mouse testis Life Sci 2003 72 1159 1169 12505546 10.1016/S0024-3205(02)02364-0
Tena-Sempere M Navarro J Pinilla L Gonzalez LC Huhtaniemi I Aguilar E Neonatal exposure to estrogen differentially alters estrogen receptor alpha and beta mRNA expression in rat testis during postnatal development J Endocrinol 2000 165 345 357 10810299 10.1677/joe.0.1650345
vom Saal FS Timms BG Montano MM Palanza P Thayer KA Nagel SC Dhar MD Ganjam VK Parmigiani S Welshons WV Prostate enlargement in mice due to fetal exposure to low doses of estradiol or diethylstilbestrol and opposite effects at high doses Proc Natl Acad Sci U S A 1997 94 2056 2061 9050904 10.1073/pnas.94.5.2056
Wilcox AJ Baird DD Weinberg CR Hornsby PP Herbst AL Fertility in men exposed prenatally to diethylstilbestrol N Engl J Med 1995 332 1411 1416 7723797 10.1056/NEJM199505253322104
Strohsnitter WC Noller KL Hoover RN Robboy SJ Palmer JR Titus-Ernstoff L Kaufman RH Adam E Herbst AL Hatch EE Cancer risk in men exposed in utero to diethylstilbestrol J Natl Cancer Inst 2001 93 545 551 11287449 10.1093/jnci/93.7.545
Leffers H Naesby M Vendelbo B Skakkebaek NE Jorgensen M Oestrogenic potencies of Zeranol, oestradiol, diethylstilboestrol, Bisphenol-A and genistein: implications for exposure assessment of potential endocrine disrupters Hum Reprod 2001 16 1037 1045 11331657 10.1093/humrep/16.5.1037
Swerdlow AJ De Stavola BL Swanwick MA Maconochie NE Risks of breast and testicular cancers in young adult twins in England and Wales: evidence on prenatal and genetic aetiology Lancet 1997 350 1723 1728 9413462 10.1016/S0140-6736(97)05526-8
Skakkebaek NE Testicular dysgenesis syndrome: new epidemiological evidence Int J Andrology 2004 27 189 191 10.1111/j.1365-2605.2004.00488.x
Virtanen HE Raipert-De Meyts E Main KM Skakkebaek NE Toppari J Testicular dysgenesis syndrome and the development and occurrence of male reproductive disorders Toxicol Appl Pharmacol 2005 16005920
Rivas A Fisher JS McKinnell C Atanassova N Sharpe RM Induction of reproductive tract developmental abnormalities in the male rat by lowering androgen production or action in combination with a low dose of diethylstilbestrol: evidence for importance of the androgen-estrogen balance Endocrinology 2002 143 4797 4808 12446607 10.1210/en.2002-220531
Sharpe RM Rivas A Walker M McKinnell C Fisher JS Effect of neonatal treatment of rats with potent or weak (environmental) oestrogens, or with a GnRH antagonist, on Leydig cell development and function through puberty into adulthood Int J Androl 2003 26 26 36 12534935 10.1046/j.1365-2605.2003.00385.x
Atanassova N McKinnell C Turner KJ Walker M Fisher JS Morley M Millar MR Groome NP Sharpe RM Comparative effects of neonatal exposure of male rats to potent and weak (environmental) estrogens on spermatogenesis at puberty and the relationship to adult testis size and fertility: evidence for stimulatory effects of low estrogen levels Endocrinology 2000 141 3898 3907 11014247 10.1210/en.141.10.3898
Thigpen JE Setchell KD Ahlmark KB Locklear J Spahr T Caviness GF Goelz MF Haseman JK Newbold RR Forsythe DB Phytoestrogen content of purified, open- and closed-formula laboratory animal diets Lab Anim Sci 1999 49 530 536 10551455
Thigpen JE Setchell KD Saunders HE Haseman JK Grant MG Forsythe DB Selecting the appropriate rodent diet for endocrine disruptor research and testing studies ILAR J 2004 45 401 416 15454679
Munro IC Harwood M Hlywka JJ Stephen AM Doull J Flamm WG Adlercreutz H Soy isoflavones: a safety review Nutr Rev 2003 61 1 33 12638461 10.1301/nr.2003.janr.1-33
Cooke B Hegstrom CD Villeneuve LS Breedlove SM Sexual differentiation of the vertebrate brain: principles and mechanisms Front Neuroendocrinol 1998 19 323 362 9799588 10.1006/frne.1998.0171
Lindzey J Wetsel WC Couse JF Stoker T Cooper R Korach KS Effects of castration and chronic steroid treatments on hypothalamic gonadotropin-releasing hormone content and pituitary gonadotropins in male wild-type and estrogen receptor-alpha knockout mice Endocrinology 1998 139 4092 4101 9751487 10.1210/en.139.10.4092
Oettel M Hubler D Patchev V Selected aspects of endocrine pharmacology of the aging male Exp Gerontol 2003 38 189 198 12543277 10.1016/S0531-5565(02)00185-7
Zhao L Bakke M Hanley NA Majdic G Stallings NR Jeyasuria P Parker KL Tissue-specific knockouts of steroidogenic factor 1 Mol Cell Endocrinol 2004 215 89 94 15026179 10.1016/j.mce.2003.11.009
Das SK Tan J Raja S Halder J Paria BC Dey SK Estrogen targets genes involved in protein processing, calcium homeostasis, and Wnt signaling in the mouse uterus independent of estrogen receptor-alpha and -beta J Biol Chem 2000 275 28834 28842 10893236 10.1074/jbc.M003827200
Gaido KW Leonard LS Maness SC Hall JM McDonnell DP Saville B Safe S Differential interaction of the methoxychlor metabolite 2,2-bis-(p-hydroxyphenyl)-1,1,1-trichloroethane with estrogen receptors alpha and beta Endocrinology 1999 140 5746 5753 10579340 10.1210/en.140.12.5746
Marquez DC Pietras RJ Membrane-associated binding sites for estrogen contribute to growth regulation of human breast cancer cells Oncogene 2001 20 5420 5430 11571639 10.1038/sj.onc.1204729
Razandi M Alton G Pedram A Ghonshani S Webb P Levin ER Identification of a structural determinant necessary for the localization and function of estrogen receptor alpha at the plasma membrane Mol Cell Biol 2003 23 1633 1646 12588983 10.1128/MCB.23.5.1633-1646.2003
Stokes WS Introduction: selecting appropriate animal models and experimental designs for endocrine disruptor research and testing studies ILAR J 2004 45 387 393 15454677
Bigsby RM Caperell-Grant A Madhukar BV Xenobiotics released from fat during fasting produce estrogenic effects in ovariectomized mice Cancer Res 1997 57 865 869 9041187
Nagel SC vom Saal FS Thayer KA Dhar MG Boechler M Welshons WV Relative binding affinity-serum modified access (RBA-SMA) assay predicts the relative in vivo bioactivity of the xenoestrogens bisphenol A and octylphenol Environ Health Perspect 1997 105 70 76 9074884
Nagel SC vom Saal FS Welshons WV The effective free fraction of estradiol and xenoestrogens in human serum measured by whole cell uptake assays: physiology of delivery modifies estrogenic activity Proc Soc Exp Biol Med 1998 217 300 309 9492339
|
16188042
|
PMC1262773
|
CC BY
|
2021-01-04 16:37:12
|
no
|
Reprod Biol Endocrinol. 2005 Sep 27; 3:51
|
utf-8
|
Reprod Biol Endocrinol
| 2,005 |
10.1186/1477-7827-3-51
|
oa_comm
|
==== Front
Reprod HealthReproductive Health1742-4755BioMed Central London 1742-4755-2-71619754610.1186/1742-4755-2-7ResearchEvaluation of a strict protocol approach in managing women with severe disease due to hypertension in pregnancy: A before and after study Lombaard Hennie [email protected] Robert C [email protected] Fèbè [email protected] Peter [email protected] Department of Obstetrics and Gynaecology, Kalafong Hospital, Private Bag X396, Pretoria 0001, South Africa2 MRC Maternal and Infant Health Care Strategies Research Unit and Obstetrics and Gynaecology Department, University of Pretoria, South Africa2005 30 9 2005 2 7 7 3 5 2005 30 9 2005 Copyright © 2005 Lombaard et al; licensee BioMed Central Ltd.2005Lombaard et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
To evaluate whether the introduction of a strict protocol based on the systemic evaluation of critically ill pregnant women with complications of hypertension affected the outcome of those women.
Method
Study group: Indigent South African women managed in the tertiary hospitals of the Pretoria Academic Complex. Since 1997 a standard definition of women with severe acute maternal morbidity (SAMM), also referred to as a Nearmiss, has been used in the Pretoria Academic Complex. All cases of SAMM and maternal deaths (MD) were entered on the Maternal Morbidity and Mortality Audit System programme (MaMMAS). A comparison of outcome of severely ill women who had complications of hypertension in pregnancy was performed between 1997–1998 (original protocol) and 2002–2003 (strict protocol). Data include women referred from outside the Pretoria Academic Complex area to the tertiary hospitals.
Results
Between 1997–1998 there were 79 women with SAMM and 18 maternal deaths due to complications of hypertension, compared with 91 women with SAMM and 13 maternal deaths in 2002–2003. The mortality index (MI) declined from 18.6% to 12.5% (OR 0.62, 95% CI 0.27–1.45). Statistically significant fewer women had renal failure (RR 0.37, 95% CI 0.21 – 0.66) and cerebral complications (RR 0.52, 95%CI 0.34 – 0.81) during the second period, and liver dysfunction (RR 0.27 95%CI 0.06 – 1.25) tended to be lower. However, there tended to be an increase in the number of women, who had immune system failure (RR 4.2 95%CI 0.93 – 18.94) and respiratory failure (RR 1.42 95%CI 0.88 – 2.29) although it did not reach significance. Cardiac failure remained constant (RR 0.84 95%CI 0.54 – 1.30).
Conclusion
The strict protocol approach based on the systemic evaluation of severely ill pregnant women with complications of hypertension and an intensive, regular feedback mechanism has been associated with a reduction in the number of patients with renal failure and cerebral compromise.
==== Body
Background
Complications due to hypertensive disease in pregnancy are the most common direct cause of maternal death (MD) in South Africa [1,2]. They are also the most common cause of acute severe morbidity in a survey of three clearly defined geographical areas in South Africa [3]. The final and contributory causes of deaths were most commonly cerebral complications (40% and 50%), cardiac failure (40% and 35%), respiratory failure (15% and 16%) and renal failure (10% and 18%) in 1998 and 1999–2001 respectively [1,2]. Any strategy for improving the outcome of severely ill pregnant women with complications due to hypertension would have to address ways of preventing these organ systems from failing.
In 1997, the Pretoria Academic Complex initiated the use of a standard definition of severe acute maternal morbidity (SAMM) [4] and the routine collection of data since. Table 1 shows the definition for organ system dysfunction for each specific organ system. After the initial survey in 1997–1998, protocols were reviewed and a new approach to managing severely ill pregnant women was introduced. This system was based on the systematic routine evaluation of all organ systems and, where an abnormality was detected, this served as a trigger to further investigate and support that organ system [5]. This policy was adopted for managing critically ill pregnant women with specific common conditions and as a group became known as "the strict protocols".
Table 1 The criteria for a near-miss case for each specific organ system according to Mantel.
Organ system-based Markers
Cardiac dysfunction Pulmonary oedema
Cardiac arrest
Vascular dysfunction Hypovolaemia requiring ≥ 5 units of blood products
Immunological dysfunction ICU admission for sepsis
Emergency hysterectomy for sepsis
Respiratory dysfunction Intubation and ventilation for any reason other than general anaesthesia
Oxygen saturation of less than 90% for more than 60 min
The ratio of the partial pressure of oxygen in arterial blood to the percentage oxygen in inspired air is ≤ 3 (paO2/FiO2 ≤ 3)
Renal dysfunction Oliguria, ≤ 400 ml/24 hr that does not respond to careful fluid replacement or attempts at inducing with dopamine or infusion
Acute deterioration in urea > 15 mmol/l or of creatinine to > 400 mmol/l
Liver dysfunction Jaundice in the presence of pre-eclampsia
Metabolic dysfunction Diabetic keto-acidosis
Coagulation dysfunction Acute thrombocytopenia requiring a platelet transfusion
Cerebral dysfunction Coma lasting > 12 hours
Subarachnoid or intracerebral haemorrhage
The implementation of new policies can be difficult to achieve [6]. The change in policy was effected by regular reinforcement at daily audit meetings at the individual hospitals and weekly departmental meetings where all cases of SAMM and MD were discussed. New trainees were introduced to the protocols from the start, face-to-face meetings were held when problems were detected, and special training sessions were introduced to stress specific points. Finally, a report was compiled and presented to the provincial administration with specific recommendations appropriate to the region. There has been ongoing regular contact since then.
This study examines the effect of the introduction of this strict protocol on the outcome of critically ill pregnant women with complications of hypertension in pregnancy and attempts to identify where further research or change in policy is required.
Methods
The Pretoria Academic Complex consists of two academic hospitals (Pretoria Academic and Kalafong) and two district hospitals (Mamelodi Day and Pretoria West). The area receives referrals from other areas within Gauteng Province and the surrounding provinces. The Pretoria Academic Complex serves a population of 3 million people. The population served is mainly an indigent general South African population. The female population was 554 000 in 1996 with an annual growth of 10%, the fertility rate for Gauteng is 2.3% for women aged 15 to 49 and the current pregnancy rate is 2.2% [17]. The number of births within the immediate area has increased by twenty percent over the last five years and is currently approximately 15 000 per year. For the period of 1997/1998, the total number of births was 27 025, the maternal mortality ratio was 133.2/100 000 births and the perinatal mortality rate was 37/1000 for babies weighing more than 1000 gr. For the period of 2002/2003 the total number of births was 32 814, the maternal mortality ratio was 115.8/100 000 births and the perinatal mortality rate for babies weighing more than 1000 gr was 27/1000 births.
The criteria for SAMM have been defined by Mantel et al. [4] as a woman with organ system dysfunction or failure who would probably have died if un- or inadequately treated. The criteria are based on clinical evaluation with limited specific investigations that are readily available at all regional, secondary and higher level hospitals.
Data on women with SAMM and MD were collected every morning at the respective hospitals and a near miss form was completed for each woman with SAMM and the maternal death notification form for all maternal deaths. The data were entered into the Maternal Morbidity and Mortality Audit System (MaMMAS) database, developed by Johan Coetzee (Simply Software). Data from 1997/1998 (initial project) [3] served as the platform for developing the database. This initial project compared Maternal Mortality and Nearmiss to determine if disease pattern was the same. In the second time period (2002/2003) the strict protocol had been implemented and became established.
Standard statistical techniques were used to compare the two time periods. The chi square test was used to compare categorical data. The outcome measures were the Mortality Index, defined as the number of maternal deaths divided by the sum of women with SAMM and maternal deaths, expressed as a percentage [7]. The criteria for each organ system dysfunction/failure were specified in the definition of a woman with SAMM [4] and the rates of each organ system dysfunction were compared between the two time periods. The relative risk was used to compare the two time periods as it gives a more realistic indication of the changes.
During the first time period (1997/1998) there was no strict protocol available and standard care was provided. This consisted of magnesium sulphate for eclamptic fits and hydralazine 1,25 mg ivi every 15 min for the acute treatment of hypertension if the diastolic BP was ≥ 110 mmHg. A fluid balance chart was not routinely kept and the management of the patient was mainly based on the opinion of the attending consultant.
In contrast, the strict management protocol, implemented during the second time period (2002/2003) included the following:
Stabilization
The stabilization of the patient is summarised in Table 2.
Table 2 Summary of stabilising the severely ill women with complications of Hypertension after admission to a High Care Obstetrics unit
Organ system Acute management Maintenance Management of complications
Fluid management Start IV line give 300 ml fluid bolus: 100 ml Ringers lactate
200 ml normal saline with loading dose of magnesium sulphate Urinary Catheter Give Ringers lactate 125 ml/hr iv.
Start a fluid balance chart If poor output repeat fluid bolus. If still poor output and positive fluid balance start low-dose dopamine infusion
Magnesium Sulphate 4 g magnesium sulphate in 200 ml saline over 20 min iv
5 g magnesium sulphate with 1 ml lignocaine im in each buttock Maintenance: 5 g four hourly iv
Check before next dosage: Urine output > 30 ml/hr
Tendon reflexes present
Respiratory rate more than 16/min In case of magnesium sulphate overdose give calcium gluconate
Blood pressure control Repeat blood pressure after 20 min and if diastolic ≥ 110 or systolic ≥ 160 treat according to the antihypertensive drug protocol Use either nifedipine or labetolol
Neurological status If still confused check saturation and blood pressure Abnormal saturation: Give oxygen via mask
Abnormal blood pressure: treat with appropriate drugs If both are normal: give haloperidol
Systemic Evaluation
Table 3 summarizes the systemic evaluation of the mother.
Table 3 Summary of the systemic evaluation and special investigations of critical ill women with complications of hypertension
Organ system evaluated Clinical examination Special investigations
Central nervous system Glasgow coma scale
Lateralising signs
Reflexes
Pupil reflexes If any abnormalities consider CT Scan
Respiratory system Respiratory rate
Blood gas
Check for dullness on percussion, crepitations or wheezes If any abnormalities do blood gas and Chest X-ray
Cardiovascular system: Pulse, Blood pressure
Heart sounds
Heart size
Radio-femoral delay
Gastro intestinal system: Check for epigastric tenderness, hepatomegaly Check AST and for jaundice. 4 hourly blood glucose test if raised AST
Renal system: Check for renal angle tenderness, macroscopic hematuria
Listen for murmurs over the renal artery Check creatinine and fluid balance. If signs of kidney dysfunction do full kidney function tests
Haematological system: Check for anaemia, purpera, bleeding tendency Check hematocrit and platelets
Immune system: Body temperature
Check for generalized lymphadenopathy, splenomegaly, signs of immune system failure Voluntary counselling and HIV testing if CD4 and ESR above 100
Musculosceletal System Check for signs of DVT
Check for spinal problems that might influence the type of anaesthesia
Gynaecological system: Abdominally: measure symphysis-fundus height, lie & position of the foetus, check for uterine tenderness or contractions, estimate foetal weight, measure amniotic fluid, check for foetal heart rate
Vaginal exam: assess the Bishop score
Fundoscopy: Check for silver wiring, papillar oedema and signs of bleeding
Foetal Evaluation
An ultrasonographic evaluation of the foetus was performed, once the mother had been stabilised, and included the following:
• Estimated foetal weight
• Doppler of the umbilical artery
• Amniotic fluid index
• Transcerebellar diameter
• Middle cerebral artery Doppler
• Ductus venosus waveform
• Detection of possible structural abnormalities
If the expected foetal weight was more than 800 g or the foetus was known to have a gestational age of 28 weeks or more, the foetus was regarded as viable. In these cases corticosteroids were administered and the foetal heart rate pattern was monitored six hourly with a cardiotocograph (CTG).
After all this information had been gathered a management plan was formulated. In most cases the laboratory blood results (initially only aspartate-amino transferase (AST), creatinine, haematocrit and platelets) were available within the hour and during the stabilisation phase of the patient. To make a final decision four questions needed to be answered by the clinician, namely:
1. Is it safe for the mother to continue the pregnancy?
2. Is it safe for the foetus to continue the pregnancy?
3. What is the risk to neonate if the foetus was born?
4. What is the risk to the mother if the foetus was born?
In case of expectant management, the woman was transferred to a high care (high dependency unit) obstetric unit, with daily evaluation by the registrar and routine blood testing (haematocrit, platelets, creatinine and AST) twice weekly. Cardiotocogram (CTG) (6 hourly) and ultrasound (two-weekly) were used to assess foetal wellbeing and growth. If there was concern about foetal growth, ultrasound was repeated more often [10].
The indications for delivery were:
• Foetal distress
• Intra uterine death
• Expected birth weight more than 2 kg
• Expected birth weight less than 500 g
• Maternal organ system failure
• Uncontrollable hypertension
• Eclampsia
• Proven foetal lung maturity
• Foetal abnormality
The indications for elective caesarean section were:
• Unfavourable Bishop score (4 or less)
• Foetal distress as diagnosed on the basis of spontaneous decelerations on CTG
• Absent end diastolic flow of the uterine artery on Doppler
• An abnormal ductus venosus waveform. The ductus venosus wave form was regarded as abnormal if it was absent or reversed.
Antihypertensive drug protocol
Alpha methyldopa was used as the first line oral antihypertensive agent [10] (an initial dose of 500 mg 8 hourly was used increasing to a maximum of 750 mg 8 hourly). In cases of severe hypertension (blood pressure more than 160 mmHg systolic or 110 mmHg diastolic) 10 mg nifedipine [11] was administered orally. Nifedipine (10 mg) has replaced dihydralazine in the strict protocol as it is more effective in controlling blood pressure and is safer for the foetus [11,12]. If the blood pressure remained higher than 160/110 mmHg after one hour a repeat dose of nifedipine was given. If the woman was unable to swallow or had a tachycardia of more than 120 beats per minute, labetolol [13] was administered: the patient was started with 20 mg iv and if the blood pressure remained above 160/110 mmHg after 10 min she would receive 40 mg iv. If her blood pressure still remained above 160/110 mmHg she would receive 80 mg iv. This would be repeated another two times 10 minutes apart if needed.
Fluid management
After admission to the High Care Obstetric Unit, the woman was maintained on 125 ml/hr Ringers lactate intravenously. A strict input and output chart was recorded by the nursing staff. The attending registrar regularly determined the fluid balance taking into account insensible and blood loss. If the patient was excreting less than 30 ml per hour and was in a negative balance a 300 ml Ringers lactate bolus was administered. If the patient had positive fluid balance a low dose dopamine infusion was started [9].
Ethical considerations
The Ethics Committee of the Faculty of Health Sciences gave approval for the initial study and the programme remains registered with the Ethics Committee. The hospital administration at each hospital continues to give approval for the audit. Patient information is anonymised once entered into the database.
Results
In 1997/1998 there were 27,025 births in the Pretoria Academic Complex and 32,814 births in 2002/2003, an increase of 21.4%. This excludes all referrals and patients born in private institutions. The Maternal Mortality Ratio (MMR) for indigent patients from the Pretoria area, dying from complications of hypertension in pregnancy remained unchanged with 9 deaths in 1997/1998 (MMR 33.3/100000 births) and 8 deaths in 2002/2003 (MMR 24.4/100000 births). However the prevalence of women severely ill from complications of hypertension in pregnancy in our population rose from 40 patients (0.15%) in 1997/1998 to 64 patients (0.20%) in 2002/2003. The denominators for referred patients are unknown.
Auditing all patients with complicated hypertension demonstrated that in 1997/1998 there were 79 SAMM and a further 18 maternal deaths compared to 91 women with SAMM and 13 maternal deaths in 2002/2003. The distribution of age and parity in both groups was similar: age less than 20 years (11% and 9.6%), 20 to 29 years (50.6% and 52.8%), 30 to 39 years (28.9% and 31.7%) and primigravid (42.3% and 43.2%).
Table 4 compares the distribution of the different categories of women with severe complications of hypertension in pregnancy. Statistically significantly more women had severe complications following eclampsia during 1997/1998 compared to 2002/2003, but the opposite was true for women with severe complications following HELLP syndrome. It is important to note that the diagnosis of eclampsia or HELLP syndrome does not automatically classify a woman as having severe morbidity as there also has to be organ system dysfunction as defined above. Overall, the Mortality Index declined from 18.6% in 1997/1998 to 12.5% in 2002/2003 (Odds Ratio 0.62, 95% CI 0.27–1.45). There were no differences in Mortality Indices within the individual disease categories.
Table 4 Comparison between the sub-categories of complications of hypertension in pregnancy and their Mortality Indices.
1997–1998 2002–2003 p MI
MD SAMM Total % MI MD SAMM Total % MI
Chronic Hypertension 1 2 3 3.1 33.3 1 4 5 4.8 20.0
Proteinuric Hypertension 6 22 28 28.9 21.4 4 30 34 32.7 11.8 0.49
Eclampsia* 9 38 47 48.5 19.1 6 22 28 26.9 21.4 0.52
HELLP** 2 17 19 19.6 10.5 1 35 36 34.6 2.8 0.27
Liver rupture 0 0 0 0.0 0.0 1 0 1 1.0 100.0
All Hypertension 18 79 97 100 18.5 13 91 104 100 12.5 0.32
MD – Maternal Death; SAMM – Severe acute maternal morbidity; MI – Mortality Index
* – Significant decline in proportion of eclampsia from 1997/8 to 2002/3, p = 0.0026
** – Significant increase in proportion of women with HELLP syndrome 1997/8 to 2002/3, p = 0.026
Table 5 compares the organ system dysfunction/failure for severely ill pregnant women with complications of hypertension. Significantly fewer women had renal failure (34.0% vs 12.5%, RR 0.37, 95% CI 0.21 – 0.66) and cerebral complications (35.1% vs 14.4%, RR 0.52 95%CI 0.34 – 0.81) than before, and liver dysfunction (7.2% vs 1.9%, RR 0.27 95%CI 0.06 – 1.25) tended to be lower. However, there was a trend towards an increase in number of women who had immune system failure (2.0% to 8.7%, RR 4.2 95%CI 0.93 – 18.94) and respiratory failure (21.6% to 30.8%, RR 1.42 95%CI 0.88 – 2.29) in the second period compared to the first although it did not reach significance. Cardiac failure remained constant (30.9% and 26.0% RR 0.84 95%CI 0.54 – 1.30).
Table 5 Comparison of the prevalence of organ system dysfunction/failure per severely ill pregnant women with complications due to hypertension.
Organ system 1997–1998 2002–2003 RR (95% CI)
SAMM n = 79 MD n = 18 SAMM+MD N = 97 % OSD SAMM N = 91 MD N = 13 SAMM+MD N = 104 % OSD
Hypovolaemic shock 7 1 8 8.2 5 1 6 5.8 0.7 (0.25 – 1.94)
Respiratory failure 17 4 21 21.6 29 3 32 30.8 1.42 (0.88 – 2.29)
Cardiac failure 25 5 30 30.9 23 4 27 26.0 0.84 (0.54 – 1.30)
Renal failure 29 4 33 34.2 11 2 13 12.5 0.37 (0.21 – 0.66)
Liver failure 5 2 7 7.2 1 1 2 1.9 0.27 (0.06 – 1.25)
Cerebral complications 24 10 34 35.1 9 6 15 14.4 0.52 (0.34 – 0.81)
Haematological dysfunction 25 4 29 29.9 26 1 27 26.0 0.87 (0.56 – 1.36)
Immune system failure* 1 1 2 2.1 6 3 9 8.7 4.2 (0.93 – 18.94)
% OSD – Percentage of severely ill women who developed that organ system dysfunction/failure
Note: A patient can have more than one organ system dysfunction/failure
* Fisher exact: 2 sided 0.060
: 1 sided 0.038
The average number of organ systems that failed or were dysfunctional for all critically ill women were 1.69 in 1997/1998 and 1.25 in 2002/2003. This indicates that per critically patient there were less severely compromised organ systems in the second period compared to the first.
Table 6 examines the patients that came from the Pretoria area only. The pattern of organ system dysfunction/failure was the same but the numbers were too small to have sufficient power to detect significant differences. The Mortality Index in Pretoria for 1997/1998 was 22.5% and 12.5% for 2002/2003 (OR 0.49, 95% CI 0.15–1.57).
Table 6 Comparison of the prevalence of organ system dysfunction/failure per severely ill pregnant women with complications due to hypertension for patients only from the Pretoria area.
Organ system 1997–1998 2002–2003
SAMM n = 31 MD n = 9 SAMM+MD N = 40 % OSD SAMM N = 56 MD N = 8 SAMM+MD N = 64 % OSD
Hypovolaemic shock 2 0 2 5.0 3 1 4 6.3
Respiratory failure 7 3 10 25.0 16 2 18 28.1
Cardiac failure 12 4 16 40.0 13 3 16 25.0
Renal failure 6 1 7 17.5 5 1 6 9.4
Liver failure 0 0 0 0.0 1 1 2 3.1
Cerebral complications 8 4 12 30.0 4 3 7 10.9
Haematological dysfunction 9 1 10 25.0 19 1 20 31.3
Immune system failure 0 1 0 2.5 2 2 4 6.3
% OSD – Percentage of severely ill women who developed that organ system dysfunction/failure
Note: A patient can have more than one organ system dysfunction/failure
Discussion
This report of managing critically ill pregnant women with complications of hypertension in pregnancy during two time periods five years apart has shown a change in pattern of organ system dysfunction/failure. Fewer women developed renal failure and cerebral complications during the later time period. However, other common complications of the respiratory, cardiovascular and immune systems did not seem to be affected by the implementation of the strict protocol.
The possible explanations for the change are better identification of patients with complications and better management, and fewer referrals of critically ill patients from outside areas. By increasing the number of cases identified with improved surveillance, one would have expected a sudden increase in the rate of critically ill women being treated followed by a levelling off. However, there has been a steady increase in the rate from 84/100000 births in 1999, 108/100000 births in 2000 and 138/100000 births in 2001 [14].
The reduction in critically ill patients referred from outside the Pretoria area could also account for the change in our findings. The longer time to get into a tertiary unit is likely to increase the severity of the complications. Thus the shift from 61% of critically ill women being referred in 1997/1998 as opposed to only 48% in 2002/2003 could account for the change. Examining only the cases from the Pretoria area (Table 6), the same shift in pattern is seen although the number of cases was too few to detect significant changes.
The final explanation is that the management of the patients has improved. This is supported by the decreased number of dysfunctional organ systems per patient; it fell from 1.69 in 1997/1998 to 1.25 in 2002/2003. This change may be associated with the introduction and adherence to a strict protocol for managing such critically ill women. It is postulated that using the strict protocol may have prevented further organ systems from becoming involved. The strict protocol was based on the audit in 1997/1998 and the introduction of treatment research findings such as the use of magnesium sulphate for women with severe pre-eclampsia [8]. The feedback system of regular audit meetings to discuss cases, one by one meetings and regular teaching sessions to discuss specific issues confirms the success found in previous randomised trials [6]. The rapid turnover of staff necessitates regular reinforcement of protocols.
One of the limitations of our study was that we used a 'before and after' design to evaluate the effect of the implementation of a strict protocol with no control patients in any of the time periods. Further, the study could not explain the fact that there was no reduction in cardiac and respiratory problems.
The lack of change in the respiratory and cardiovascular system organ dysfunction/failure suggests that there is still significant room for improving the protocol. The confidential enquiries into maternal deaths in the UK revealed that 37% of maternal deaths due to eclampsia or pre-eclampsia during the 1985–1999 period are being ascribed to "pulmonary" [15]. Adult respiratory distress syndrome has been described as contributing factor in 28% [16] of deaths associated with HELLP syndrome. Cardiac and respiratory functions in severe pre-eclampsia are not well understood and future research must attempt to identify the changes that can improve the outcome. The increasing incidence of HIV infection in the area accounts for the rise in immune system failure. This may also affect pulmonary and cardiac function.
The 33% increase of women who were critically ill due to complications of hypertension from the Pretoria area is disturbing. The reason for this is unknown. There has been a steady influx of people into the Gauteng area during the last 5 years. This is confirmed by the 2001 Census [17] and the increase of more than twenty percent of births in the Pretoria area. This influx of many impoverished people looking for employment would be expected to bring more unhealthy people to the region. This will lead to increased pressure on the primary health care services to identify these patients early (at antenatal clinics) and refer appropriately, and increase the pressure on secondary and tertiary services that must deal with the increased load of critically ill patients. Appropriate plans for distribution of resources need to be made to deal with this challenge.
Conclusion
The strict protocol approach based on the systemic evaluation of critically ill pregnant women due to complications of hypertension has been associated with a reduction in the preventable complications, such as renal failure and cerebral compromise by improved fluid management and blood pressure control. However, there has been no change in the prevalence of cardiac or respiratory failure. Cardiac and respiratory function in women with severe hypertension in pregnancy needs further investigation and strategies need to be developed to improve its management.
List of Abbreviations
SAMM: Severe acute maternal morbidity
MaMMAS: Maternal Morbidity and Mortality Audit System programme
MD: Maternal Deaths
AIDS: Acquired Immune Deficiency Syndrome
HELLP: Haemolysis, Elevated Liver Enzymes, Low Platelets
AST: Aspartate-amino transferase
UK: United Kingdom
OR: Odds Ratio
MMR: Maternal Mortality Ratio
CTG: Cardio tochography
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
RCP and PM wrote the initial protocol and coordinated the collection of the SAMM data. FB entered the data on the MaMMAS programme. HL and FB collected the results from the MaMMAS programme. HL, FB and RCP did the statistical analysis. HL wrote the article. RCP and PM substantially improved the manuscript.
Acknowledgements
Dr J Makin help with the statistical analysis of the study. Mrs Sue Sellers and Dr SS Smith helped with editing the article.
==== Refs
Moodley J Nyasulu D Pattinson RC Hypertensive disorders of pregnancy Saving Mothers: Report on confidential enquiries into maternal deaths in South Africa, 1998 1999 Government Printer, Pretoria 28 39
Moodley J Molefe N Pattinson RC Hypertensive disorders of pregnancy Saving Mothers 1999–2001: Second report on confidential enquiries into maternal deaths in South Africa 2003 Government Printer, Pretoria 37 54
Pattinson RC Buchmann EJ Mantel G Schoon M Rees H Can enquiries into severe acute maternal morbidity act as a surrogate for maternal death enquiries? BJOG 2003 110 889 893 14550357
Mantel GD Buchmann E Rees H Pattinson RC Severe acute maternal morbidity: a pilot study of a definition for a near-miss BJOG 1998 105 985 990
Pattinson RC Mantel G Moodley J A systematic approach to examining an ill pregnant woman Saving Mothers: Policy and management guidelines for common causes of maternal deaths 2001 Government Printer 6 23
Grimshaw JM Shirran L Thomas R Changing provider behaviour: An overview of systematic reviews of interventions Medical Care 2001 39 II-2 II-45 11583120
Vandecruis H Pattinson RC Macdonald AP Changing patterns of severe acute maternal morbidity and mortality in the Pretoria Region Eur J Obstet Gynaecol Reprod Biol 2002 102 6 10 10.1016/S0301-2115(01)00558-9
The Magpie Collaborative Group Do women with pre-eclmpsia, and their babies, benefit from magnesium sulphate? The Magpie Trail: A randomised placebo controlled trail The Lancet 2002 359 1877 90 12057549 10.1016/S0140-6736(02)08778-0
Mantel GD Makin JD Low dose Dopamine in postpartum pre-eclamptic women with oliguria: a double-blind, placebo controlled, randomised trail BJOG 1997 104 1180 1183
Hall DR Odendaal HJ Steyn DW Expectant management severe pre-eclampsia in the mid-trimester European J Obstet Gynecol Reprod Biol 2001 96 168 172 11384801 10.1016/S0301-2115(00)00449-8
Smith P Anthony J Johnson R Nifedipine in pregnancy BJOG 2000 107 299 307 10740323
Hall DR Odendaal HJ Steyn DW Smith M Nifedipine or prazosin as a second agent to control early severe hypertension in pregnancy: a randomised controlled trial BJOG 2000 107 759 765 10847232
Collins R Duley L Enkin MW, Keirse MJNC, Renfrew MJ, Neilson JP Labetolol vs hydralazine in severe pregnancy induced hypertension Pregnancy and Childbirth Module of the Cochrane Database of Systemic Reviews 1995 BMJ Publishing group, London
Pattinson RC Hall M Near Misses. A useful adjunct to maternal mortality audits Br Med Bull 2003 67 231 243 14711767 10.1093/bmb/ldg007
Drife J Lewis G editors Why Mothers Die Confidential Enquiries into Maternal Deaths in the United Kingdom 1997–1999 2001 London: RCOG Press
Isler CM Rinehart BK Terrone DA Martin RW Magam EF Martin JN Maternal mortality associated with HELLP (hemolysis, elevated liver enzymes, and low platelets) syndrome Am J Obstet Gynecol 1999 181 924 8 10521755
Lehohla P Census 2001. Census in Brief 2003 Second Published Statistics South Africa
|
16197546
|
PMC1262774
|
CC BY
|
2021-01-04 16:38:16
|
no
|
Reprod Health. 2005 Sep 30; 2:7
|
utf-8
|
Reprod Health
| 2,005 |
10.1186/1742-4755-2-7
|
oa_comm
|
==== Front
RetrovirologyRetrovirology1742-4690BioMed Central London 1742-4690-2-571617658810.1186/1742-4690-2-57ResearchNatural history of the ERVWE1 endogenous retroviral locus Bonnaud Bertrand [email protected] Jean [email protected] Olivier [email protected] Guy [email protected] Laurent [email protected] François [email protected] UMR 2714 CNRS-bioMérieux, IFR128 BioSciences Lyon-Gerland Ecole Normale Supérieure de Lyon, 46 allée d'Italie, 69364 Lyon cedex 07, France2 Laboratoire de Biométrie et Biologie Evolutive, UMR CNRS 5558, Université Claude Bernard – Lyon 1, 43 Bd du 11 Novembre 1918, 69622 Villeurbanne Cedex, France2005 22 9 2005 2 57 57 21 7 2005 22 9 2005 Copyright © 2005 Bonnaud et al; licensee BioMed Central Ltd.2005Bonnaud et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background
The human HERV-W multicopy family includes a unique proviral locus, termed ERVWE1, whose full-length envelope ORF was preserved through evolution by the action of a selective pressure. The encoded Env protein (Syncytin) is involved in hominoid placental physiology.
Results
In order to infer the natural history of this domestication process, a comparative genomic analysis of the human 7q21.2 syntenic regions in eutherians was performed. In primates, this region was progressively colonized by LTR-elements, leading to two different evolutionary pathways in Cercopithecidae and Hominidae, a genetic drift versus a domestication, respectively.
Conclusion
The preservation in Hominoids of a genomic structure consisting in the juxtaposition of a retrotransposon-derived MaLR LTR and the ERVWE1 provirus suggests a functional link between both elements.
==== Body
Background
The infectious retrovirus founding the contemporary HERV-W family [1] entered the genome of a Catarrhine ancestor 25–40 million years ago [2,3]. The spread of the HERV-W family into the genome essentially results from autonomous and non-autonomous events of intracellular retrotransposition of transcriptionally active copies [4,5]. The HERV-W family contains a unique locus, termed ERVWE1, which encodes an envelope glycoprotein expressed in the placenta [3,6]. This envelope, also dubbed Syncytin, exhibits fusogenic properties in vitro and is directly involved in trophoblast differentiation [6-8]. The functional conservation of the ERVWE1 locus among Hominoids [9] and the identification of selective constraints on the env gene [10] strongly suggest that this retroviral locus has been recruited to play a role in placental physiology. In order to decipher the natural history of the ERVWE1 locus, we performed a comparative genomic analysis of the eutherian chromosomal regions syntenic to a portion of human chromosome 7q21.2 containing the (H)ERVWE1 locus. We observe in this region that the content in transposable elements varies between species, notably with a progressive enrichment of LTR-elements in the Platyrrhine and Catarrhine lineages. Based on an ancestral mosaic of LTR-elements, this retroviral cluster followed two opposed evolutionary pathways, a genetic drift versus a domestication, in Cercopithecidae and Hominidae lineages, respectively.
Results and Discussion
The initial failure to isolate the ERVWE1 integration site in Old World Monkeys [9] suggested that this region was shaped by complex recombination events. The comparative analysis of human ERVWE1 flanking sequences with the mouse genome has revealed two syntenic anchor points in the ERVWE1 provirus vicinity. Thus, the peroxisome biogenesis factor 1 gene (PEX1) and the ocular development-associated gene (ODAG) are located upstream and downstream from ERVWE1, respectively. In genomic databases, the genetic linkage between both boundary genes was found in 14 mammals and 2 birds (Figure 1a). In addition, to fill in the evolutionary gap of this dataset, we PCR amplified and sequenced the intergenic region of two primates, Macaca mulatta and Ateles fusciceps robustus.
Figure 1 Comparative analysis of PEX1-ODAG orthologous locus. (a) Length of PEX1-ODAG intergenic region. Species with an identified PEX1-ODAG intergenic region (either extracted from databases or sequenced in the lab) are indicated on the tree. Clades are redrawn from a previous mammalian phylogeny [23]. Branches are not drawn to scale. The length of the PEX1-ODAG intergenic region is indicated for each species. (b) Length and TEs composition of PEX1 and ODAG intergenic and intronic regions. Species were selected regarding the quality of TEs description in RepBase) [18].
The length of the PEX1-ODAG intergenic region varies among species (17.8 ± 7.9 kb), ranging from 2.6 kb to 30.9 kb for rat and human, respectively (Figure 1a). The length variation of the intergenic region is generally due to the presence of various transposable elements (TEs) (Figure 1b). The particularly short intergenic regions of rodents may result from the general deletion mechanisms previously proposed to account for rodent small genome size [11]. The herein described region suggests that the rodent deletion process show no bias towards TEs (Figure 1b). In comparison, the length of PEX1 and ODAG intronic regions is homogenous (PEX1 : 38.5 ± 13.4 kb ; ODAG : 8.1 ± 2.5 kb), the variability relying mostly upon one species for each gene (Figure 1b). For example, the largest intronic region of PEX1 orthologous gene is observed in Bos taurus and corresponds to the presence of about 40 kb of TEs as compared to 10–20 kb in other species (Figure 1b).
TEs contents differ quantitatively and qualitatively between lineages and between intergenic and intronic regions (Figure 1b). In introns, SINEs then LINEs represent the majority of TEs among all species. The singular large LINE content of Bos taurus PEX1 introns is compatible with the huge amount of specific LINE elements in the genome of this species [12]. The absence of such specific LINE elements in Bos taurus ODAG introns may be due to the shorter length of this gene. Within the intergenic regions, first LINEs and second SINEs predominate in Carnivores, Artiodactyls and Rodents. In primates, the intergenic regions consist largely of LTR elements and Alus. The LTR-elements are clustered in a 20 kb region just downstream from the PEX1 gene and the Alu elements are spread within the 10 kb region upstream from the ODAG gene. This local LTR concentration in primates is particularly high as compared to previous comparative analysis over several megabases [12]. The 30 kb human PEX1-ODAG intergenic region contains 11%, 2% and 64% of Alus, LINE-1s and LTR-elements, respectively.
The picture obtained from the comparison of the syntenic PEX1-ODAG intergenic regions between mammalian species is informative about the putative composition of this region in common ancestors, depicted at the nodes of the phylogenic tree (Figure 2). In addition, LTR-element flanking sequences indicate whether the retrotransposition process was autonomous, i.e. mediated by an HERV-specific reverse transcriptase (RT), or non-autonomous, i.e. mediated by the LINE RT which contributes to pseudogene formation. The autonomous events leads to the duplication of a genomic 4–6 bp sequence, flanking consequently the proviral 5' and 3' LTRs. In the case of LINE RT retrotransposition, a longer flanking repeat of 10–16 bp is observed together with an mRNA typical structure (absence of promoter element and presence of a 3' poly(A) tail) [13,14]. By merging all this information, we infer the natural history of this region.
Figure 2 Phylogenetic analysis of the PEX1-ODAG intergenic region in 9 mammal species. Flanking black boxes correspond to the 24th exon and the 5th exon of the PEX1 and ODAG genes, respectively. LTR elements are depicted as red boxes (MaLR-e1 LTR), green boxes (ERV-P LTR) and empty boxes (ERVWE1 and ERV-H proviruses). The ERVWE1 provirus is labeled W, ERV-H Platyrrhini and Catarrhini lineage specific proviruses are labeled H(p) and H(c), respectively. env smaller boxes refer to the ERVWE1 env gene. Proposed ancestral chromosomal structure are drawn in grey cartouches. The cross-box within the H(c) ancestor represents a pol/env deletion as referenced to the HERV-H repbase consensus. Dash lines represent the evolutionary processes leading to Cercopitheque vs. Hominoid lineages. The double slashes indicate the truncation of longest sequences. Clades are derived from previous phylogeny [23] and branches are not drawn to scale.
The first step of the parsimonious scenario consists in the integration of mammalian apparent LTR-retrotransposon (MaLR) element in the PEX1-ODAG intergenic region of a primitive mammalian ancestor, followed by a local recombination between the 5' and 3' paired LTRs)[15], generating the MaLR isolated LTR. However, the absence among species of flanking duplicated sequences as a vestige of the original integration does not support this hypothesis, although this 100 million years-old signature may have vanished. In human, only two short 57 bp and 106 bp segments were identified (Figure 3), presenting 75.4 % and 67.9% similarity with MLT1J2 and MLT1J subfamilies of MaLR elements)[15], respectively. The 260 bp remaining parts of the MaLR LTR exhibits no similarity with previously defined MaLR consensus sequences, suggesting the identification of a new MaLR subfamily named MaLR-e1. In addition, similarity search (threshold 60%) of MaLR-e1 human and dog sequences on their respective genomes indicate only one other full-length element and a vast majority of elements consisting roughly in either the 5' or the 3' half part of MaLR-e1. The location of one end of these MaLR partial sequences within a 40 bp region (Figure 3) bordered on each side by the MLT1J and MLT1J2 identified regions suggests an authentic chimerical origin for this MaLR-e1 LTR. The paucity of the MaLRs bipartition reflect an unsuccessful propagation of this form. Strikingly, the deduced junction area of both parts of the chimera corresponds to a functional sequence consisting of a trophoblast specific enhancer (TSE) [16].
Figure 3 Alignments of orthologous and paralogous MaLR-e1 LTR sequences from mammalian species. Sequences were assembled using the human sequence (HOM) as reference. Orthologous sequences are from the following origin: HOM: Homo sapiens, PAN: Pan troglodytes, GOR: Gorilla gorilla, ORA:Pongo pygmaeus, GIB: Hylobatides pileatus, ATE: Ateles fusciceps robustus, CAL: Callithrix jacchus, OTO: Otolemur garnetti, CAN: Canis familiaris, DAS: Dasypus novemcictus. Hs12, hs5 and hs20 correspond to MaLR-e1 putative paralogous sequences isolated in the human genome. 5' and 3' openboxes corresponds to MLT1J2 and MLT1J Repbase consensuses, respectively. The region with grey background indicates the 5' or 3' boundaries zone of most partial MaLR-e1 in human and dog genomes. Four putative 3' boundaries of the MaLR-e1 LTR are shown as vertical bars. The double head arrow delimits the trophoblastic specific enhancer (TSE). * correspond to the location of the omitted ERVWE1 provirus. Direct repeats flanking the ERVWE1 integration site are underlined. Bold gray characters in the 3' end of DAS and CAN sequences precede large insertions (1,3 kb and 4,5 kb, respectively) omitted in the alignment.
Second, a 633 bp ERV-P element was acquired by the common ancestor of the Platyrrhines and Catarrhines more than 40 million years ago [17]. As for the MaLR-e1 element, the absence of trivial duplication of the integration site shades the origin of the contemporary isolated ERV-P LTRs. In any case, the putative primary recombination between paired LTRs may have occurred rapidly after integration as no ERV-P internal sequence can be detected in any of the studied species. The LTR sequence is complete as referred to the consensus sequence)[18], although the 5' first ten nucleotides largely diverged.
Third, ERV-H and ERV-W proviruses integrated in the germ line of a Catarrhine ancestor, within the ERV-P and MaLR-e1 LTRs, respectively. Note that an ERV-H sequence is identified in the Platyrrhines (ERV-H(p)), distinct from the Catarrhines ERV-H provirus (ERV-H(c)) described above, as located about 2 kb upstream from the ERV-P LTR. The ERV-W element corresponds to the ERVWE1 provirus as it contains the locus-specific signature (a 12 bp deletion in the 3' end of the env gene) previously identified by comparing (H)ERVWE1 and paralogous HERV-W copies [10]. The presence in several species of degenerated direct repeat at both ends of ERV-H(c) [A(C/T)(G/A)AC] and ERVWE1 [CA(A/G)(C/T)] proviruses attests that retrovirus-like integration events occurred. Whether these proviral insertions derived from re-infection or cis- or trans-retrotransposition processes remains unknown. Nevertheless, the duplication of the integration site indicates the existence at that time of functional H- and W-specific reverse transcriptases. The accumulation of independent substitutions in 5' and 3' paired LTRs, identical when the provirus integrated, is informative about the chronology of these events. Thus, the comparison of paired LTRs distances between the ERV-H(c) and the ERVWE1 proviruses (0.84 and 0.65, respectively) suggests that ERV-H(c) integrated earlier than ERVWE1.
Then the Catarrhine ancestor genomic structure followed two divergent evolutionary pathways in Cercopitheques and Hominoids (Figure 2). An about 9 kb fragment was deleted in the Cercopitheque lineage, consisting of a 3.8 kb pol-env-LTR ERV-H(c) sequence, a 4.3 kb LTR-gag-pol ERVWE1 sequence and the 0.9 kb inter-proviral region. This large deletion produced an hybrid ERV-(H/W) defective proviral structure. Surprisingly, as both ERV-H(c) 5' and ERVWE1 3' flanking sequences were also deleted, the Cercopitheque lineage is devoid of MaLR-e1 and ERV-P LTRs elements. This global inactivation of all four LTR elements was followed by the genetic drift of the env gene as revealed by the presence of different inactivating substitutions in the baboon and macaque ERVWE1 remnants, a stop codon in position 181 and a frameshift in position 498, respectively. In Hominoids, the overall 30 kb structure was preserved as confirmed by overlapping LD-PCR amplification of gorilla, orangutan and gibbon genomic DNA (data not shown). In Hominoids, the ERV-H(c) element contains a locus specific signature that consists in a unique pol-env junction. An accurate dating of this deletion event would require an extended panel of species as the region of interest is absent from the Macaca mulatta and Papio anubis genomes. The presence of the env 12 bp deletion (crucial for the Env fusogenic activity) in Hominoids [10] and Cercopitheques ERVWE1 proviruses suggests that this deletion occurred originally in a primary Catarrhine ancestor possibly soon after integration, in the youth of the ERV-W family. Furthermore, the ERVWE1 env signature was found to be unique in human and chimpanzee genomes, what shows an absence of retrotransposition of this element. This suggests an absence of expression of the ERVWE1 locus in the Hominoid germ line, as opposed to many other HERV-W loci that were shown to retrotransposed using mainly LINE-RT [5].
ERVWE1 was shown to be a bona fide gene involved in hominoid placental physiology [9]. The concomitant conservation in Hominoids of the surrounding LTR elements suggests that they were either required for ERVWE1 activity or hitchhiked during the purifying ERVWE1 selection process [10]. The substitution profile along the whole region does not rule out any hypotheses. Nevertheless, it reveals the strict identity of the MaLR-e1 portion located upstream from ERVWE1 in human, chimpanzee and gorilla, as opposed to a MaLR-e1 3' part different for each species. The regulation of the expression of ERVWE1 env was shown to be a bipartite element [16] composed of (i) a cyclic AMP (cAMP)-inducible retroviral promoter, the ERVWE1 5' LTR, and (ii) a 436 bp upstream regulatory element (URE), encompassing the MaLR-e1 5' part, that contains the trophoblast specific enhancer (TSE) cited above, conferring high level of expression and placental tropism [16]. Although efficient, the cooperation between the URE and the LTR seemed complex due to an interference phenomenon, probably resulting from the presence of AP-2 and Sp-1 binding sites on the TSE and the cAMP-responsive elements of the LTR [16]. Interestingly, the gibbon transcriptional regulatory elements shows an in vitro biased behavior as compared to human, chimpanzee, gorilla and orangutan orthologous elements, i.e. the ERVWE1 5' LTR exhibits a higher placental promoter activity [9] and the URE is deficient in enhancer activity [16]. This feature of the gibbon URE seems associated with two specific mutations in AP-2 and Sp-1, an enhancer activity equivalent to the human one being restored after the modification of the two corresponding residues [16]. Although we cannot exclude the possibility that these observations are partially due to the specific context of a human trophoblastic cell line, this functional analysis supports the very recent recruitment of the elderly MaLR-e1 5' half as proposed in this work. Thus, a LTR of retrotransposon MaLR element and a LTR of a (H)ERV-W proviral locus were co-opted to regulate syncytin expression in placenta. Interestingly, the newly identified murine syncytin-B env gene which triggers cell-cell fusion in vitro and is expressed specifically in placenta in vivo displays an upstream MaLR LTR [19]. Whether this represents an additional element to the puzzling convergent physiological role of primate and rodent syncytins remains to be determined.
Conclusion
We observe in the region syntenic to a portion of human chromosome 7q21.2 containing the (H)ERVWE1 locus a progressive enrichment of LTR-elements in the Platyrrhine and Catarrhine lineages. Based on an ancestral mosaic of LTR-elements, two opposed evolutionary pathways are followed, a genetic drift versus a domestication, in Cercopithecidae and Hominidae lineages, respectively. The domestication process includes the ERVWE1 locus in Hominoid species, and putatively a retrotransposon-derived MaLR LTR strictly conserved in the Homo/Pan/Gorilla subgroup. We propose that both elements were recruited to achieve the regulation of syncytin expression in placenta.
Methods
Syntenic sequences to PEX1-ODAG intergenic regions are extracted from the high throughput genomic sequences (HTGS) division of GenBank using BLAST [20]. The query sequence is composed of exons of PEX1 and ODAG genes, as described in the ensembl repository as vega transcript OTTHUMT00000060247 and OTTHUMG00000023913, respectively. We obtain the following GenBank accession nos., [GenBank:AC092510.2]: Papio anubis, [GenBank:AC148267.2] and [GenBank:AC148269.3]: Callithrix jacchus, [GenBank:AC148127.3] and [GenBank:AC149006.1]: Otolemur garnettii, [GenBank:AC147739.3]: Dasypus novemcinctus, [GenBank:AC148524.3]: Rhinolophus ferrumequinum, [GenBank:AC145009.2] and [GenBank:AC108896.2]: Bos taurus, [GenBank:AC105371.2]: Sus scrofa, [GenBank:AC147729.2]: Oryctolagus cuniculus, [GenBank:AC148352.2]: Sorex araneus, [GenBank:AC097829.7], [GenBank:AC079989.2], [GenBank:AC127809.3] and [GenBank:AC079998.2]: Rattus norvegicus, [GenBank:AC092872.2]: Pan troglodytes, [GenBank:AC114335.3]: Canis familiaris, [GenBank:AC148249.3]: Otolemur garnettii, [GenBank:AC148380.2] and [GenBank:AC148379.2]: Taeniopygia guttata, [GenBank:AC148423.3] and [GenBank:AC148421.2]: Meleagris gallopavo, [GenBank:AC138736.2]: Gallus gallus.
We use RepeatMasker (Smit, AFA, Hubley, R & Green, P. RepeatMasker Open-3.0. 1996–2004 ) to identify transposable elements in all the studied species. Sequence alignments were computed with ClustalW [21] and refined manually using Seaview [22].
We have sequenced Ateles fusciceps robustus and Macaca mulatta genomic PEX1-ODAG region. Sequences are provided in genomic databases with the following accession number : [GenBank:AY925147] for Ateles fusciceps robustus and [GenBank:AY925148] for Macaca mulatta.
List of Abbreviations
HERV: human endogenous retrovirus
ORF: open reading frame
LTR: long terminal repeat
MaLR: mammalian apparent LTR-retrotransposon
SINE: short interspersed element
LINE: long interspersed element
LD-PCR: long distance PCR
Competing interests
The author(s) declare that there are no competing interests.
Authors' contributions
BB designed this study and edited the manuscript. JB, OB and GO isolated and sequenced Macaca mulatta and Ateles fusciceps robustus PEX1-ODAG regions. They also participated to the sequence analysis. LD and FM conceived of the study, and participated in its design and coordination and helped to draft the manuscript.
Acknowledgements
BB is supported by a doctoral fellowship from bioMérieux and Centre National de la Recherche Scientifique and a grant from "La fondation pour la recherche médicale (FRM)". The work was partially supported by INTAS 01-0759. We thank G. Hunsmann for Ateles DNA samples.
==== Refs
Blond JL Beseme F Duret L Bouton O Bedin F Perron H Mandrand B Mallet F Molecular characterization and placental expression of HERV-W, a new human endogenous retrovirus family J Virol 1999 73 1175 1185 9882319
Kim HS Takenaka O Crow TJ Isolation and phylogeny of endogenous retrovirus sequences belonging to the HERV-W family in primates J Gen Virol 1999 80 2613 2619 10573154
Voisset C Bouton O Bedin F Duret L Mandrand B Mallet F Paranhos-Baccala G Chromosomal distribution and coding capacity of the human endogenous retrovirus HERV-W family AIDS Res Hum Retroviruses 2000 16 731 740 10826480 10.1089/088922200308738
Costas J Characterization of the intragenomic spread of the human endogenous retrovirus family HERV-W Mol Biol Evol 2002 19 526 533 11919294
Pavlicek A Paces J Elleder D Hejnar J Processed pseudogenes of human endogenous retroviruses generated by LINEs: their integration, stability, and distribution Genome Res 2002 12 391 399 11875026 10.1101/gr.216902. Article published online before print in February 2002
Blond JL Lavillette D Cheynet V Bouton O Oriol G Chapel-Fernandes S Mandrand B Mallet F Cosset FL An envelope glycoprotein of the human endogenous retrovirus HERV-W is expressed in the human placenta and fuses cells expressing the type D mammalian retrovirus receptor J Virol 2000 74 3321 3329 10708449 10.1128/JVI.74.7.3321-3329.2000
Mi S Lee X Li X Veldman GM Finnerty H Racie L LaVallie E Tang XY Edouard P Howes S Syncytin is a captive retroviral envelope protein involved in human placental morphogenesis Nature 2000 403 785 789 10693809 10.1038/35001608
Frendo JL Olivier D Cheynet V Blond JL Bouton O Vidaud M Rabreau M Evain-Brion D Mallet F Direct involvement of HERV-W Env glycoprotein in human trophoblast cell fusion and differentiation Mol Cell Biol 2003 23 3566 3574 12724415 10.1128/MCB.23.10.3566-3574.2003
Mallet F Bouton O Prudhomme S Cheynet V Oriol G Bonnaud B Lucotte G Duret L Mandrand B The endogenous retroviral locus ERVWE1 is a bona fide gene involved in hominoid placental physiology Proc Natl Acad Sci U S A 2004 101 1731 1736 14757826 10.1073/pnas.0305763101
Bonnaud B Bouton O Oriol G Cheynet V Duret L Mallet F Evidence of Selection on the Domesticated ERVWE1 env Retroviral Element Involved in Placentation Mol Biol Evol 2004 21 1895 1901 15254254 10.1093/molbev/msh206
Waterston RH Lindblad-Toh K Birney E Rogers J Abril JF Agarwal P Agarwala R Ainscough R Alexandersson M An P Initial sequencing and comparative analysis of the mouse genome Nature 2002 420 520 562 12466850 10.1038/nature01262
Thomas JW Touchman JW Blakesley RW Bouffard GG Beckstrom-Sternberg SM Margulies EH Blanchette M Siepel AC Thomas PJ McDowell JC Comparative analyses of multi-species sequences from targeted genomic regions Nature 2003 424 788 793 12917688 10.1038/nature01858
Wei W Gilbert N Ooi SL Lawler JF Ostertag EM Kazazian HH Boeke JD Moran JV Human L1 retrotransposition: cis preference versus trans complementation Mol Cell Biol 2001 21 1429 1439 11158327 10.1128/MCB.21.4.1429-1439.2001
Esnault C Maestre J Heidmann T Human LINE retrotransposons generate processed pseudogenes Nat Genet 2000 24 363 367 10742098 10.1038/74184
Smit AF Identification of a new, abundant superfamily of mammalian LTR-transposons Nucleic Acids Res 1993 21 1863 1872 8388099
Prudhomme S Oriol G Mallet F A retroviral promoter and a cellular enhancer define a bipartite element which controls env ERVWE1 placental expression J Virol 2004 78 12157 12168 15507602 10.1128/JVI.78.22.12157-12168.2004
Goodman M Porter CA Czelusniak J Page SL Schneider H Shoshani J Gunnell G Groves CP Toward a phylogenetic classification of Primates based on DNA evidence complemented by fossil evidence Mol Phylogenet Evol 1998 9 585 598 9668008 10.1006/mpev.1998.0495
Jurka J Repbase update: a database and an electronic journal of repetitive elements Trends Genet 2000 16 418 420 10973072 10.1016/S0168-9525(00)02093-X
Dupressoir A Marceau G Vernochet C Benit L Kanellopoulos C Sapin V Heidmann T Syncytin-A and syncytin-B, two fusogenic placenta-specific murine envelope genes of retroviral origin conserved in Muridae Proc Natl Acad Sci U S A 2005 102 725 730 15644441 10.1073/pnas.0406509102
Altschul SF Gish W Miller W Myers EW Lipman DJ Basic local alignment search tool J Mol Biol 1990 215 403 410 2231712 10.1006/jmbi.1990.9999
Thompson JD Higgins DG Gibson TJ CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position specific gap penalties and weight matrix choice Nucleic Acids Res 1994 22 4673 4680 7984417
Galtier N Gouy M Gautier C SEA VIEW and PHYLO_WIN: two graphic tools for sequence alignment and molecular phylogeny Comput Appl Biosci 1996 12 543 548 9021275
Murphy WJ Eizirik E O'Brien SJ Madsen O Scally M Douady CJ Teeling E Ryder OA Stanhope MJ de Jong WW Resolution of the early placental mammal radiation using Bayesian phylogenetics Science 2001 294 2348 2351 11743200 10.1126/science.1067179
|
16176588
|
PMC1262775
|
CC BY
|
2021-01-04 16:36:39
|
no
|
Retrovirology. 2005 Sep 22; 2:57
|
utf-8
|
Retrovirology
| 2,005 |
10.1186/1742-4690-2-57
|
oa_comm
|
==== Front
RetrovirologyRetrovirology1742-4690BioMed Central London 1742-4690-2-581618803510.1186/1742-4690-2-58Short ReportMutations affecting cleavage at the p10-capsid protease cleavage site block Rous sarcoma virus replication Vana Marcy L [email protected] Aiping [email protected] Peter [email protected] Irene [email protected] Dalbinder [email protected] Eric [email protected] Jonathan [email protected] Department of Microbiology and Immunology, Feinberg School of Medicine, Northwestern University, Chicago, Illinois 60611, USA2 Department of Biology, Georgia State University, Atlanta, GA 30303, USA3 Biochemistry and Molecular Biology Department, Medical and Health Sciences Center, University of Debrecen, Debrecen, Hungary4 Vollum Institute and Department of Microbiology, Oregon Health and Science University, Portland, OR, 97201, USA2005 27 9 2005 2 58 58 1 2 2005 27 9 2005 Copyright © 2005 Vana et al; licensee BioMed Central Ltd.2005Vana et al; licensee BioMed Central Ltd.This is an Open Access article distributed under the terms of the Creative Commons Attribution License (), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
A series of amino acid substitutions (M239F, M239G, P240F, V241G) were placed in the p10-CA protease cleavage site (VVAM*PVVI) to change the rate of cleavage of the junction. The effects of these substitutions on p10-CA cleavage by RSV PR were confirmed by measuring the kinetics of cleavage of model peptide substrates containing the wild type and mutant p10-CA sites. The effects of these substitutions on processing of the Gag polyprotein were determined by labeling Gag transfected COS-1 cells with 35S-Met and -Cys, and immunoprecipitation of Gag and its cleavage products from the media and lysate fractions. All substitutions except M239F caused decreases in detectable Gag processing and subsequent release from cells. Several of the mutants also caused defects in production of the three CA proteins. The p10-CA mutations were subcloned into an RSV proviral vector (RCAN) and introduced into a chick embryo fibroblast cell line (DF-1). All of the mutations except M239F blocked RSV replication. In addition, the effects of the M239F and M239G substitutions on the morphology of released virus particles were examined by electron microscopy. While the M239F particles appeared similar to wild type particles, M239G particles contained cores that were large and misshapen. These results suggest that mutations affecting cleavage at the p10-CA protease cleavage site block RSV replication and can have a negative impact on virus particle morphology.
==== Body
Findings
The structural proteins of retroviruses are encoded by the gag gene and are translated as a single polyprotein. During or subsequent to virus budding, the Gag polyprotein is cleaved by the viral protease (PR), thereby releasing the mature structural proteins. Gag processing leads to morphological changes in the virus particle, including condensation of the capsid core, and is associated with the appearance of infectious particles [1]. It has previously been demonstrated that proper processing at several protease sites throughout RSV Gag is required for production of infectious virus [2,3]. However, the protease site separating the C-terminus of p10 and the N-terminus of CA has not been examined.
Multiple studies have highlighted the importance of cleavage at the N-terminus of retrovirus CA proteins in particle assembly and maturation. Structural studies have identified a β hairpin structure at the N-terminus of RSV CA that is thought to form after proteolysis at the p10-CA site and liberation of the N-terminus of CA [4]. Moreover, a conserved Pro residue at the extreme N-terminus of RSV CA forms a salt bridge with an internal Asp residue, thereby stabilizing the β-hairpin structure [4]. These Pro and Asp residues are highly conserved among many retrovirus CA proteins, suggesting that the β-hairpin is a common structural feature of retrovirus CA proteins [5-8]. Mutating the conserved Asp residue in HIV-1 CA (Asp51) or murine leukemia virus CA (MLV, Asp63) causes a loss in virus infectivity [8]. In addition, blocking protease cleavage at the N-terminus of MLV CA results in the production of virus that is non-infectious [9]. It has also been demonstrated that the N-terminus of CA and the residues immediately upstream of CA have a role in determining the shape of assembling retrovirus particles [8,10-13]. More specifically for RSV, it has been demonstrated that the presence of p10 on the N-terminus of CA-NC converts the in vitro assembly phenotype from cylindrical particles to spherical particles that resemble wild type immature RSV particles [10,11].
In this study, amino acid substitutions were made in the first two N-terminal residues of CA and the last C-terminal amino acid of p10 in order to alter cleavage at the p10-CA site and examine the role of p10-CA cleavage in Gag processing and RSV replication (Fig. 1A). Previous studies focusing on the RSV NC-PR or HIV-1 MA-CA cleavage sites showed that substituting Gly at any of the P2-P2' positions resulted in greatly reduced in vitro hydrolysis of the peptides [14,15]. Phe substitutions of P1 provided good cleavage of the RSV NC-PR or HIV-1 MA-CA peptides, while Phe substitutions of P1' were tolerated in the RSV NC-PR peptide, but not in the HIV-1 MA-CA peptide. The ability of RSV PR to cleave peptides containing the p10-CA amino acid substitutions compared to a peptide containing the wild type p10-CA site was tested using an in vitro protease assay [16]. All of the substitutions except M239F led to a decrease in the rate of peptide cleavage (Fig. 1A). Substituting Phe for Met in the P1 position (M239F) had a small stimulatory effect on peptide cleavage by PR, while changing the same Met to Gly (M239G) resulted in a complete block in peptide cleavage. Similarly, changing the P1' Pro to Phe (P240F) caused a severe if not complete loss in peptide cleavage and replacement of the Val in the P2' position with Gly (V241G) resulted in a 200-fold decrease in peptide cleavage. Thus, mutating residues on either side of the cleavage junction significantly altered processing of the site.
Figure 1 A. Schematic diagram of the RSV Gag polyprotein and the amino acid substitutions placed in the p10-CA protease cleavage site within Gag. The rectangle represents the RSV Gag polyprotein with the encoded protein sequences indicated by the standard nomenclature. The horizontal lines represent the PR cleavage sites. SP is the spacer peptide. The L domain of RSV Gag resides in the p2b peptide. In the box below, the P4-P1 and P1'-P4' amino acid sequence of the wild type p10-CA protease cleavage site is shown. The p10-CA mutants (underlined bold text) are shown below the wild type sequence. The results of in vitro protease assays examining RSV PR-mediated cleavage of peptides containing the wild type (PVVAM*PVVIKRR) and mutant p10-CA sites are also indicated. The site of p10-CA cleavage is designated with an asterisk. B. Top, Effect of p10-CA amino acid substitutions on processing of RSV Gag. COS-1 cells were transfected with wild type Gag or the p10-CA mutants in pSV.Myr0(HpaI). 48 hours after transfection, cells were labeled with [35S]-Met and Cys and Gag proteins were immunoprecipitated with an anti-RSV rabbit antiserum from the media (right panel) and lysate (left panel) fractions. Immunoprecipitated proteins were separated by SDS-PAGE and exposed to film. Lane 1, untransfected cells. Cells transfected with wild type, lane 2; M239F, lane 3; M239G, lane 4; P240F, lane 5; V241G, lane 6. B. Bottom. Effect of p10-CA amino acid substitutions on Gag release in the context of a protease inactivating substitution (PR-D37S). COS-1 cells were transfected and full-length Gag proteins were immunoprecipitated and separated by SDS-PAGE as above. Cells transfected with M239F/PR-D37S, lane 1; M239G/PR-D37S, lane 2; P240F/PR-D37S, lane 3; V241G/PR-D37S, lane 4; untransfected cells, lane 5; PR-D37S, lane 6.
The effects of the p10-CA substitutions on Gag processing were tested by introduction of the mutations into the context of full-length Gag and expressing the wild type or mutant Gag proteins in COS-1 cells [2,3]. Gag and its cleavage products were immunoprecipitated from the media and lysate fractions from transfected cells following metabolic labeling and were separated using SDS-PAGE (Fig. 1B, top). By comparison to wild type (Fig. 1B, top lanes 2), all of the p10-CA substitutions except M239F caused processing defects. The banding pattern in the lysate and media fractions from cells transfected with M239F (Fig. 1B, top, lanes 3) was very similar to wild type, suggesting that the M239F substitution did not affect Gag processing. In contrast, a novel and stable band representing a p10-CA fusion protein was present in the lysate and media fractions from cells transfected with the M239G (Fig. 1B, top, lanes 4) and P240F (Fig. 1B, top, lanes 5) mutants that was not present in fractions from cells transfected with wild type Gag (lanes 2 top). The presence of a p10-CA fusion indicated that these mutations resulted in a reduction in the ability of PR to cleave the p10-CA site within Gag.
In cells transfected with wild type Gag, three CA species were detected (CA1, CA2, and CA3) in the media and lysate fractions (Fig. 1B, top, lanes 2) [2,3]. These species are the result of processing of CA at its C-terminus at different sites. In contrast, in cells transfected with the M239G mutant, CA2 and CA3 were detected in the media fraction, but CA1 was not (Fig. 1B, top, lanes 4). Furthermore, mature CA proteins were not detected in the lysate. Similarly, none of the mature CA proteins were detected in the media or lysate fractions from cells transfected with the P240F (Fig. 1B, top, lanes 5) mutant, and CA1 made up the majority of the CA protein in the media and lysate fractions from cells transfected with the V241G (Fig. 1B, top, lanes 6) mutant. There also appeared to be a reduction in the amount of Gag released into the media from cells transfected with the V241G mutant compared to cells transfected with wild type Gag (Fig. 1B, top, lanes 6 and 2). This effect was most apparent when examining the signal of PR in the lysate and media fractions. The amount of PR in the lysate fraction from cells transfected with the V241G mutant was similar to wild type, but the amount of PR in the media fraction from cells transfected with the V241G mutant was greatly reduced compared to wild type. In order to determine whether the reduction in particle release observed with the V241G mutant was due to impaired Gag processing, a D37S mutation in the PR domain was constructed in the context of the p10-CA Gag mutants. COS-1 cells were transfected with the p10-CA/PR-D37S mutants and full-length Gag was immunoprecipitated from the media and lysate fractions. A similar level of Gag release was observed with all of the p10-CA/PR-D37S mutants when compared to PR-D37S (Fig. 1B, bottom), suggesting that the particle release defect observed with the V241G substitution was due to impaired Gag processing. Taken together, these results indicate that mutations to the p10-CA site of Gag affect processing of the C-terminus of CA.
In order to determine the effects of the p10-CA substitutions on RSV replication, the p10-CA mutations were sub-cloned into the RCAN proviral vector [17]. DF-1 cells were transfected with each of the mutants, and reverse transcriptase (RT) activity was monitored in the media of transfected cells at regular intervals [18]. All of the p10-CA mutations except M239F had a detrimental effect on RSV replication (Fig. 2). The M239F mutation caused an initial delay in replication with an approximate four-fold reduction in RT activity but reached a similar peak in virus production to wild type by day six. In contrast, all of the other p10-CA mutations led to a severe block in viral replication (Fig. 2). The RT activity of these mutants could not be detected above control levels of 5TE buffer (data not shown), media from untransfected cells (data not shown), or media from cells transfected with an L domain deletion mutant (ΔPY/RCAN).
Figure 2 Effect of p10-CA substitutions on ability of RSV to replicate in transfected DF-1 cells. DF-1 cells were transfected with wild type RCAN, RCAN constructs containing the p10-CA mutations, or an RCAN construct containing an L domain deletion (ΔPY). At the indicated times after transfection, the RT activity in the culture medium was determined by quantification of [α-32P]-dTTP incorporation during reverse transcription using a polyadenylic acid (poly rA) template and a oligodeoxythymidylate (p(dT)12–18) primer. Wild type (◇), L-domain deletion (□), M239F (X), M239G (*), P240F (-), and V241G (+).
To better understand the effect of the p10-CA mutations on RSV replication, wild type and p10-CA virus particles were examined using electron microscopy. Virus particles were harvested three days post transfection and viewed at a magnification of 11,000× (Fig. 3, left) and 37,000× (Fig. 3, right). We were only able to examine wild type, M239F and M239G particles by EM, as we were unable to obtain high enough amounts of P240F and V241G particles. M239F particles appeared to be similar to wild type particles in diameter (wild type; 119 ± 7 nm, MF; 118 ± 11 nm). The ratio of the cross-sectional areas of the virus core and the entire virus particle were also similar between the wild type (Fig. 3, top left and right) and M239F (Fig. 3, middle left and right) particles (wild type; 28 ± 3%, MF; 26 ± 3%). In contrast, the M239G particles (Fig. 3, bottom left and right) were larger in diameter (MG; 125 ± 5 nm) compared to the wild type and M239F particles, and had a higher ratio of core to particle cross-sectional area (MG; 45 ± 5%). It is likely that the defect in particle morphology observed with the M239G mutant played a role in the loss of replication capacity of this mutant. Together, these results highlight the importance of proper processing at the p10-CA site in RSV replication, and support previous findings demonstrating the importance of this region in retrovirus replication [4-13].
Figure 3 Effect of p10-CA mutations on virus particle morphology. WT (top left and right), M239F (middle left and right), and M239G (bottom left and right) viruses from transfected cells were sedimented through 20% sucrose cushions, resuspended, and processed for electron microscopy. At low magnification (left; top, middle and bottom), WT and M239F cores appeared conical or bullet-shaped, whereas M239G cores sometimes appeared conical (left-bottom, leftmost virus), but more often appeared with large misshapen cores. At higher magnification (right; top, middle and bottom), internal cores were difficult to discern without significant adjustment of image contrast levels. Size bars for the two magnifications of images appear in bottom left and right panels, and correspond to 100 nm.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
M. L. V. constructed the p10-CA mutations, performed the Gag processing and replication assays, purified virus particles for EM analysis, and wrote the paper. A. C. constructed the D37S mutations and performed the budding assay with the D37S mutants. P. B. performed the in vitro protease assay. D. C. and E. B. performed the EM analysis.
Acknowledgements
This work was supported in part by United States Public Health Service grant CA52047 (to J.L.), CA58166 (to I. W.), and GM60170 (to E.B.), the Hungarian Science and Research Fund, OTKA F35191 (to P.B.), and the Cancer Biology Fellowship Program, Chicago Baseball Cancer Charities, from the Robert H. Lurie Comprehensive Cancer Center (to M.L.V.). Peptides were a generous gift of Dr. Terry Copeland, NCI, Frederick, Maryland.
==== Refs
Wills JW Craven RC Form, function, and use of retroviral gag proteins Aids 1991 5 639 654 1883539
Xiang Y Ridky TW Krishna NK Leis J Altered Rous sarcoma virus Gag polyprotein processing and its effects on particle formation J Virol 1997 71 2083 2091 9032340
Xiang Y Thorick R Vana ML Craven R Leis J Proper processing of avian sarcoma/leukosis virus capsid proteins is required for infectivity J Virol 2001 75 6016 6021 11390603 10.1128/JVI.75.13.6016-6021.2001
Kingston RL Fitzon-Ostendorp T Eisenmesser EZ Schatz GW Vogt VM Post CB Rossmann MG Structure and self-association of the Rous sarcoma virus capsid protein Structure Fold Des 2000 8 617 628 10873863 10.1016/S0969-2126(00)00148-9
Gamble TR Vajdos FF Yoo S Worthylake DK Houseweart M Sundquist WI Hill CP Crystal structure of human cyclophilin A bound to the amino-terminal domain of HIV-1 capsid Cell 1996 87 1285 1294 8980234 10.1016/S0092-8674(00)81823-1
Gitti RK Lee BM Walker J Summers MF Yoo S Sundquist WI Structure of the amino-terminal core domain of the HIV-1 capsid protein Science 1996 273 231 235 8662505
Momany C Kovari LC Prongay AJ Keller W Gitti RK Lee BM Gorbalenya AE Tong L McClure J Ehrlich LS Summers MF Carter C Rossmann MG Crystal structure of dimeric HIV-1 capsid protein Nat Struct Biol 1996 3 763 770 8784350 10.1038/nsb0996-763
von Schwedler UK Stemmler TL Klishko VY Li S Albertine KH Davis DR Sundquist WI Proteolytic refolding of the HIV-1 capsid protein amino-terminus facilitates viral core assembly Embo J 1998 17 1555 1568 9501077 10.1093/emboj/17.6.1555
Oshima M Muriaux D Mirro J Nagashima K Dryden K Yeager M Rein A Effects of blocking individual maturation cleavages in murine leukemia virus gag J Virol 2004 78 1411 1420 14722296 10.1128/JVI.78.3.1411-1420.2004
Campbell S Vogt VM In vitro assembly of virus-like particles with Rous sarcoma virus Gag deletion mutants: identification of the p10 domain as a morphological determinant in the formation of spherical particles J Virol 1997 71 4425 4435 9151833
Joshi SM Vogt VM Role of the Rous sarcoma virus p10 domain in shape determination of gag virus-like particles assembled in vitro and within Escherichia coli J Virol 2000 74 10260 10268 11024160 10.1128/JVI.74.21.10260-10268.2000
Gross I Hohenberg H Huckhagel C Krausslich HG N-Terminal extension of human immunodeficiency virus capsid protein converts the in vitro assembly phenotype from tubular to spherical particles J Virol 1998 72 4798 4810 9573245
Rumlova-Klikova M Hunter E Nermut MV Pichova I Ruml T Analysis of Mason-Pfizer monkey virus Gag domains required for capsid assembly in bacteria: role of the N-terminal proline residue of CA in directing particle shape J Virol 2000 74 8452 8459 10954545 10.1128/JVI.74.18.8452-8459.2000
Tozser J Bagossi P Weber IT Copeland TD Oroszlan S Comparative studies on the substrate specificity of avian myeloblastosis virus proteinase and lentiviral proteinases J Biol Chem 1996 271 6781 6788 8636100 10.1074/jbc.271.12.6781
Cameron CE Grinde B Jacques P Jentoft J Leis J Wlodawer A Weber IT Comparison of the substrate-binding pockets of the Rous sarcoma virus and human immunodeficiency virus type 1 proteases J Biol Chem 1993 268 11711 11720 8389361
Mahalingam B Louis JM Reed CC Adomat JM Krouse J Wang YF Harrison RW Weber IT Structural and kinetic analysis of drug resistant mutants of HIV-1 protease Eur J Biochem 1999 263 238 245 10429209 10.1046/j.1432-1327.1999.00514.x
Hughes SH Greenhouse JJ Petropoulos CJ Sutrave P Adaptor plasmids simplify the insertion of foreign DNA into helper-independent retroviral vectors J Virol 1987 61 3004 3012 3041020
Miller JT Ge Z Morris S Das K Leis J Multiple biological roles associated with the Rous sarcoma virus 5' untranslated RNA U5-IR stem and loop J Virol 1997 71 7648 7656 9311847
|
16188035
|
PMC1262776
|
CC BY
|
2021-01-04 16:36:38
|
no
|
Retrovirology. 2005 Sep 27; 2:58
|
utf-8
|
Retrovirology
| 2,005 |
10.1186/1742-4690-2-58
|
oa_comm
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.